Search

The present paper aims to conduct an experiment that compares different methods of detecting objects in images. Programs were developed to evaluate the efficiency of SURF, BRISK, MSER, and ORB object detection methods. Four static gray images with sufficiently different histograms were used. The experiment also highlighted the need for image preprocessing to improve feature extraction and detection. Thus, a programmed method for adjusting pixel groups was developed. This method proved useful when one of the listed algorithms failed to detect the object in the original image, but succeeded after adjustment. The effectiveness of detection methods and the evaluation of their performance depend on the application, image preparation, algorithms used, and their implementation. Results of the detection methods were presented numerically (similarities, gradients, distances, etc.) and graphically.

The present paper aims to conduct an experiment that compares different methods of detecting objects in images. Programs were developed to evaluate the efficiency of SURF, BRISK, MSER, and ORB object detection methods. Four static gray images with sufficiently different histograms were used. The experiment also highlighted the need for image preprocessing to improve feature extraction and detection. Thus, a programmed method for adjusting pixel groups was developed. This method proved useful when one of the listed algorithms failed to detect the object in the original image, but succeeded after adjustment. The effectiveness of detection methods and the evaluation of their performance depend on the application, image preparation, algorithms used, and their implementation. Results of the detection methods were presented numerically (similarities, gradients, distances, etc.) and graphically. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Computer-Aided Design & Computer Graphics / Jisuanji Fuzhu Sheji Yu Tuxingxue Xuebao is the property of Gai Kan Bian Wei Hui and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The problem of data enrichment to train visual SLAM and VO construction models using deep learning (DL) is an urgent problem today in computer vision. DL requires a large amount of data to train a model, and more data with many different contextual and conditional conditions will create a more accurate visual SLAM and VO construction model. In this paper, we introduce the TQU-SLAM benchmark dataset, which includes 160,631 RGB-D frame pairs. It was collected from the corridors of three interconnected buildings comprising a length of about 230 m. The ground-truth data of the TQU-SLAM benchmark dataset were prepared manually, including 6-DOF camera poses, 3D point cloud data, intrinsic parameters, and the transformation matrix between the camera coordinate system and the real world. We also tested the TQU-SLAM benchmark dataset using the PySLAM framework with traditional features such as SHI_TOMASI, SIFT, SURF, ORB, ORB2, AKAZE, KAZE, and BRISK and features extracted from DL such as VGG, DPVO, and TartanVO. The camera pose estimation results are evaluated, and we show that the ORB2 features have the best results ( E r r d = 5.74 mm), while the ratio of the number of frames with detected keypoints of the SHI_TOMASI feature is the best ( r d = 98.97 % ). At the same time, we also present and analyze the challenges of the TQU-SLAM benchmark dataset for building visual SLAM and VO systems. [ABSTRACT FROM AUTHOR]

Facial expressions can convey the internal emotions of a person within a certain scenario and play a major role in the social interaction of human beings. In automatic Facial Expression Recognition (FER) systems, the method applied for feature extraction plays a major role in determining the performance of a system. In this regard, by drawing inspiration from the Swastik symbol, three texture based feature descriptors named Symbol Patterns (SP1, SP2 and SP3) have been proposed for facial feature extraction. SP1 generates one pattern value by comparing eight pixels within a 3 × 3 neighborhood, whereas, SP2 and SP3 generates two pattern values each by comparing twelve and sixteen pixels within a 5 × 5 neighborhood respectively. In this work, the proposed Symbol Patterns (SP) have been evaluated with natural, fibonacci, odd, prime, squares and binary weights for determining the optimal recognition accuracy. The proposed SP methods have been tested on MUG, TFEID, CK+, KDEF, FER2013 and FERG datasets and the results from the experimental analysis demonstrated an improvement in the recognition accuracy when compared to the existing FER methods. [ABSTRACT FROM AUTHOR]

