Search

Image segmentation applied to medical image analysis is still a critical and important task. Although there exist several segmentation algorithms that have been widely studied in literature, these are subject to segmentation problems such as over- and under-segmentation as well as non-closed edges. In this paper, a simple method that combines well-known segmentation algorithms is presented. This method is applied to detect acid-fast bacilli (AFB) in bacilloscopies used to diagnose pulmonary tuberculosis (TB). This diagnosis can be performed through different tests, and the most used worldwide is smear microscopy because of its low cost and effectiveness. This diagnosis technique is based on the analysis and counting of the bacilli in the bacilloscopy observed under an optical microscope. The proposed method is used to segment the bacilli in digital images from bacilloscopies processed using Ziehl-Neelsen (ZN) staining. The proposed method is fast, has a low computational cost and good efficiency compared to other methods. The bacilli image segmentation is performed by image processing and analysis techniques, probability concepts and classifiers. In this work, a Bayesian classifier based on a Gaussian mixture model (GMM) is used. The segmentations' results are validated by using the Jaccard index, which indicates the efficiency of the classifier. [ABSTRACT FROM AUTHOR]

Different blade species have different blade characteristics information, whether it is based on leaf information to identify tree species or accurately detect leaf diseases, image segmentation is particularly important. This paper proposes a blade image segmentation method based on the new addition algorithm, which is used to add the edge contour of the binary image extracted by the Canny detector to the original gray image, and the inverted binary image is quadratic with the first added image. Before and after adding, the image format and size of the blade image should be kept consistent, and the image addition condition can be satisfied to realize the segmentation of the blade image. The experimental results show that the method retains the complete feature information of the blade and the blade is completely separated from the background. It provides solid theoretical support for subsequent image feature extraction and tree species identification, the addition method can also be applied to other fields of image segmentation. [ABSTRACT FROM AUTHOR]

As one of the most popular image segmentation techniques, multi-level thresholding is widely used. Much work has been done to improve the efficiency of multi-level thresholding, but satisfied affect is hard to achieve. In this paper, a multi-level image segmentation method using between-class variance (Otsu) based on improved bat algorithm (DWBA) with dynamically adjusting inertia weight and velocity stratification theory is proposed. DWBA algorithm has strong global search ability at the beginning. Then, the local search capability is enhanced with numbers increasing of iterations. More importantly, the performance of DWBA further improved, because bats with different fitness values have diverse velocities. Furthermore, an improved local search strategy is proposed to avoid the current best solution being replaced during iterations. The experimental results established that the proposed DWBA algorithm obtains better outcome than other algorithms. [ABSTRACT FROM AUTHOR]

Lung cancer is one of the most diagnosable forms of cancer worldwide. The early diagnoses of pulmonary nodules in computed tomography (CT) chest scans are crucial for potential patients. Recent researches have showed that the methods based on deep learning have made a significant progress for the medical diagnoses. However, the achievements on identification of pulmonary nodules are not yet satisfactory enough to be adopted in clinical practice. It is largely caused by either the existence of many false positives or the heavy time of processing. With the development of fully convolutional networks (FCNs), in this study, we proposed a new method of identifying the pulmonary nodules. The method segments the suspected nodules from their environments and then removes the false positives. Especially, it optimizes the network architecture for the identification of nodules rapidly and accurately. In order to remove the false positives, the suspected nodules are reduced using the 2D models. Furthermore, according to the significant differences between nodules and non-nodules in 3D shapes, the false positives are eliminated by integrating into the 3D models and classified via 3D CNNs. The experiments on 1000 patients indicate that our proposed method achieved 97.78% sensitivity rate for segmentation and 90.1% accuracy rate for detection. The maximum response time was less than 30 s and the average time was about 15 s. Graphical Abstract This paper has proposed a new method of identifying the pulmonary nodules. The method segments the suspected nodules from CT images and removes the false positives. As shown in the above, the proposed approach consists of three stages. In stage I, raw data are filtered and normalized. The clean normalized data are then segmented in stage II to extract the suspected nodular lesions through 2D FCNs. Stage III is to remove some false positives generated at stage II via 3D CNNs and outputs the final results. The experiments on 1000 patients indicate that our proposed method has achieved 97.78% sensitivity rate for segmentation and 90.1% accuracy rate for detection. The maximum response time was less than 30 s and the average time was about 15 s. [ABSTRACT FROM AUTHOR]

In order to realize the visual positioning for Hangzhou white chrysanthemums harvesting robot in natural environment, a color image segmentation method for Hangzhou white chrysanthemum based on least squares support vector machine (LS-SVM) was proposed. Firstly, bilateral filter was used to filter the RGB channels image respectively to eliminate noise. Then the pixel-level color feature and texture feature of the image, which was used as input of LS-SVM model (classifier) and SVM model (classifier), were extracted via RGB value of image and gray level co-occurrence matrix. Finally, the color image was segmented with the trained LS-SVM model (classifier) and SVM model (classifier) separately. The experimental results showed that the trained LS-SVM model and SVM model could effectively segment the images of the Hangzhou white chrysanthemums from complicated background taken under three illumination conditions such as front-lighting, back-lighting and overshadow, with the accuracy of above 90%. When segmenting an image, the SVM algorithm required 1.3 s, while the LS-SVM algorithm proposed in this paper just needed 0.7 s, which was better than the SVM algorithm obviously. The picking experiment was carried out and the results showed that the implementation of the proposed segmentation algorithm on the picking robot could achieve 81% picking success rate. [ABSTRACT FROM AUTHOR]

Clustering large and complex data sets whose partitions may adopt arbitrary shapes remains a difficult challenge. Part of this challenge comes from the difficulty in defining a similarity measure between the data points that captures the underlying geometry of those data points. In this paper, we propose an algorithm, DCG++ that generates such a similarity measure that is data-driven and ultrametric. DCG++ uses Markov Chain Random Walks to capture the intrinsic geometry of data, scans possible scales, and combines all this information using a simple procedure that is shown to generate an ultrametric. We validate the effectiveness of this similarity measure within the context of clustering on synthetic data with complex geometry, on a real-world data set containing segmented audio records of frog calls described by mel-frequency cepstral coefficients, as well as on an image segmentation problem. The experimental results show a significant improvement on performance with the DCG-based ultrametric compared to using an empirical distance measure. [ABSTRACT FROM AUTHOR]

