Search

The present study addresses the segmentation and the 3D reconstruction of the corpus callosum from MRI scans. Accurate segmentation of the corpus callosum is essential in order to enable its reconstruction and 3D visualization to facilitate early diagnosis. In fact, many studies have established a strong correlation between the shape of the corpus callosum and several pathological conditions. However, the segmentation is made difficult by regions of similar intensity within the MRI images. To overcome this challenge, we propose an automated method that relies mainly on a probabilistic neural network applied to superpixels. The proposed scheme involves segmenting the corpus callosum within the MRI scans, followed by the application of the marching cubes technique in order to generate 3D volumes. The effectiveness of the proposed method has been extensively validated on four challenging datasets (OASIS, ABIDE, MIRIAD, and SBD), and the obtained results demonstrate its superior performance compared to other state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Hyperspectral image (HI) analysis approaches have recently become increasingly complex and sophisticated. Recently, the combination of spectral-spatial information and superpixel techniques have addressed some hyperspectral data issues, such as the higher spatial variability of spectral signatures and dimensionality of the data. However, most existing superpixel approaches do not account for specific HI characteristics resulting from its high spectral dimension. In this work, we propose a multiscale superpixel method that is computationally efficient for processing hyperspectral data. The Simple Linear Iterative Clustering (SLIC) oversegmentation algorithm, on which the technique is based, has been extended hierarchically. Using a novel robust homogeneity testing, the proposed hierarchical approach leads to superpixels of variable sizes but with higher spectral homogeneity when compared to the classical SLIC segmentation. For validation, the proposed homogeneity-based hierarchical method was applied as a preprocessing step in the spectral unmixing and classification tasks carried out using, respectively, the Multiscale sparse Unmixing Algorithm (MUA) and the CNN-Enhanced Graph Convolutional Network (CEGCN) methods. Simulation results with both synthetic and real data show that the technique is competitive with state-of-the-art solutions. [ABSTRACT FROM AUTHOR]

In order to enhance the accuracy and efficiency of color difference detection for dyed fabrics, this article introduces a novel superpixel‐based color difference detection algorithm, which exhibits three significant improvements over the previous algorithm. First, it proposes a more rational color feature extraction method to extract more detailed color features from fabric images, which uses superpixels as basic units for both color feature extraction and color difference detection. Second, the algorithm employs the CIEDE2000 color difference formula to calculate color difference values. This eliminates the need for obtaining standard fabric images as references in advance, making the process more convenient and practical. Finally, an area‐based color difference calculation method is proposed based on the human visual mechanism, ensuring that the evaluation results are more consistent with human eye observations. Experimental results demonstrate that the proposed algorithm provides more accurate and visually consistent color difference assessments compared to the previous algorithm. [ABSTRACT FROM AUTHOR]

The Covid-19 disease has spread widely over the whole world since the beginning of 2020. Following the epidemic which started in Wuhan, China on January 30, 2020 the World Health Organization (WHO) declared a global health emergency and a pandemic. Researchers of different disciplines work along with public health officials to understand the SARS-CoV-2 pathogenesis and jointly with the policymakers urgently develop strategies to control the spread of this new disease. Recent findings have observed specific image patterns from computed tomography (CT) for patients infected by SARS-CoV-2 which are distinct from the other pulmonary diseases. In this paper, we propose an explainable-by-design that has an integrated image segmentation mechanism based on SLIC that improves the algorithm performance and the interpretability of the resulting model. In order to evaluate the proposed approach, we used the SARS-CoV-2 CT scan dataset that we published recently and has been widely used in the literature. The proposed Super-xDNN could obtain statistically better results than traditional deep learning approaches as DenseNet-201 and Resnet-152. Furthermore, it also improved the explainability and interpretability of its decision mechanism when compared with the xDNN basis approach that uses the whole image as prototype. The segmentation mechanism of Super-xDNN favored a decision structure that is more close to the human logic. Moreover, it also allowed the provision of new insights as a heat-map which highlights the areas with highest similarities with Covid-19 prototypes, and an estimation of the area affected by the disease. [ABSTRACT FROM AUTHOR]