The use of Synthetic Aperture Sonar (SAS) in autonomous underwater vehicle (AUV) surveys has found applications in archaeological searches, underwater mine detection and wildlife monitoring. However, the easy confusability of natural objects with the target object leads to high false positive rates. To improve detection, the combination of SAS and optical images has recently attracted attention. While SAS data provides a large-scale survey, optical information can help contextualize it. This combination creates the need to match multimodal, optical–acoustic image pairs. The two images are not aligned, and are taken from different angles of view and at different times. As a result, challenges such as the different resolution, scaling and posture of the two sensors need to be overcome. In this research, motivated by the information gain when using both modalities, we turn to statistical exploration for feature analysis to investigate the relationship between the two modalities. In particular, we propose an entropic method for recognizing matching multimodal images of the same object and investigate the probabilistic dependency between the images of the two modalities based on their conditional probabilities. The results on a real dataset of SAS and optical images of the same and different objects on the seafloor confirm our assumption that the conditional probability of SAS images is different from the marginal probability given an optical image, and show a favorable trade-off between detection and false alarm rate that is higher than current benchmarks. For reproducibility, we share our database. [ABSTRACT FROM AUTHOR]

Coarse registration of point cloud is a necessary step for object digitization. However, insufficient overlapping, large pose difference and the existence of noise and outliers seriously reduce the result. In this paper, several improvements were made to improve the registration effect under the above conditions. Firstly, a lightweight network for feature point detection based on local extremum is proposed to improve the repeatability and robustness of feature detection; Secondly, a feature description network combined with attention mechanism is constructed to generate highly differentiated descriptors for the feature points; Finally, a transformation parameters calculation strategy based on only two feature points is proposed, which improves the success probability under low overlapping. Experiments show that our feature detection, description and registration methods achieved satisfactory results in various challenging scenes and perform better than current mainstream methods. [ABSTRACT FROM AUTHOR]

Accurately classifying petrol and diesel fuel using an image processing method is crucial for fuel-related industries such as petrol pumps, refineries, and fuel storage facilities. However, distinguishing between these fuels using traditional methods can be challenging due to their similar visual characteristics. This article introduces a novel multi-scale and multi-level modified histogram of oriented gradients (MHOG) feature descriptors for robust classification of fuel images. Our proposed method involves extracting distinctive features from the images using the novel multi-scale and multi-level MHOG feature descriptor. These features are then utilized to train a range of machine learning classifiers with different hyperparameter settings for an ablation study. To the best of our knowledge, this is the first ablation study for this fuel classification application. To evaluate the effectiveness of our approach, we conduct experiments on a carefully labeled dataset consisting of petrol and diesel fuel images. The results demonstrate the high accuracy of our proposed method, achieving a classification accuracy of 98% using the light gradient boosting machine (LGBM). Furthermore, our method surpasses existing state-of-the-art techniques for fuel image classification. With its superior performance, this approach holds great potential for efficient and effective fuel classification in diverse fuel-related industries.

This research seeks to assess the effect of different selected feature descriptors on the accuracy of an automatic image registration scheme. Three different feature descriptors were selected based on their peculiar characteristics, and implemented in the process of developing the image registration scheme. These feature descriptors (Modified Harris and Stephens corner detector (MHCD), the Scale Invariant Feature Transform (SIFT) and the Speeded Up Robust Feature (SURF)) were used to automatically extract the conjugate points common to the overlapping image pairs used for the registration. Random Sampling Consensus (RANSAC) algorithm was used to exclude outliers and to fit the matched correspondences, Sum of Absolute Differences (SAD) which is a correlation-based feature matching metric was used for the feature match, while projective transformation function was used for the computation of the transformation matrix (T). The obtained overall result proved that the SURF algorithm outperforms the other two feature descriptors with an accuracy measure of -0.0009 pixels, while SIFT with a cumulative signed distance of 0.0328 pixels also proved to be more accurate than MHCD with a cumulative signed distance of 0.0457 pixels. The findings affirmed the importance of choosing the right feature descriptor in the overall accuracy of an automatic image registration scheme. [ABSTRACT FROM AUTHOR]

This research seeks to assess the effect of different selected feature descriptors on the accuracy of an automatic image registration scheme. Three different feature descriptors were selected based on their peculiar characteristics, and implemented in the process of developing the image registration scheme. These feature descriptors (Modified Harris and Stephens corner detector (MHCD), the Scale Invariant Feature Transform (SIFT) and the Speeded Up Robust Feature (SURF)) were used to automatically extract the conjugate points common to the overlapping image pairs used for the registration. Random Sampling Consensus (RANSAC) algorithm was used to exclude outliers and to fit the matched correspondences, Sum of Absolute Differences (SAD) which is a correlation-based feature matching metric was used for the feature match, while projective transformation function was used for the computation of the transformation matrix (T). The obtained overall result proved that the SURF algorithm outperforms the other two feature descriptors with an accuracy measure of -0.0009 pixels, while SIFT with a cumulative signed distance of 0.0328 pixels also proved to be more accurate than MHCD with a cumulative signed distance of 0.0457 pixels. The findings affirmed the importance of choosing the right feature descriptor in the overall accuracy of an automatic image registration scheme.