Interval valued fuzzy c-means (IVFCM) clustering algorithm is one of effective clustering algorithms. When applied to image segmentation, IVFCM includes three problems as follows: (1) It is sensitive to the initial values of algorithm and may easily fall into the local optimal. (2) The algorithm is sensitive to the image noise and cannot obtain the satisfying performance on images corrupted by noise. (3) It always performs image segmentation under one objective function, therefore it cannot meet multiple practical needs. In order to address these problems, a multi-objective interval valued fuzzy clustering algorithm is proposed in this paper. This method constructs two novel interval valued fuzzy fitness functions which utilize the non-local spatial information of the image. Then a new mutation operator combining the interval valued fuzzy information of image is designed. Furthermore, an effective interval valued fuzzy cluster validity index using the non-local spatial information of image is presented to select a single solution from the non-dominated solution set. Experimental results show that the proposed method behaves well in noisy image segmentation. [ABSTRACT FROM AUTHOR]

A plethora of algorithms have been defined for foreground segmentation, a fundamental stage for many computer vision applications. In this paper, we propose a post-processing framework to improve the foreground segmentation performance of background subtraction algorithms. We define a hierarchical framework for extending segmented foreground pixels to undetected foreground object areas and for removing erroneously segmented foreground. First, we create a motion-aware hierarchical image segmentation of each frame that prevents merging foreground and background image regions. Then, we estimate the quality of the foreground mask through the fitness of the binary regions in the mask and the hierarchy of segmented regions. Finally, the improved foreground mask is obtained as an optimal labeling by jointly exploiting foreground quality and spatial color relations in a pixel-wise fully connected conditional random field. Experiments are conducted over four large and heterogeneous data sets with varied challenges (CDNET2014, LASIESTA, SABS, and BMC) demonstrating the capability of the proposed framework to improve background subtraction results. [ABSTRACT FROM AUTHOR]

Image segmentation is a very important activity in computer vision, where critical applications are highly dependent on the efficacy of such activity. To enhance the efficiency of such automated activity, meta-heuristic algorithms to optimally elucidate multi-level image segmentation problems have been proposed in the literature. Because of the advantages in terms of efficiency and convergence speed of the bat algorithm, this paper presents a novel improvement of such algorithm for solving the image multi-thresholding problem. The algorithm leads to speed up the convergence and increase diversity through the utilization of an appropriate crossover operator and chaotic sequences, with the use of Kapur's entropy as the optimized objective function. The proposed method produces segmented images with optimal values for the threshold in few iterations. Through the comparative analysis based on standard deviation, peak signal to noise ratio (PSNR) and segmented image quality, it is observed that the effectiveness of the proposed method, validated using different standard test images, outperforms well-known metaheuristic-based optimization techniques. [ABSTRACT FROM AUTHOR]

In this paper, a new variational model is proposed for image segmentation based on active contours, nonlinear diffusion and level sets. It includes a Chan-Vese model-based data fitting term and a regularized term that uses the potential functions (PF) of nonlinear diffusion. The former term can segment the image by region partition instead of having to rely on the edge information. The latter term is capable of automatically preserving image edges as well as smoothing noisy regions. To improve computational efficiency, the implementation of the proposed model does not directly solve the high order nonlinear partial differential equations and instead exploit the efficient alternating direction method of multipliers (ADMM), which allows the use of fast Fourier transform (FFT), analytical generalized soft thresholding equation, and projection formula. In particular, we creatively propose a new fast algorithm, normal vector projection method (NVPM), based on alternating optimization method and normal vector projection. Its stability can be the same as ADMM and it has faster convergence ability. Extensive numerical experiments on grey and colour images validate the effectiveness of the proposed model and the efficiency of the algorithms. [ABSTRACT FROM AUTHOR]

Highlights • Propose indirect diffusion model for extraction of deeply concave edges and sharp corners. • Model is defined as two PDEs, one driving level set motion and one smoothing indirectly level set. • Level set function can be initialized to zero function, so allowing for flexible initialization. • It is easy to set a termination criterion for the proposed algorithm. Abstract In this paper, we put forward an idea of indirect diffusion and further develope an indirect diffusion-based level set model for image segmentation. This model is based on the dynamic process of diffusion that is posed indirectly on level set function by way of auxiliary function, coupled with a transition region-based force that exhibits the desired sign-changing property. It is formulated as a coupled system of two evolution equations, in which the first equation drives the motion of zero level set toward the object edges and makes it possible to set a termination criterion on the algorithm, while the second equation (indirect diffusion) smoothens the auxiliary function and keeps the auxiliary function as close to the level set function as possible. The derived model can effectively be solved purely by the simplest explicit finite difference. Experimental results show that the proposed model not only has the strong capability of noise immunity, but it also can much better conduce to extraction of deeply concave edges and preservation of sharp corners, compared with the direct diffusion-based counterpart. [ABSTRACT FROM AUTHOR]

Fuzzy C-means has been adopted for image segmentation, but it is sensitive to noise and other image artifacts due to not considering neighbor information. In order to solve this problem, many improved algorithms have been proposed, such as fuzzy local information C-means clustering algorithm (FLICM). However, the segmentation results of FLICM are unsatisfactory when performed on complex images. To overcome this, a novel fuzzy clustering algorithm is proposed in this paper, and more information is utilized to guide the procedure of image segmentation. In the proposed algorithm, pixel relevance based on patch similarity will be investigated firstly, by which all information over the whole image can be considered, not limited to local context. Compared with Zhang et al. (Multimed Tools Appl 76(6):7869–7895, 2017a. 10.1007/s11042-016-3399-x) pixel relevance is unnecessary to be normalized, and much more information can play positive role in image segmentation. Experimental results show that the proposed algorithm outperforms current fuzzy algorithms, especially in enhancing the robustness of corresponding fuzzy clustering algorithms. [ABSTRACT FROM AUTHOR]

A shape prior-based object segmentation is developed in this paper by using a shape transformation distance to constrain object contour evolution. In the proposed algorithm, the transformation distance measures the dissimilarity between two unaligned shapes by cyclic shift, which is called 'circulant dissimilarity'. This dissimilarity with respect to transformation of the object shape is represented by circular convolution, which could be efficiently computed by using fast Fourier transform. Given a set of training shapes, the kernel density estimation is adopted to model shape prior. By integrating low-level image feature, high-level shape prior and transformation distance, a variational segmentation model is proposed to solve the transformation invariance of shape prior. Numerical experiments demonstrate that circulant dissimilarity-based shape registration outperforms the iterative optimization on explicit pose parameters, and show promising results and highlight the potential of the method for object registration and segmentation. [ABSTRACT FROM AUTHOR]

Abstract Uncertainty is a fundamental and unavoidable feature in daily life, which is the same for a single classifier. Thus, combining the predictions of many different classifiers is a very successful way to reduce the uncertainty. In this paper, we present a Correcting Reliability Level (CRL) supervised three-way decision (3WD) cascade model to implement image classification tasks. Our model simulates the human decision process by using 3WD to judge "certainty" or "uncertainty" of the classification result. When judged as "uncertainty", CRL will supervise the 3WD and learn more information to make the final prediction. In addition, we introduce two Class Grouping methods to mining the relation between labels, which help us to train several expert ConvNets for different types of images. Experimental results show that our model can effectively reduce the classification error rate compared with the base classifier. [ABSTRACT FROM AUTHOR]