The superpixel usually serves as a region-level feature in various image processing tasks, and is known for segmentation accuracy, spatial compactness and running efficiency. However, since these properties are intrinsically incompatible, there is still a compromise within the overall performance of existing superpixel algorithms. In this work, the property constraint in superpixels is relaxed by in-depth understanding of the image content, and a novel two-stage superpixel generation framework is proposed to produce content-aware superpixels. In the global processing stage, a diffusion-based online average clustering framework is introduced to efficiently aggregate image pixels into multiple superpixel candidates according to color and spatial information. During this process, a centroid relocation strategy is established to dynamically guide the region updating. According to the area feature in manifold space, several superpixel centroids are then split or merged to optimize the regional representation of image content. Subsequently, local updating is adopted on pixels in those superpixel regions to further improve the performance. As a result, the dynamic centroid relocating strategy offers online averaging clustering the property of content awareness through coarse-to-fine label updating. Extensive experiments verify that the produced superpixels achieve desirable and comprehensive performance on boundary adherence, visual satisfactory and time consumption. The quantitative results are on par with existing state-of-the-art algorithms in terms with several common property metrics. [ABSTRACT FROM AUTHOR]

Abstract The expansion of scientific image data holds great promise to quantify individuals, size distributions and traits. Computer vision tools are especially powerful to automate data mining of images and thus have been applied widely across studies in aquatic and terrestrial ecology. Yet marine benthic communities, especially infauna, remain understudied despite their dominance of marine biomass, biodiversity and playing critical roles in ecosystem functioning. Here, we disaggregated infauna from sediment cores taken throughout the spring transition (April–June) from a near‐natural mesocosm setup under experimental warming (Ambient, +1.5°C, +3.0°C). Numerically abundant mudsnails were imaged in batches under stereomicroscopy, from which we automatically counted and sized individuals using a superpixel‐based segmentation algorithm. Our segmentation approach was based on clustering superpixels, which naturally partition images by low‐level properties (e.g., colour, shape and edges) and allow instance‐based segmentation to extract all individuals from each image. We demonstrate high accuracy and precision for counting and sizing individuals, through a procedure that is robust to the number of individuals per image (5–65) and to size ranges spanning an order of magnitude (40k individuals and revealed overall growth in response to springtime warming. We illustrate that image processing and segmentation workflows can be built upon existing open‐access R packages, underlining the potential for wider adoption of computer vision tools among ecologists. The image‐based approach also generated reproducible data products that, alongside our scripts, we have made freely available. This work reinforces the need for next‐generation monitoring of benthic communities, especially infauna, which can display differential responses to average warming.