The problem of data enrichment to train visual SLAM and VO construction models using deep learning (DL) is an urgent problem today in computer vision. DL requires a large amount of data to train a model, and more data with many different contextual and conditional conditions will create a more accurate visual SLAM and VO construction model. In this paper, we introduce the TQU-SLAM benchmark dataset, which includes 160,631 RGB-D frame pairs. It was collected from the corridors of three interconnected buildings comprising a length of about 230 m. The ground-truth data of the TQU-SLAM benchmark dataset were prepared manually, including 6-DOF camera poses, 3D point cloud data, intrinsic parameters, and the transformation matrix between the camera coordinate system and the real world. We also tested the TQU-SLAM benchmark dataset using the PySLAM framework with traditional features such as SHI_TOMASI, SIFT, SURF, ORB, ORB2, AKAZE, KAZE, and BRISK and features extracted from DL such as VGG, DPVO, and TartanVO. The camera pose estimation results are evaluated, and we show that the ORB2 features have the best results (Errd = 5.74 mm), while the ratio of the number of frames with detected keypoints of the SHI_TOMASI feature is the best (rd=98.97%). At the same time, we also present and analyze the challenges of the TQU-SLAM benchmark dataset for building visual SLAM and VO systems.

The use of Synthetic Aperture Sonar (SAS) in autonomous underwater vehicle (AUV) surveys has found applications in archaeological searches, underwater mine detection and wildlife monitoring. However, the easy confusability of natural objects with the target object leads to high false positive rates. To improve detection, the combination of SAS and optical images has recently attracted attention. While SAS data provides a large-scale survey, optical information can help contextualize it. This combination creates the need to match multimodal, optical–acoustic image pairs. The two images are not aligned, and are taken from different angles of view and at different times. As a result, challenges such as the different resolution, scaling and posture of the two sensors need to be overcome. In this research, motivated by the information gain when using both modalities, we turn to statistical exploration for feature analysis to investigate the relationship between the two modalities. In particular, we propose an entropic method for recognizing matching multimodal images of the same object and investigate the probabilistic dependency between the images of the two modalities based on their conditional probabilities. The results on a real dataset of SAS and optical images of the same and different objects on the seafloor confirm our assumption that the conditional probability of SAS images is different from the marginal probability given an optical image, and show a favorable trade-off between detection and false alarm rate that is higher than current benchmarks. For reproducibility, we share our database.

In this article, we propose a hybrid framework for cross-resolution 3D face recognition which utilizes a Streamed Attention Network (SAN) that combines handcrafted features with Convolutional Neural Networks (CNNs). It consists of two main stages: first, we process the depth images to extract low-level surface descriptors and derive the corresponding Descriptor Images (DIs), represented as four-channel images. To build the DIs, we propose a variation of the 3D Local Binary Pattern (3DLBP) operator that encodes depth differences using a sigmoid function. Then, we design a CNN that learns from these DIs. The peculiarity of our solution consists in processing each channel of the input image separately, and fusing the contribution of each channel by means of both self- and cross-attention mechanisms. This strategy showed two main advantages over the direct application of Deep-CNN to depth images of the face; on the one hand, the DIs can reduce the diversity between high- and low-resolution data by encoding surface properties that are robust to resolution differences. On the other, it allows a better exploitation of the richer information provided by low-level features, resulting in improved recognition. We evaluated the proposed architecture in a challenging cross-dataset, cross-resolution scenario. To this aim, we first train the network on scanner-resolution 3D data. Next, we utilize the pre-trained network as feature extractor on low-resolution data, where the output of the last fully connected layer is used as face descriptor. Other than standard benchmarks, we also perform experiments on a newly collected dataset of paired high- and low-resolution 3D faces. We use the high-resolution data as gallery, while low-resolution faces are used as probe, allowing us to assess the real gap existing between these two types of data. Extensive experiments on low-resolution 3D face benchmarks show promising results with respect to state-of-the-art methods. [ABSTRACT FROM AUTHOR]