Abstract RANSAC is a popular robust model estimation algorithm in various computer vision applications. However, the speed of RANSAC declines dramatically as the inlier rate of the measurements decreases. In this paper, a novel Adaptively Ranked Sample Consensus(ARSAC) algorithm is presented to boost the speed and robustness of RANSAC. The algorithm adopts non-uniform sampling based on the ranked measurements to speed up the sampling process. Instead of a fixed measurement ranking, we design an adaptive scheme which updates the ranking of the measurements, to incorporate high quality measurements into sample at high priority. At the same time, a geometric constraint is proposed during sampling process to select measurements with scattered distribution in images, which could alleviate degenerate cases in epipolar geometry estimation. Experiments on both synthetic and real-world data demonstrate the superiority in efficiency and robustness of the proposed algorithm compared to the state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Weather describes the condition of our atmosphere during a specific period of time, and climate represents a composite of day to day weather over longer period of time. Climatology attempts to analyze and explain the impact of climate so that the society can plan accordingly. Climatology analysis is often done on radar images representing various climatic conditions. These images contain varying scale of severity for any specific climatic parameter of study. The climatologists often find it convenient to analyze climatic conditions if tools are available to segment the weather images based on the severity scale which is represented by different colors. Segmentation of the weather radar image is also used for automated analysis of weather conditions. Differential evolution (DE) approach instead is used for fast selection of optimal threshold. In present paper, we have applied DE with multilevel thresholding for weather image segmentation which results in minimum computational time and excellent image quality. A new mutation strategy for DE named reconstructed differential evolution (RDE) strategy is suggested for better performance over image segmentation. Using fuzzy entropy and RDE for multilevel thresholding provides better results in comparison with last suggested methods. [ABSTRACT FROM AUTHOR]

Highlights • A unified deep learning framework for segmentation of objects (cell nuclei, cells, and multi-cellular objects such as glandular structures) in two main types of microscopy images: fluorescence and histology. • The proposed Micro-Net is aimed at better object localization in the face of varying intensities and is robust to noise. • Detailed experimentation & comparative evaluation on publicly available data sets and a new image dataset that is made public with this paper. • Demonstration of robustness of the algorithm to high levels of noise. Graphical abstract Abstract Object segmentation and structure localization are important steps in automated image analysis pipelines for microscopy images. We present a convolution neural network (CNN) based deep learning architecture for segmentation of objects in microscopy images. The proposed network can be used to segment cells, nuclei and glands in fluorescence microscopy and histology images after slight tuning of input parameters. The network trains at multiple resolutions of the input image, connects the intermediate layers for better localization and context and generates the output using multi-resolution deconvolution filters. The extra convolutional layers which bypass the max-pooling operation allow the network to train for variable input intensities and object size and make it robust to noisy data. We compare our results on publicly available data sets and show that the proposed network outperforms recent deep learning algorithms. [ABSTRACT FROM AUTHOR]

The amount of wheat ears per unit area is an important parameter for assessing wheat yield and wheat planting density. Generally, phenotype parameters should be obtained by manual count technique which is time-consuming and needs great effort. Aiming at the traditional image segmentation method based on color feature, texture feature and Haar feature, in this paper, we proposed an improved K-means algorithm for estimating the number of wheat ears. This method established a direct mapping relationship from the low-level features of the image to the number of wheat ears in the image through the color feature so that the target did not need to be segmented or detected. This type of method was more suitable for complex lighting and dense wheat ears, and had higher computational efficiency than counting methods based on image segmentation. This method made full use of the color feature of wheat ear image, took the area feature of local region extracted from local region as the basis of wheat ear judgment, and used the number of sub-regions in clustering region as the estimation value of wheat ear number, thus avoiding the task of target detection and location, and greatly improving the accuracy of Wheat ear counting. According to classification results of K-means method, the wheat ear image was divided into three regions. The green area pixel value representing the wheat eat region was set to 255, the pixel value of other areas was set to 0, so the binary image of wheat ears was obtained. Most of the binarized areas in the image were wheat ears, and a small part of the binarized area that was too small or too large was caused by the leaves. Following the steps were used to count wheat ears: 1) extracting the connected regions of the binary image and labeling them; 2) calculating the area of each connected region, the area was represented by the number of pixels in the region; 3) using the following method to filter out the connected region where the area was too small or too large; 4) finally, the number of wheat ears was obtained by counting the retained binary regions. This method established a direct classification relation from the image color feature of the lower layer to the image containing wheat ears so that the image did not need to be segmented or detected. Based on color feature clustering, this method could estimate the number of wheat ears in local area of space, and it was more scalable without training. The average prediction precision of the total wheatear number in a wheatear image for 12 wheatear images was 94.69%. And the Statistical error of the total wheatear number was between 2.17% and 7.41%, the average statistical error was 5.31%. The statistical experiment results showed that the accuracy of this algorithm to wheat ear counting was better than the traditional method based on image color feature and texture feature. [ABSTRACT FROM AUTHOR]

With the increase of the camera resolution, the number of pixels contained in froth image is increased, which brings many challenges to image segmentation. Froth size and distribution are the important index in froth flotation. The segmentation of froth images is always a problem in building flotation model. In segmenting froth images, Otsu method is usually used to get a binary image for classification of froth images, this method can get a satisfactory segmentation result. However, each gray level is required to calculate each of the between-class variance, it takes a longer time in froth images with a large number of pixels. To solve this problem, an improved method is proposed in this paper. Most froth images have the pixel distribution characteristic that the gray histogram curve is a sawtooth shape. The proposed method uses polynomial to fit the curve of gray histogram and takes the characteristic of gray histogram's valley into consideration in Otsu method. Two performance comparison methods are introduced and used. Experimental comparison between Otsu method and the proposed method shows that the proposed method has a satisfactory image segmentation with a low computing time. [ABSTRACT FROM AUTHOR]

This paper proposed segmentation method optimized by ABC algorithm in ASCS algorithm for the problem of ASCS algorithm, which adopted uniform segmentation instead of optimal segmentation. The ASCS algorithm only considered an influencing factor and ignored the influence of other factors. The ratio of compressing long sequence of uneven distribution was unsatisfactory with ASCS algorithm, it adopted the process integer sequence with sorting and differential encoding before ASCS algorithm. Simulation experiments show that the improved algorithm has significantly increased compression ratio of inverted index file comparing with ASCS algorithm. [ABSTRACT FROM AUTHOR]