The increased interest in deep learning applications, and their hard-to-detect biases result in the need to validate and explain complex models. However, current explanation methods are limited as far as both the explanation of the reasoning process and prediction results are concerned. They usually only show the location in the image that was important for model prediction. The lack of possibility to interact with explanations makes it difficult to verify and understand exactly how the model works. This creates a significant risk when using the model. The risk is compounded by the fact that explanations do not take into account the semantic meaning of the explained objects. To escape from the trap of static and meaningless explanations, we propose a tool and a process called LIMEcraft. LIMEcraft enhances the process of explanation by allowing a user to interactively select semantically consistent areas and thoroughly examine the prediction for the image instance in case of many image features. Experiments on several models show that our tool improves model safety by inspecting model fairness for image pieces that may indicate model bias. The code is available at: http://github.com/MI2DataLab/LIMEcraft. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Remote Sensing is the property of Editorial Office of Journal of Remote Sensing & Science Publishing Co. 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 correct segmentation of brain tissues in magnetic resonances images helps the development of higher accuracy diagnoses. A model that allows segmentation soft tissue in brain magnetic resonance images is proposed, where are discriminated between grey matter, white matter and background. To segment images, the single linear iterative clustering of superpixels algorithm was used, as well as the classification between the soft matters of the images was carried out with an artificial neural network of the multilayer perceptron type. To measure the quality of the classification in the segmentation, the Sorensen-Dice coefficient was used, as well as the results of the proposed methodology was compared against the k-means algorithm. The proposed methodology reached a Sorensen-Dice coefficient of 0.921 and 0.963, for gray and white matter, respectively, for images with 0% noise, while the k-means only reached 0.911 and 0.958 under the same conditions. In images with 9% noise, the difference was even greater, with a 0.863 and 0.937 classification between grey and white matter, respectively, while for the k-means algorithm it decreased to 0.805 and 0.901, respectively. When the image is dividing in superpixels the structure of the brain regions is better preserved than other methodologies as k- means. Therefore, the neural network works individually in the classification with each superpixel. Where a multilayer perceptron network can be used as accuracy classifier of brain tissue in magnetic resonance imaging was proven. [ABSTRACT FROM AUTHOR]

The expansion of scientific image data holds great promise to quantify individuals, size distributions and traits. Computer vision tools are especially powerful to automate data mining of images and thus have been applied widely across studies in aquatic and terrestrial ecology. Yet marine benthic communities, especially infauna, remain understudied despite their dominance of marine biomass, biodiversity and playing critical roles in ecosystem functioning.Here, we disaggregated infauna from sediment cores taken throughout the spring transition (April–June) from a near‐natural mesocosm setup under experimental warming (Ambient, +1.5°C, +3.0°C). Numerically abundant mudsnails were imaged in batches under stereomicroscopy, from which we automatically counted and sized individuals using a superpixel‐based segmentation algorithm. Our segmentation approach was based on clustering superpixels, which naturally partition images by low‐level properties (e.g., colour, shape and edges) and allow instance‐based segmentation to extract all individuals from each image.We demonstrate high accuracy and precision for counting and sizing individuals, through a procedure that is robust to the number of individuals per image (5–65) and to size ranges spanning an order of magnitude (<750 μm to 7.4 mm). The segmentation routine provided at least a fivefold increase in efficiency compared with manual measurements. Scaling this approach to a larger dataset tallied >40k individuals and revealed overall growth in response to springtime warming.We illustrate that image processing and segmentation workflows can be built upon existing open‐access R packages, underlining the potential for wider adoption of computer vision tools among ecologists. The image‐based approach also generated reproducible data products that, alongside our scripts, we have made freely available. This work reinforces the need for next‐generation monitoring of benthic communities, especially infauna, which can display differential responses to average warming. [ABSTRACT FROM AUTHOR]

Graph neural networks (GNNs) offer a potent framework for learning and representing data with graph structures. In the context of hyperspectral images (HSIs) classification tasks, GNNs have been successfully employed. Among the diverse variations of GNNs, graph attention networks (GATs) stand out, demonstrating cutting-edge performance in node prediction. GATs excel by acquiring the capability to assign dense attention coefficients to each neighboring node during the process of feature aggregation. First, the intricate nature of land distribution, coupled with the high dimensionality of hyperspectral imaging (HSI) data, poses a hurdle for the direct application of GATs in effectively distinguishing between various coverage categories. Second, the insufficiency of spectral information further compounds the issue by adversely affecting classification accuracy, contributing to GATs’ susceptibility to overfitting during training. Additionally, prevailing classification approaches commonly derive features either at the pixel or superpixel levels. The utilization of single-scale feature extraction techniques proves inadequate in concurrently addressing both local and global features of land cover. This limitation contributes to less-than-optimal classification outcomes. To address this issue, this article proposes a double-branch multilevel skip sparse GAT for HSI classification. The network comprises two parallel branches of sparse GATs, extracting features from the perspectives of pixels and superpixels. Via the construction of a multilevel skip graph, these two branches adeptly gather both local and global context information. Subsequently, a fully connected network is deployed to classify the integrated features derived from both branches. The effectiveness of the proposed method is validated through experimental results on three benchmark HSI datasets, showcasing its superior performance.