In a previous publication [1], we created a dataset of feature patches for detection model training. In this paper, we use the same patches to create a new large synthetic dataset of feature pairs, similar and different, in order to perform, thanks to a siamese convolutional model, the description and matching of the detected features. We thus complete the entire matching pipeline. The accurate manual labeling of image features being very difficult because of their large number and the various associated parameters of position, scale and rotation, recent deep learning models use the result of handcrafted methods for training. Compared to existing datasets, ours avoids model training with false detections of the extraction of feature patches by other algorithms, or with inaccuracy errors of manual labeling. The other advantage of synthetic patches is that we can control their content (corners, edges, etc.), as well as their geometric and photometric parameters, and therefore we control the invariance of the model. The proposed datasets thus allow a new approach to train the different matching modules without using traditional methods. To our knowledge, these are the first feature datasets based on generated synthetic patches for image matching.

Local feature descriptor plays a significant role in texture classification. However, in the traditional local binary pattern method, image pixels are converted into a binary pattern based on the relationship between center and neighborhood pixels. This paper introduces a novel feature extraction method named local neighbourhood edge responsive binary pattern to extract and categorize reliable texture features from images for achieving an automated classification of different types of ocean bottom sediments. Initially, the basic magnitude variance pixel values are derived depending on an odd and even pixel value of a 3 × 3 image patch. Further, the edge information is extracted using the local directional pattern method from all images. The edge response of the image is obtained using a kirsch mask in all eight directions. The encoding condition is then applied to both the local intensity and the edge information to create a unique descriptor value. Finally, a new learning algorithm called GMJAYA-ELM combines the Gaussian mutated JAYA (GMJAYA) with an extreme learning machine (ELM) for texture classification. The GMJAYA is used to optimize the input weights and hidden biases of single-hidden-layer feed-forward neural networks. Experimental findings indicate that the suggested method yields better accuracy and sensitivity efficiency among various groups. The proposed algorithm is tested by contrasting outcomes with standard learning schemes such as PSO-ELM, GA-ELM, ABC-ELM, Birdswarm-ELM, and JAYA-ELM, suggesting the dominance of GMJAYA-ELM. [ABSTRACT FROM AUTHOR]

A type of coronavirus disease called COVID‐19 is spreading all over the globe. Researchers and scientists are endeavoring to find new and effective methods to diagnose and treat this disease. This article presents an automated and fast system that identifies COVID‐19 from X‐ray radiographs of the chest using image processing and machine learning algorithms. Initially, the system extracts the feature descriptors from the radiographs of both healthy and COVID‐19 affected patients using the speeded up robust features algorithm. Then, visual vocabulary is built by reducing the number of feature descriptors via quantization of feature space using the K‐means clustering algorithm. The visual vocabulary train the support vector machine (SVM) classifier. During testing, an X‐ray radiograph's visual vocabulary is sent to the trained SVM classifier to detect the absence or presence of COVID‐19. The study used the dataset of 340 X‐ray radiographs, 170 images of each Healthy and Positive COVID‐19 class. During simulations, the dataset split into training and testing parts at various ratios. After training, the system does not require any human intervention and can process thousands of images with high precision in a few minutes. The performance of the system is measured using standard parameters of accuracy and confusion matrix. We compared the performance of the proposed SVM‐based classier with the deep‐learning‐based convolutional neural networks (CNN). The SVM yields better results than CNN and achieves a maximum accuracy of up to 94.12%. [ABSTRACT FROM AUTHOR]

With the development of the Internet, malware has become one of the most significant threats. Recognizing specific types of malware is an important step toward effective removal. Malware visualization is an important branch of malware static analysis techniques, where a piece of malware is turned into an image for visualization and classification. Despite great success, it is still difficult to extract effective texture feature representations for challenging datasets. The existing methods use global image features which are sensitive to relative code locations. In this paper, we present a new learning framework to obtain more discriminative and robust feature descriptors. The proposed method works with the existing local descriptors such as local binary patterns and dense scale-invariant feature transform, by grouping them into blocks and by using a new bag-of-visual-words model to obtain robust features, which are more flexible than global features and more robust than local features. We evaluate the proposed method on three malware databases. The experimental results demonstrate that the obtained descriptors lead to the state-of-the-art classification performance.