Highlights • A bit-plane-dependence probability model is developed. • Statistical analysis of the bit-plane-dependence probability model is performed. • A fuzzy bit-plane-dependence segmentation algorithm is proposed. Abstract This paper presents a novel fuzzy bit-plane-dependence image segmentation methodology. We propose a probability model for characterizing the distributions of image variations based on bit-plane probabilities and dependencies between bit-planes. Compared with the current state-of-the-art image variation models which assume the distributions have specific structures (e.g., symmetry, monotone and periodicity), the proposed model provides a universal parametric representation that can be used to model random distributions without enforcing any specific restrictions on the distributions. In addition, we show that the maximum likelihood estimators of model parameters are joint sufficient statistics, which, in turn, justify the theoretical basis for their use. To effectively segment images with various textures, we propose a fuzzy bit-plane-dependence image segmentation algorithm. The proposed algorithm integrates the bit-plane-dependence probability model into the agglomerative fuzzy algorithm, and incorporates neighboring information and boundary correction for image segmentation applications. Experiments demonstrate the superior performance of the proposed method. [ABSTRACT FROM AUTHOR]

Among traditional Light Detection And Ranging (LIDAR) data representations such as raster grid, triangulated irregular network, point clouds and octree, the explicit 3D nature of voxel-based representation makes it a promising alternative. Despite the benefit of voxel-based representation, voxel-based algorithms have rarely been used for building detection. In this paper, a voxel segmentation-based 3D building detection algorithm is developed for separating building and nonbuilding voxels. The proposed algorithm first voxelizes the LIDAR point cloud into a grayscale voxel structure in which the grayscale of the voxel corresponds to the quantized mean intensity of the LIDAR points within the voxel. The voxelized dataset is segmented into multiple 3D-connected regions depending on the connectivity and grayscale similarity among voxels. The 3D-connected regions corresponding to the building roof and facade are detected sequentially according to characteristics such as their area, density, elevation difference and location. The obtained results for the detected buildings are evaluated by the LIDAR data provided by working group III/4 of ISPRS, which demonstrate a high rate of success. Average completeness, correctness, quality, and kappa coefficient indexes values of 90.0%, 96.0%, 88.1% and 88.7%, respectively, are obtained for buildings. [ABSTRACT FROM AUTHOR]

Milano Retinex is a family of spatial color algorithms inspired by Retinex and mainly devoted to the image enhancement. In the so-called point-based sampling Milano Retinex algorithms, this task is accomplished by processing the color of each image pixel based on a set of colors sampled in its surround. This paper presents STAR, a segmentation based approximation of the point-based sampling Milano Retinex approaches: it replaces the pixel-wise image sampling by a novel, computationally efficient procedure that detects once for all the color and spatial information relevant to image enhancement from clusters of pixels output by a segmentation. The experiments reported here show that STAR performs similarly to previous point-based sampling Milano Retinex approaches and that STAR enhancement improves the accuracy of the well-known algorithm scale-invariant feature transform on the description and matching of photographs captured under difficult light conditions. [ABSTRACT FROM AUTHOR]

Retrieving information from fish hatcheries is a key need in the Colombian fishing industry because it allows hatchery managers to determine necessary food amounts and measure other population parameters. However, traditional measurement methods involve extracting live fish from the ponds. This results in stress and the possibility of injury. Researchers have proposed automated measuring systems for shortening measurement times and reducing fish stress, but they must fulfill several prerequisites before they can retrieve fish information. These include mounting underwater camera systems and applying image enhancement and segmentation algorithms. In this paper, the literature revolving around these issues is reviewed and a novel approach is proposed. It is shown that using a single camera for image acquisition in a controlled setup is appropriate because it enables better management of sample size and image acquisition conditions. Furthermore, a combination of homomorphic filtering, contrast limited adaptive histogram equalization (CLAHE) and guided filtering for fish image enhancement was used. The fish were then segmented using a combination of 2D saliency detection and morphological operators. Finally, fish length was obtained using a third-degree polynomial regression on the fish mid-points. The length was calculated to estimate the weight with several regression algorithms. This approach was shown to be the most appropriate method for regression of fish weight based on length. [ABSTRACT FROM AUTHOR]

Highlights • we use background weight map to provide seeds for manifold ranking. • we extend the traditional manifold ranking to second-order formula. • we establish a third-order smoothness framework to optimize the saliency map. Abstract Graph-based approaches for saliency detection have attracted much attention and been exploited widely in recent years. In this paper, we present a new method to promote the performance of existing manifold ranking algorithms. Initially, we use background weight map to provide seeds for manifold ranking; Next, we extend the traditional manifold ranking to second-order formula and add a weight mask to its fitting term. Finally, for further improvement of the performance, we establish a third-order smoothness framework to optimize the saliency map. In the experiments, we compare two versions (manifold ranking with and without optimization) of our model with seven previous methods and test them on several benchmark datasets. Different kinds of strategies are also adopted for evaluation and the results demonstrate that our method achieves the state-of-the-art. [ABSTRACT FROM AUTHOR]

Highlights • The samples balanced lung nodule segmentation dataset based on CT slice image with labels was rebuilt. • Residual network is added to U-NET network, which resembles an ensemble approach to improve our final predictions. • Add the Batch Normalization layer into U-NET network, which can speed up training, prevent over-fitting. Abstract Since pulmonary nodules in CT images are very small and easily confusing with other tissues, there are still many problems in the pulmonary nodule segmentation. This paper presents an improved lung nodule segmentation algorithm based on U-NET network. Firstly, CT images are transformed and normalized, and the lung parenchyma is obtained by simple and efficient morphological method. Then, the U-NET network is improved, which mainly includes the dataset rebuilding, convolutional layer, pooling layer and upsampled layer. And we introduced residual network, which has improved the network training effect. Besides, we designed batch standardization operation, which has speeded up the network training and improves the network stability. Finally, we used the new dataset to train and test the improved U-NET network. A large number of experiments show that the proposed method can effectively improve the segmentation accuracy of pulmonary nodules. It is a great work with theoretical and practical value. [ABSTRACT FROM AUTHOR]

Highlights • A method of extracting key frames from surveillance videos was put forward. • A foreground segmentation method to detect the foreground of edge areas was proposed. • Combined with spatio-temporal continuity, a new region proposal method was proposed. • The rate of accuracy and efficiency of the method was verified by experiments. Abstract As one of the most famous algorithms, Faster R-CNN has been applied to many fields. However, it is unlikely to be used to surveillance videos directly because of its low efficiency and precision. To deal with these problems, this paper puts forward an object detection method which is based on Faster CNN to detect dairy goats in the surveillance video. It includes key frames extraction, foreground segmentation, region proposal and Fast R-CNN. The experimental results show that our method is more than twice as fast as Faster R-CNN and obtains 92.49% average precision. Our results suggest that our key frame extraction and region proposal method are helpful for detecting dairy goats. [ABSTRACT FROM AUTHOR]