Traditionally, the indirect visual motion estimation and simultaneous localization and mapping (SLAM) systems were based on point features. In recent years, several SLAM systems that use lines as primitives were suggested. Despite the extra robustness and accuracy brought by the line segment matching, the line segment descriptors used in such systems were hand-crafted, and therefore sub-optimal. In this paper, we suggest applying descriptor learning to construct line segment descriptors optimized for matching tasks. We show how such descriptors can be constructed on top of a deep yet lightweight fully-convolutional neural network. The coefficients of this network are trained using an automatically collected dataset of matching and non-matching line segments. The use of the fully-convolutional network ensures that the bulk of the computations needed to compute descriptors is shared among the multiple line segments in the same image, enabling efficient implementation. We show that the learned line segment descriptors outperform the previously suggested hand-crafted line segment descriptors both in isolation (i.e., for the subtask of distinguishing matching and non-matching line segments), but also when built into the SLAM system. We construct a new line based SLAM pipeline built upon a state-of-the-art point-only system. We demonstrate generalization of the learned parameters of the descriptor network between two well-known datasets for autonomous driving and indoor micro aerial vehicle navigation.

The accuracy and efficiency of the object pose estimation is one of the key performance to realize the automation of production process and the intellectualization of warehousing and logistics. Some general mechanical parts pose estimation methods are studied. This work proposes a pose recognition method based on area sensing for mechanical parts characterized by abundant varieties of shapes and obvious structural features for mechanical parts pose estimation. This is achieved by combining the state-of-the-art semantic segmentation method and the point pair feature. In order to verify the effectiveness of the method，we conduct vast experiments and comparisons using multiple sets of data. The experimental results show that the method has higher pose recognition accuracy and can effectively improve the recognition efficiency. [ABSTRACT FROM AUTHOR]

In this paper, a novel benchmark is introduced for evaluating local image descriptors. We demonstrate limitations of the commonly used datasets and evaluation protocols, that lead to ambiguities and contradictory results in the literature. Furthermore, these benchmarks are nearly saturated due to the recent improvements in local descriptors obtained by learning from large annotated datasets. To address these issues, we introduce a new large dataset suitable for training and testing modern descriptors, together with strictly defined evaluation protocols in several tasks such as matching, retrieval and verification. This allows for more realistic, thus more reliable comparisons in different application scenarios. We evaluate the performance of several state-of-the-art descriptors and analyse their properties. We show that a simple normalisation of traditional hand-crafted descriptors is able to boost their performance to the level of deep learning based descriptors once realistic benchmarks are considered. Additionally we specify a protocol for learning and evaluating using cross validation. We show that when training state-of-the-art descriptors on this dataset, the traditional verification task is almost entirely saturated. [ABSTRACT FROM AUTHOR]

In this paper, we examine a novel data augmentation (DA) method that transforms an image into a new image containing multiple rotated copies of the original image. The DA method creates a grid of $n \times n$ cells, in which each cell contains a different randomly rotated image and introduces a natural background in the newly created image. We investigate the use of deep learning to assess the classification performance on the rotation matrix or original dataset with colour constancy versions of the datasets. For the colour constancy methods, we use two well-known retinex techniques: the multi-scale retinex and the multi-scale retinex with colour restoration for enhancing both original (ORIG) and rotation matrix (ROT) images. We perform experiments on three datasets containing images of animals, from which the first dataset is collected by us and contains aerial images of cows or non-cow backgrounds. To classify the Aerial UAV images, we use a convolutional neural network (CNN) architecture and compare two loss functions (hinge loss and cross-entropy loss). Additionally, we compare the CNN to classical feature-based techniques combined with a k-nearest neighbour classifier or a support vector machine. The best approach is then used to examine the colour constancy DA variants, ORIG and ROT-DA alone for three datasets (Aerial UAV, Bird-600 and Croatia fish). The results show that the rotation matrix data augmentation is very helpful for the Aerial UAV dataset. Furthermore, the colour constancy data augmentation is helpful for the Bird-600 dataset. Finally, the results show that the fine-tuned CNNs significantly outperform the CNNs trained from scratch on the Croatia fish and the Bird-600 datasets, and obtain very high accuracies on the Aerial UAV and Bird-600 datasets.

In the field of visual tracking, the methods of Discriminative Correlation Filters (DCF) have showed excellent performance, which rely heavily on the choice of feature descriptors. The Continuous Convolution Operator Tracker (C-COT) is a novel correlation filter to track the target position in the continuous domain, which achieved significant effects. However, as for various visual scenes, different feature descriptors are suitable to different environments. If each feature channel is given the same confidence during the tracking phase, it would limit the performance of some good features. To address this problem, this paper proposes an improved C-COT algorithm that can adaptively perform feature channel weighting. The Average Peak Correlation Energy (APCE) is used to evaluate the corresponding response map of each feature channel, guiding the target appearance model to give different weights to different features. Then, we can obtain the final weighted feature response map whose peak value is applied to locate the target. In addition, the C-COT updates the appearance model rigorously in every frame, which may lead to over-fitting and increase computional complexity. Therefore, in order to reduce the redundancy of the online training sample and avoid similar background interference, we adopts the method of Peak Side Lobe Ratio (PSLR) to update the model. We perform comprehensive experiments on OTB50 and OTB100. The results show that the improved tracker achieves better accuracy, especially in some specific video scenes. In addition, speed has also improved.

Rotary kilns are important industrial plants to process material. In this article, we address the problem of detecting material rings formed in alumina rotary kilns, which leads to high waste and substantial economic loss. This necessitates the use of artificial intelligence to help kiln operators detect material rings promptly. We describe a recently developed system that extracts useful features and detects the presence of material rings efficiently and accurately in real time. Our major contribution is a novel feature extraction method based on geometric transformation and peak localization exploiting the domain knowledge. We also contribute a large data set covering many thousand labeled frames for the evaluation of ring detection. On this data set, we demonstrate that our novel feature extraction method outperforms other alternatives in the literature. Our method is also demonstrated as favorable over deep learning approaches. Our system produces an overall accuracy of nearly 96%, which is acceptable for deployment. [ABSTRACT FROM AUTHOR]

Optical-based navigation for space is a field growing in popularity due to the appeal of efficient techniques such as Visual Simultaneous Localisation and Mapping (VSLAM), which rely on automatic feature tracking with low-cost hardware. However, low-level image processing algorithms have traditionally been measured and tested for ground-based exploration scenarios. This paper aims to fill the gap in the literature by analysing state-of-the-art local feature detectors and descriptors with a taylor-made synthetic dataset emulating a Non-Cooperative Rendezvous (NCRV) with a complex spacecraft, featuring variations in illumination, rotation, and scale. Furthermore, the performance of the algorithms on the Long Wavelength Infrared (LWIR) is investigated as a possible solution to the challenges inherent to on-orbit imaging in the visible, such as diffuse light scattering and eclipse conditions. The Harris, GFTT, DoG, Fast-Hessian, FAST, CenSurE detectors and the SIFT, SURF, LIOP, ORB, BRISK, FREAK descriptors are benchmarked for images of Envisat. It was found that a combination of Fast-Hessian with BRISK was the most robust, while still capable of running on a low resolution and acquisition rate setup. For large baselines, the rate of false-positives increases, limiting their use in model-based strategies. [ABSTRACT FROM AUTHOR]

Industry 4.0 requires flexible and fast solutions to automatically identify and handle different parts during the production process. For such a multiclassification (MCC) task, usually feature detectors or machine learning approaches are used. However, increasing product variety and consequently rising number of classes impel the expansion of data set sizes, a mandatory procedure for accurate object identification. The acquisition of a sufficiently large number of sample images is a costly and time-consuming issue. This work presents an iterative reference data set creation method of small size image data sets for feature descriptor-based MCC. The novel method compares the shape of an object in a candidate image with the shapes from a growing reference image set consisting of all previously accepted candidates. An image is assigned to or rejected from the reference data set depending on the found shape similarity. Rejected images form a grid search set that is later used to optimize the feature descriptors' hyperparameters and enable the addition of new classes. The benefits of this method are the small number of images to be acquired for MCC, the possibility of adding new parts without re-training, the little overhead for new applications, and its compatibility with most commonly used feature descriptors. When compared with a small control data set provided by an inexperienced user with an accuracy of 49% for a five classes MCC, the reference data set built by the novel method gains 20% on the accuracy (69%, 21 images in total) and can be performed by the same inexperienced user. [ABSTRACT FROM AUTHOR]