Visual tracking is an essential technique in computer vision. Even though the notable improvement has been achieved during last few years, tracking multiple objects still remains as a challenging task. In this paper, a novel method for tracking multiple players in soccer videos, which include severe occlusions between players and nonlinear motions by their complex interactions, is introduced. Specifically, we first extract moving objects (i.e., players) by refining results of background subtraction via the edge information obtained from the frame differencing result. Then, we conduct multiscale sampling in foreground regions, which are spatially close to each tracked player, and subsequently computing the dissimilarity between sampled image blocks and each tracked player. Based on the best-matched case, the state of each tracked player (e.g., center position, color, etc.) is consistently updated using the online interpolation scheme. Experimental results in various soccer videos show the efficiency and robustness of our method compared to previous approaches introduced in the literature. [ABSTRACT FROM AUTHOR]

Object-based hyperspectral image classification (OBHIC) converts the basic unit from ‘pixel’ to ‘object’ by image segmentation, in order to take advantage of the spatial distribution law of geographical substances, as well as increase classification performances. However, it involves the problem of scale selection, i.e. the segmentation parameters are set manually by empirical values. In this paper, a novel OBHIC algorithm based on adaptive segmentation is proposed. Here, hyperspectral images (HSIs) are first segmented through a new segmentation scheme with constraint ability, and the thresholds for segmentation are calculated adaptively by utilizing training samples. And then K-nearest neighbor algorithm (KNN) is applied to classify the centers of each region after segmentation. In addition, based on the semisupervised idea, semi-known samples are obtained to further improve the classification performance. Experimental results are presented on two HSI datasets. For different HSIs, the adaptive thresholds calculated are consistent with empirical ones, and the developed classification algorithm has achieved good classification results, thus demonstrating strong robustness of the algorithm. For the HSI Indian Pines from AVIRIS sensor, the Overall Accuracy (OA) and kappa are 95.13% and 0.9444 respectively with 10% training samples, and for the HSI Pavia University from ROSIS sensor, the OA and kappa are 95.52% and 0.9416 respectively with 2% training samples. And good classification performance is still maintained for small number of training samples. [ABSTRACT FROM AUTHOR]

The automatic reconstruction of single neurons from microscopic images is essential to enable large-scale data-driven investigations in neuron morphology research. However, few previous methods were able to generate satisfactory results automatically from 3-D microscopic images without human intervention. In this paper, we developed a new algorithm for automatic 3-D neuron reconstruction. The main idea of the proposed algorithm is to iteratively track backward from the potential neuronal termini to the soma centre. An online confidence score is computed to decide if a tracing iteration should be stopped and discarded from the final reconstruction. The performance improvements comparing with the previous methods are mainly introduced by a more accurate estimation of the traced area and the confidence controlled back-tracking algorithm. The proposed algorithm supports large-scale batch-processing by requiring only one user specified parameter for background segmentation. We bench tested the proposed algorithm on the images obtained from both the DIADEM challenge and the BigNeuron challenge. Our proposed algorithm achieved the state-of-the-art results. [ABSTRACT FROM AUTHOR]

This paper presents a lightweight synthesis algorithm, named adaptive region segmentation based piecewise linear (ARSPL) algorithm, for reconstructing standard 12-lead electrocardiogram (ECG) signals from a 3-lead subset (I, II and V2). Such a lightweight algorithm is particularly suitable for healthcare mobile devices with limited resources for computing, communication and data storage. After detection of R-peaks, the ECGs are segmented by cardiac cycles. Each cycle is further divided into four regions according to different cardiac electrical activity stages. A personalized linear regression algorithm is then applied to these regions respectively for improved ECG synthesis. The proposed ARSPL method has been tested on 39 subjects randomly selected from the PTB diagnostic ECG database and achieved accurate synthesis of remaining leads with an average correlation coefficient of 0.947, an average root-mean-square error of 55.4μV, and an average runtime performance of 114ms. Overall, these results are significantly better than those of the common linear regression method, the back propagation (BP) neural network and the BP optimized using the genetic algorithm. We have also used the reconstructed ECG signals to evaluate the denivelation of ST segment, which is a potential symptom of intrinsic myocardial disease. After ARSPL, only 10.71% of the synthesized ECG cycles are with a ST-level synthesis error larger than 0.1mV, which is also better than those of the three above-mentioned methods. [ABSTRACT FROM AUTHOR]

In this paper, we proposed a novel method for text line segmentation of Tibetan historical document image with uchen script based on contour tracking. Our method is mainly to segment the text lines from the image documents using the contour curve of the text lines, which consists of three parts: First, we calculate the barycentre coordinates of the connected components for the text regions, and then the barycentre of each text line is connected in order, so that the main part of each text line is connected and a new connected component is formed; then the contour curve of the connected component is obtained using the contour tracing algorithm; Second, the contour curve and the barycentre gravity are used to assign key elements (such as the syllable point, the upper vowel, the lower vowel, and the broken strokes and so on) of the text lines, and next the candidate text lines are obtained based on these connected components; Finally, the contour tracking algorithm is used to calculate the contour curve of the candidate text lines and segment the text lines. We evaluated our text line segmentation method on the 200 document image data sets. Experimental results show that the proposed method based on contour curve tracing can accurately segment the text lines of image documents and achieve the encouraging results. [ABSTRACT FROM AUTHOR]

Image segmentation is the most important operation in image processing that changes the representation of an image to be useful in much computer vision applications. During the last years, several image segmentation algorithms have been developed, consequently, a great number of techniques for evaluating segmentation results have been proposed. Unfortunately, the majority of these methods is subjective and cannot be used to judge the performance of different segmentation algorithms. In this paper, we propose a novel evaluation criterion based on the Gini index and the entropy calculation. This new method permits evaluating the conjunction of the regions arrangement in the segmented image, and measuring the pixels homogeneity within each region. In experiments, benchmark images segmented by the multilevel thresholding technique based on particle swarm optimization are used to conclude the strength, the effectiveness and the rapidity of the proposed evaluation criterion. [ABSTRACT FROM AUTHOR]

Edge detection is a fundamental step in image segmentation and one of the most important steps to recognize the features of objects. In this paper, we proposed an edge detection method using modified Moore-Neighbor algorithm. The Modified Moore-neighbor algorithm is used as boundary tracing for detection and feature extraction. Further, range filter is incorporated with modified Moore-Neighbor to detect the edges of objects. The proposed edge detection can extract the true edges as well as features of an object and is robust to noise. We have considered both objective evaluation and subjective evaluation to verify the results. The image analysis parameters are considered for subjective evaluation. Further, a total number of 24 BSD (Berkeley Segmentation Database) images are considered for subjective evaluation using the proposed method and compared with the other existing methods. The results obtained using the proposed modified Moore-Neighbor algorithm for different test images and BSD images are found to be promising. Therefore, proposed edge detection method warrants further investigation. [ABSTRACT FROM AUTHOR]

Semantic segmentation, also called scene labeling, refers to the process of assigning a semantic label (e.g. car, people, and road) to each pixel of an image. It is an essential data processing step for robots and other unmanned systems to understand the surrounding scene. Despite decades of efforts, semantic segmentation is still a very challenging task due to large variations in natural scenes. In this paper, we provide a systematic review of recent advances in this field. In particular, three categories of methods are reviewed and compared, including those based on hand-engineered features, learned features and weakly supervised learning. In addition, we describe a number of popular datasets aiming for facilitating the development of new segmentation algorithms. In order to demonstrate the advantages and disadvantages of different semantic segmentation models, we conduct a series of comparisons between them. Deep discussions about the comparisons are also provided. Finally, this review is concluded by discussing future directions and challenges in this important field of research. [ABSTRACT FROM AUTHOR]

Semantic segmentation tasks can be well modeled by Markov Random Field (MRF). This paper addresses semantic segmentation by incorporating high-order relations and mixture of label contexts into MRF. Unlike previous works that optimized MRFs using iterative algorithm, we solve MRF by proposing a Convolutional Neural Network (CNN), namely Deep Parsing Network (DPN), which enables deterministic end-to-end computation in a single forward pass. Specifically, DPN extends a contemporary CNN to model unary terms and additional layers are devised to approximate the mean field (MF) algorithm for pairwise terms. It has several appealing properties. First, different from the recent works that required many iterations of MF during back-propagation, DPN is able to achieve high performance by approximating one iteration of MF. Second, DPN represents various types of pairwise terms, making many existing models as its special cases. Furthermore, pairwise terms in DPN provide a unified framework to encode rich contextual information in high-dimensional data, such as images and videos. Third, DPN makes MF easier to be parallelized and speeded up, thus enabling efficient inference. DPN is thoroughly evaluated on standard semantic image/video segmentation benchmarks, where a single DPN model yields state-of-the-art segmentation accuracies on PASCAL VOC 2012, Cityscapes dataset and CamVid dataset. [ABSTRACT FROM AUTHOR]

With the emergence of affordable 3D scanning and printing devices, processing of large point clouds has to be performed in many applications. Several algorithms are available for surface reconstruction, smoothing, and parametrization. However, many of these require the sampling of the point cloud to be uniform or at least to be within certain controlled parameters. For non-uniformly sampled point clouds, some methods have been proposed that deal with the non-uniformity by adding additional information such as topological or hierarchical data. In this paper, we focus on point clouds sampling surfaces in R 3 . We present the notion of local directional density measure that can be intrinsically computed within the point cloud, that is without further knowledge of the geometry despite the given point samples. Specifically, we build on previous work to derive a local, directed, and discrete measure for density. Furthermore, we derive another discrete and a smooth density measure and compare these three experimentally. Each of the three considered measures gives density weights to use in discretizations of operators such that these become independent of sampling uniformity. We demonstrate the effectiveness of our method on both synthetic and real world data. [ABSTRACT FROM AUTHOR]

This paper presents a new calibration method for a 6 degree-of-freedom (DOF) parallel manipulator based on identifiable parameters separation (IPS). This method begins with the parameters separation by the principle of IPS, and then identifies the subsets of the parameters sequentially and respectively. The method aims to identify as many parameters as possible just by simple and direct measuring procedures, which can lower the number of parameters in the complex identification model and thus reduce the coupling effect and uncertainty during identification. Moreover, the reduction of the parameters will allow fewer measurements and faster convergence for the identification algorithm. Firstly, the manipulator is briefly described and its geometric errors are studied. Secondly, the concept of IPS is introduced and the principle of IPS is summarized. Then, according to the proposed method, the corresponding calibration experiment is designed. Lastly, the experiment of this manipulator is conducted and the result shows the validity of the method for accuracy improvement. The rules of IPS and the simplification of complex identification enable the use of this method on other types of parallel mechanisms. [ABSTRACT FROM AUTHOR]

Abstract In this paper, we present a new model for spatiotemporal saliency detection. Instead of previous works which combine the image saliency in the spatial domain with motion cues to build their video saliency model, we propose to apply the pattern mining (PM) algorithm. From initial saliency maps computed in spatial and temporal domains, discriminative spatiotemporal saliency patterns can be recognized and their label information is propagated to obtain the final saliency map. Our model ensures a good compromise between image saliency and motion saliency and presents an accurate prediction to estimate salient regions in comparison with other methods for video saliency detection. Finally, as an application of our algorithm, our spatiotemporal saliency map is combined with appearance models and dynamic location models into an energy minimization framework to segment salient moving object. Experiments show a good performance of our algorithm for moving object segmentation (MOS) on benchmark datasets. [ABSTRACT FROM AUTHOR]

Highlights • Two interaction mechanisms are allowed: fully-user-supervised (BnS) or machine-assisted (BnQ). • The manifold ranking is integrated with the query-based interactive mechanism. • The foreground-object-confidence update rule is efficient and effective. • The average response time per-interaction is quite short. • The initial bounding box can be imprecise. This paper presents an efficient and effective interactive segmentation scheme for extracting the region of a foreground object in an image. Our goal is to design an interactive segmentation algorithm that unifies the bounding-box-based, seed-based, and query-based interaction mechanisms for pursuing (i) high efficiency in simple interaction mechanism, (ii) few interaction rounds, and (iii) short response time. The proposed algorithm starts with a user-provided bounding box and obtains candidate background superpixels for inferring the foreground object. Our algorithm tolerates imprecise bounding boxes and provides two kinds of interactions for acquiring correct labels from the user. The user can either input the seed/scribble annotations or label the algorithm-queried regions. Our algorithm selects the most uncertain region as a query, and this query-based interaction mechanism reduces the burden of the user on deciding suitable annotation locations. The average response time per-interaction of our algorithm is merely 0.014 s. Our experiments demonstrate that the algorithm achieves an efficient unified scheme for interactive image segmentation. [ABSTRACT FROM AUTHOR]