In this study, the authors propose a novel methodology for static gesture recognition in a complex background using only depth map from Microsoft's Kinect camera. Four different types of features are extracted and analysed on two public static gesture datasets. The features extracted from the segmented hand are geometrical, local binary patterns, number of fingers (Num) raised in a gesture and distance of hand palm centre from the fingertips and the valley between the fingers. The hand region is first segmented from the image using depth data followed by the forearm removal. Four multi‐class support vector machine (SVM) kernels are also compared and used for recognition of gestures with extracted feature vector as an input. The experimental results achieved recognition accuracy of 99 and 95.7% on two public complex static gesture datasets using Gaussian SVM kernel function as a classifier. The proposed approach is found to be comparable and even outperforms some of the state‐of‐the‐art techniques in terms of high recognition accuracies, even after using a single cue for hand segmentation and extraction of features in the complex background which results in non‐dependency on too many cues and much hardware. [ABSTRACT FROM AUTHOR]

Quorum-sensing peptides (QSPs) are the signal molecules that are closely associated with diverse cellular processes, such as cell–cell communication, and gene expression regulation in Gram-positive bacteria. It is therefore of great importance to identify QSPs for better understanding and in-depth revealing of their functional mechanisms in physiological processes. Machine learning algorithms have been developed for this purpose, showing the great potential for the reliable prediction of QSPs. In this study, several sequence-based feature descriptors for peptide representation and machine learning algorithms are comprehensively reviewed, evaluated and compared. To effectively use existing feature descriptors, we used a feature representation learning strategy that automatically learns the most discriminative features from existing feature descriptors in a supervised way. Our results demonstrate that this strategy is capable of effectively capturing the sequence determinants to represent the characteristics of QSPs, thereby contributing to the improved predictive performance. Furthermore, wrapping this feature representation learning strategy, we developed a powerful predictor named QSPred-FL for the detection of QSPs in large-scale proteomic data. Benchmarking results with 10-fold cross validation showed that QSPred-FL is able to achieve better performance as compared to the state-of-the-art predictors. In addition, we have established a user-friendly webserver that implements QSPred-FL, which is currently available at http://server.malab.cn/QSPred-FL. We expect that this tool will be useful for the high-throughput prediction of QSPs and the discovery of important functional mechanisms of QSPs. [ABSTRACT FROM AUTHOR]

In military application target tracking has always been an interesting and challenging problem. Nowadays, it has also found its place in civil applications, especially concerning surveillance and monitoring. Until recently, thermal imagery (image is formed based on infrared spectrum radiation) was considered only in military applications because of the price and size of cameras. Also, thermal image quality was not as good as TV (image is formed based on visual spectrum radiation) camera image. The situation has changed, and thermal cameras are now widely used in many kinds of applications. Thermal image is different from TV camera image, as it measures temperature difference between objects and background. Therefore, it has an advantage over television cameras, since it can be used in low light conditions and in dark. This paper examines the options for a coarse and quick algorithm for rough target locating in thermal image, where the target is a small Unmanned Aerial Vehicle (UAV). Three different feature descriptor algorithms are tested in thermal imagery target tracking. Feature descriptor methods are widely used in visual imagery, but the goal of this paper is to examine their usage in thermal imagery. That is why three different feature descriptor algorithms from three different families are tested: FREAK (Fast Retina Keypoint), SURF (Speeded Up Robust Features) and MSER (Maximally Stable Extremal Regions). The algorithms are tested in case of translation, rotation, blur and size change of an object of interest, as well as in the case of noisy image. Since none of the tested methods works great in different situations, new, multi-stage algorithm is proposed. This algorithm is based on MSER and SURF algorithm combination, with a goal to use the advantages of each of them in different real situations. The obtained results show that the new, multi-stage algorithm has got the best performance among the group of the tested methods. All the algorithms are implemented in Matlab software.

Successful treatment of pulmonary tuberculosis (TB) depends on early diagnosis and careful monitoring of treatment response. Identification of acid-fast bacilli by fluorescence microscopy of sputum smears is a common tool for both tasks. Microscopy-based analysis of the intracellular lipid content and dimensions of individual Mycobacterium tuberculosis (Mtb) cells also describe phenotypic changes which may improve our biological understanding of antibiotic therapy for TB. However, fluorescence microscopy is a challenging, time-consuming and subjective procedure. In this work, we automate examination of fields of view (FOVs) from microscopy images to determine the lipid content and dimensions (length and width) of Mtb cells. We introduce an adapted variation of the UNet model to efficiently localising bacteria within FOVs stained by two fluorescence dyes; auramine O to identify Mtb and LipidTox Red to identify intracellular lipids. Thereafter, we propose a feature extractor in conjunction with feature descriptors to extract a representation into a support vector multi-regressor and estimate the length and width of each bacterium. Using a real-world data corpus from Tanzania, the proposed method i) outperformed previous methods for bacterial detection with a 8% improvement (Dice coefficient) and ii) estimated the cell length and width with a root mean square error of less than 0.01%. Our network can be used to examine phenotypic characteristics of Mtb cells visualised by fluorescence microscopy, improving consistency and time efficiency of this procedure compared to manual methods., Competing Interests: Declaration of competing interest The authors would like to declare no conflict of interest., (Copyright © 2023 The Author(s). Published by Elsevier Ltd.. All rights reserved.)