Graphical abstract Highlights • A feature visualization for CNNs method is exploited to improve co-segmentation. • Visualization is used as an auxiliary information to discriminate salient regions. • The visualization information is used as an extra energy term in the cost function. • Region occlusion sensitivity is introduced for feature visualization. • An adaptive strategy is designed to detect correct foreground/background regions. This paper addresses the co-segmentation problem using feature visualization for CNNs. Visualization is exploited as an auxiliary information to discriminate salient image regions (dubbed as “heat-regions”) from non-salient ones. Region occlusion sensitivity is proposed for feature visualization. The co-segmentation problem is formulated via a convex quadratic optimization which is initialized by the heat-regions. The information obtained through the visualization is considered as an extra energy term in the cost function. The results of the visualization demonstrate that there exist some heat-regions which are not productive in the co-segmentation. To detect helpful regions among them, an adaptive strategy in the form of an iterative algorithm is proposed according to the consistency among all images. Comparison experiments conducted on two benchmark datasets, iCoseg and MSRC, illustrate the superior performance of the proposed approach over state-of-the-art algorithms. [ABSTRACT FROM AUTHOR]

In this paper, we identify and solve a multi-join optimization problem for Arbitrary Feature-based social image Similarity JOINs(AFS-JOIN). Given two collections(i.e., R and S) of social images that carry both visual, spatial and textual(i.e., tag) information, the multiple joins based on arbitrary features retrieves the pairs of images that are visually, textually similar or spatially close from different users. To address this problem, in this paper, we have proposed three methods to facilitate the multi-join processing: 1) two baseline approaches(i.e., a naïve join approach and a maximal threshold(MT)-based), and 2) a Batch Similarity Join( BSJ) method. For the BSJ method, given m users' join requests, they are first conversed and grouped into m″ clusters which correspond to m″ join boxes, where m > m″. To speedup the BSJ processing, a feature distance space is first partitioned into some cubes based on four segmentation schemes; the image pairs falling in the cubes are indexed by the cube tree index; thus BSJ processing is transformed into the searching of the image pairs falling in some affected cubes for m″ AFS-JOINs with the aid of the index. An extensive experimental evaluation using real and synthetic datasets shows that our proposed BSJ technique outperforms the state-of-the-art solutions. [ABSTRACT FROM AUTHOR]

Image segmentation is one of the fundamental steps in image analysis for object identification. The main goal of image segmentation is to recognize homogeneous regions within an image as distinct and belonging to different objects. Inspired by the idea of the packing problem, in this paper, we propose a fast O ( N α ( N ) ) -time algorithm for image segmentation by using Non-symmetry and Anti-packing Model and Extended Shading representation, which was called the NAMES-based algorithm, where N is the number of homogenous blocks and α ( N ) is the inverse of the Ackerman’s function and it is a very slowly growing function. We first put forward four extended Lemmas and two extended Theorems. Then, we present a new scanning method used to process each NAMES block. Finally, we propose a novel NAMES-based data structure used to merge two regions. With the same experimental conditions and the same time complexity, our proposed NAMES-based algorithm, which extends the popular hierarchical representation model to a new non-hierarchical representation model, has about 86.75% and 89.47% average execution time improvement ratio when compared to the Binary Partition Tree (BPT)-based algorithm and the Quadtree Shading (QS)-based algorithm which has about 55.4% execution time improvement ratio when the QS-based algorithm itself is compared to the previous fastest region segmentation algorithm by Fiorio and Gustedt whose O ( N 2 ) -time algorithm is run on the original N × N gray image. Further, the NAMES can improve the memory-saving by 28.85% (5.04%) and simultaneously reduce the number of the homogeneous blocks by 49.05% (36.04%) more than the QS (the BPT) whereas maintaining the satisfactory image quality. Therefore, by comparing our NAMES-based algorithm with the QS-based algorithm and the BPT-based algorithm, the experimental results presented in this paper show that the former has not only higher compression ratio and less number of homogenous blocks than the latter whereas maintaining the satisfactory image quality, but also can significantly improve the execution speed for image segmentation, and therefore it is a much more effective algorithm for image segmentation. [ABSTRACT FROM AUTHOR]

Deciding a rational number of clusters for the problems to be solved and determining the parameters associated with the clustering algorithms are two critical issues in the configuration of data clustering. Usually, a manual trial-and-error manner is used to induce a feasible configuration. This paper presents an automated clustering method, which determines the clustering configuration automatically. The proposed method is based on the techniques of support-vector machines with a local-search strategy. It starts with running one-class support-vector machines (OCSVM) to partition input data into a random number of clusters. When a result is obtained, the “local-search” mechanism launches several rounds of OCSVM each of which works with a new clustering configuration. Each new configuration is from the current configuration with incrementally modifications. The clustering results obtained from the local searches are post-evaluated by specific clustering validity index and the best one is retained. The clustering configuration of the best result is used by OCSVM for the clustering afterwards. Such a clustering process iterates until no better result can be obtained. This paper describes the clustering algorithm and compares three clustering validity indices, i.e. distance-based index, Davies–Bouldin index, and Xie–Beni index, on their effectiveness. The performance of the proposed method is demonstrated on the segmentation of several aerial images. Experimental results show that the proposed approach is feasible and effective for image segmentation. [ABSTRACT FROM AUTHOR]

Pattern recognition is a key issue in the field of textile computer aided analysis and design. Focusing on the non-stationarity of the fabric texture structure, a hybrid method combining two-dimensional mode decomposition with mean shift segmentation is proposed in this paper. Two-dimensional empirical mode decomposition is an adaptive image analysing technique, which uses intrinsic spatial scales instead of predefined basis functions to decompose a given image into a number of intrinsic modes. The first intrinsic mode preserves multiple patterns for an underlying fabric image, which is subsequently submitted to the mean shift clustering process in this paper. Simulation results have proved that the proposed mean shift algorithm in empirical mode domain is in accordance with human visual effect, and possesses the capability of clustering diverse patterns inherent in the textile in a fully unsupervised way. [ABSTRACT FROM AUTHOR]

Background: As a new authentication method developed years ago, vein recognition technology features the unique advantage of bioassay.Objective: This paper studies the specific procedure for the extraction of hand back vein characteristics.Methods: There are different positions used in the collecting process, so that a suitable intravenous regional orientation method is put forward, allowing the positioning area to be the same for all hand positions. In addition, to eliminate the pseudo vein area, the valley regional shape extraction operator can be improved and combined with multiple segmentation algorithms. The images should be segmented step by step, making the vein texture to appear clear and accurate. Lastly, the segmented images should be filtered, eroded, and refined. This process helps to filter the most of the pseudo vein information.Results: Finally, a clear vein skeleton diagram is obtained, demonstrating the effectiveness of the algorithm.Conclusion: This paper presents a hand back vein region location method. This makes it possible to rotate and correct the image by working out the inclination degree of contour at the side of hand back. [ABSTRACT FROM AUTHOR]