Human perception of dimensions is usually limited to two or three degrees. Any further increase in the number of dimensions usually leads to the difficulty in visual imagination for any person. Hence, machine learning researchers often commonly have to overcome the curse of dimensionality in high dimensional feature sets with dimensionality reduction techniques. In this proposed model, two handwritten digit datasets are used: CVL Single Digit and MNIST, and two popular feature descriptors, Histogram of Oriented Gradients (HOG) and Gabor filters, are used to generate the feature sets. Investigations are carried out on linear and nonlinear transformations of the feature sets using multiple dimensionality reduction techniques such as Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA) and Isomap. The lower dimension vectors obtained, are then used to classify the numeric digits using Support Vector Machine (SVM). A conclusion arrived is that using HOG as the feature descriptor and PCA as the dimensionality reduction technique resulted in the experimental model achieving the highest accuracy of 99.29% on the MNIST dataset with the time efficiency comparable to that of a convolutional neural network (CNN). Further, it is concluded that even though the LDA model with HOG as the feature descriptor achieved a lesser accuracy of 98.34%, but it was able to capture maximum information in just 9 components in its lower dimensional subspace with 75% reduction in time efficiency of that of the PCA-HOG model and the CNN model.

We present an unsupervised framework that automatically detects objects of interest in images by formulating the general problem of semantic segmentation as community detection problem in graphs. The framework broadly follows a four-step procedure. First, we perform an over-segmentation of the original image using the well-known statistical region merging algorithm which presents the advantage of not requiring any quantization or colour space transformations. Second, we compute the feature descriptors of the resulting segmented regions. For encoding colour and other textural information, each region is described by an hybrid descriptor based on colour histograms and covariance matrix descriptor. Third, from the set of descriptors we construct different weighted graphs using various graph construction algorithms. Finally, the resulting graphs are then divided into groups or communities using a community detection algorithm based on spectral modularity maximization. This algorithm makes use of the eigenspectrum of matrices such as the graph Laplacian matrix and the modularity matrix which are more likely to reveal the community structure of the graph. Experiments conducted on large orthophotos depicting several zones in the region of Belfort city situated on the north-eastern of France provide promising results. The proposed framework can be used by semi-automatic approaches to handle the challenging problems of scene parsing. [ABSTRACT FROM AUTHOR]

Iterative back-projection (IBP) is a popular and straightforward approach applied successfully in the field of image super-resolution reconstruction (SRR). SRR using IBP (SRR-IBP) methods efficiently satisfies the basic reconstruction constraints. In spatial domain applications, it allows easy inclusion of data and is a computationally efficient method. However, inferior convergence rate, sensitivity to the initial choice of image, the presence of different degrees of ringing artifacts are some of the major disadvantages that limit the performance of SRR-IBP. To relieve these inherent limitations, an evolutionary edge preserving IBP (EEIBP) is proposed in this paper. The proposed work introduces an improved initial choice of the digital image by interpolating the low-resolution digital image via hybridizing the notion of uniform and non-uniform B-spline interpolation. Secondly, it incorporates a spatially adaptive back-projecting kernel (SABPK) and regularization constraints in the iterative process. The SABPK utilizes covariance-based adaptation to restore the lost high-frequency details and is regulated by a control parameter to make the reconstruction process robust. The regularization constraints use different low-level feature descriptors to track the information related to shape and salient visual properties of the digital image. Finally, the overall reconstruction error is minimized via GA, PSO and cuckoo search (CS) algorithms. Experimental results demonstrate the robustness and the effectiveness of the proposed EEIBP method to provide a high-resolution solution with improved visual perception and reduced artifacts. Moreover, EEIBP method optimized via CS algorithm enables a better quality of reconstruction as compared the other search algorithms (gradient, GA and PSO). [ABSTRACT FROM AUTHOR]