This paper addresses the problem of semantic segmentation, where the possible class labels are from a predefined set. We exploit top-down guidance, i.e., the coarse localization of the objects and their class labels provided by object detectors. For each detected bounding box, figure–ground segmentation is performed and the final result is achieved by merging the figure–ground segmentations. The main idea of the proposed approach, which is presented in our preliminary work, is to reformulate the figure–ground segmentation problem as sparse reconstruction pursuing the object mask in a nonparametric manner. The latent segmentation mask should be coherent subject to sparse error caused by intra-category diversity; thus, the object mask is inferred by making use of sparse representations over the training set. To handle local spatial deformations, local patch-level masks are also considered and inferred by sparse representations over the spatially nearby patches. The sparse reconstruction coefficients and the latent mask are alternately optimized by applying the Lasso algorithm and the accelerated proximal gradient method. The proposed formulation results in a convex optimization problem; thus, the global optimal solution is achieved. In this paper, we provide theoretical analysis of the convergence and optimality. We also give an extended numerical analysis of the proposed algorithm and a comprehensive comparison with the related semantic segmentation methods on the challenging PASCAL visual object class object segmentation datasets and the Weizmann horse dataset. The experimental results demonstrate that the proposed algorithm achieves a competitive performance when compared with the state of the arts. [ABSTRACT FROM PUBLISHER]

Multi-level image thresholding is a common approach to image segmentation where an image is divided into several regions based on its histogram. Otsu's method is the most popular method for this purpose, and is based on seeking for threshold values that maximise the between-class variance. This requires an exhaustive search to find the optimal set of threshold values, making image thresholding a time-consuming process. This is especially the case with increasing numbers of thresholds since, due to the curse of dimensionality, the search space enlarges exponentially with the number of thresholds. Population-based metaheuristic algorithms are efficient and effective problem-independent methods to tackle hard optimisation problems. Over the years, a variety of such algorithms, often based on bio-inspired paradigms, have been proposed. In this paper, we formulate multi-level image thresholding as an optimisation problem and perform an extensive evaluation of 23 population-based metaheuristics, including both state-of-the-art and recently introduced algorithms, for this purpose. We benchmark the algorithms on a set of commonly used images and based on various measures, including objective function value, peak signal-to-noise ratio, feature similarity index, and structural similarity index. In addition, we carry out a stability analysis as well as a statistical analysis to judge if there are significant differences between algorithms. Our experimental results indicate that recently introduced algorithms do not necessarily achieve acceptable performance in multi-level image thresholding, while some established algorithms are demonstrated to work better. [ABSTRACT FROM AUTHOR]

Image processing methods for automated concrete crack detection are often challenged by binary noise. Noise removal methods decrease the false positive pixels of crack detection results, often at the cost of a reduction in true positives. This paper proposes a novel method for binary noise removal and segmentation of noisy concrete crack images. The method applies an area threshold before reducing the pixel groups in the image to a skeleton. Each skeleton is connected to its nearest neighbour before the remaining short skeletons in the image are removed using a length threshold. A morphological reconstruction follows to remove all elements in the original noisy image that do not intersect with the skeleton. Finally, pixel groups in close proximity to the endpoints of the pixel groups in the resulting image are reinstated. Testing was conducted on a dataset of noisy binary crack images; the proposed method (Skele-Marker) obtained recall, precision, intersection over union, and F1 score results of 77%, 91%, 72%, and 84%, respectively. Skele-Marker was compared to other methods found in literature and was found to outperform other methods in terms of precision, intersection over union and F1 score. The proposed method is used to make crack detection results more reliable, supporting the ever-growing demand for automated inspections of concrete structures. • We use image processing techniques to remove noise in binary concrete crack images. • False positives in binary concrete crack images can be reduced using our method. • We compare our concrete crack noise removal method to others found in literature. • Our method performed better than other concrete crack segmentation methods. [ABSTRACT FROM AUTHOR]

For harvesting robot, fruit identification is the key step for accurate fruit positioning and successful picking. The primary task of fruit identification and picking is to separate fruit from complicated background of branches, trunk and sky by image segmentation. It is hard to accurately segment colorized bayberry image because there are fruits with low brightness or uneven illumination in nature scenes. In this study, RGB (red, green and blue) color space was transformed into HSV (hue, saturation and value) space. After that, the luminance component of image was strengthened by dynamic Butterworth homomorphism filter transfer function. Then, it was restored to RGB color space for colorized image illumination compensation and shadow removal. The bayberry image after shadow removal included red bayberry, green leave and white sky. Each pixel of colorized bayberry image to be segmented was considered as one point of data set X. These pixels were classified into red, green and white. According to the characteristics of the components a and b in Lab color space, RGB color space was transformed into CIELAB space. The K-means clustering algorithm was used for image segmentation, and the parameter K was selected as 3. In order to verify the effectiveness of the proposed algorithm, 15 bayberry images were selected from 100 images affected by different degrees of shadow under different growth conditions and uneven illumination conditions. Firstly, in order to prove effectiveness of illumination compensation, the K-means clustering algorithm was used to conduct image segmentation experiments before and after illumination compensation to shadow removal. Secondly, in order to validate segmentation effectiveness of images after illumination compensation based on different methods, this study applied adaptive 2*R-G-B grey threshold and K-means clustering segmentation algorithms to compare their effects of shadow removal. Thirdly, homomorphism filter algorithm was compared with linear enhancement and histogram equalization methods, and the K-means algorithm was applied to analyze image segmentation effectiveness based on different strengthen methods. The experiments showed that the segmentation result based on K-means clustering algorithm was without wrong segmentation after illumination compensation for shadow removal compared with that before illumination compensation. Although grey threshold based on color difference 2*R-G-B had better image segmentation effect after illumination compensation, some samples had large wrong segmentation and bright leaves were segmented and classified into bayberry. Therefore, image segmentation by grey threshold based on color difference 2*R-G-B was worse than that segmentation algorithm based on K-means clustering. Three criteria such as segmentation error, false positive rate (FPR) and false negative rate (FNR) were used to evaluate the segmentations results as quantitative analysis. Under the proposed method in this paper, the average segmentation error, FPR and FNR were 3.78%, 0.69% and 6.8%, respectively. Compared with the gray scale transform method, the segmentation error was reduced by 32.94 percent point, FPR by 6.85 percent point and FNR by 29.65 percent point for this proposed method. Then the average segmentation error was reduced by 24.92 percent point compared with the result obtained by histogram equalization method, FPR by 6.12 percent point and FNR by 20.40 percent point. All these results show that image illumination compensation by homomorphism filter algorithm presents better effect of shadow removal. K-means clustering segmentation algorithm has better image segmentation effect after shadow removal. This paper provides reference for the research on bayberry image segmentation and bayberry fruit recognition. [ABSTRACT FROM AUTHOR]