Polarimetric synthetic aperture radar (PolSAR) has attracted more attentions because of its excellent observation ability, and PolSAR image classification has become one of the significant tasks in remote sensing interpretation. Various types and sizes of land cover objects lead to misclassification, especially in the boundaries of different categories. To solve these issues, a multiscale superpixel-guided weighted graph convolutional network (MSGWGCN) is proposed for classifying PolSAR images. In the proposed MSGWGCN, multiscale superpixel features are imported into the weighted graph convolutional network to obtain higher level representation, which can make full use of land cover object information in PolSAR images. Moreover, to fuse pixel-level features at different scales, a multiscale feature cascade fusion module is built, which plays an important role in preserving classification details. Experiments on three PolSAR datasets indicate that the proposed MSGWGCN performs better than other advanced methods on PolSAR classification task.

In recent years, superpixel segmentation has been widely used in image processing tasks as a preprocessing step. Superpixel segmentation aims to group pixels into homogeneous regions while maintaining edges. This paper proposes a superpixel segmentation algorithm based on boundary preservation. In the algorithm, the side window filtering is first used to smooth the image texture area, so that the superpixel shape generated at the texture is regular. Different from other superpixel clustering algorithms, the algorithm in this paper uses a new distance measurement function for distance measurement, which can assign different weights to its color distance items and spatial distance items according to different pixels, so that the superpixel fit in the image boundary area. The boundary is regular in the flat area. The distance measurement function also takes into account the pixel information of the linear path from the pixel to the cluster center, and avoids the category error division caused by only the local information of the pixel for clustering. Finally, this paper designs a new cluster center update strategy, which uses only the weighted average of some reliable pixels in the superpixel as the new cluster center, thereby reducing the update of the cluster center of pixels that are not very similar to the cluster center. The interference makes the cluster center update more accurate. Experimental results show that our algorithm can get better results in visual effects and BR,UE,ASA indicators compared with existing algorithms.

Extracting building from high-resolution (HR) remote sensing imagery (RSI) serves a variety of areas, such as smart city, environment management, and emergency disaster services. Previous building extraction methods primarily focus on pixel-level and superpixel-level features, which do not fully utilize the superpixel-level spatial context, leaving room for performance improvement. To bridge the gap, this study incorporates spatial context of both pixels and superpixels for building extraction of HR RSI. Specifically, the proposed method develops a trainable superpixel segmentation module to segment HR RSI into superpixels by fusing pixel features and pixel-level context. And a superpixel-level context aggregation module is devised to incorporate the multiple-scale spatial context of superpixels to extract buildings. Experiments on public challenging datasets show that our method is superior to the state-of-the-art baselines in accuracy, with better building boundaries and higher integrity. This study explores a new approach for HR RSI building extraction by introducing spatial context of superpixels, and a methodological reference for the HR RSI interpretation tasks.

The application of continuous-time quantum walk in the field of image segmentation has attracted much attention due to the advantages of quantum computation. However, the proposed image segmentation algorithm constructs the continuous-time quantum walk model based on pixels, which will cause a huge burden on quantum resources, and the various feature information of the image cannot be better considered. In addition, this pixel-based processing method requires a lot of manual annotation to achieve the desirable segmentation effect. To address these issues, we propose an image segmentation algorithm using continuous-time quantum walk based on superpixels. In our segmentation algorithm, the original image is firstly segmented into superpixels, and then a weighted graph is constructed with superpixels as nodes, where the weight of edges in graph is measured by the feature similarity of two adjacent superpixels, which consists of color features and texture features. Next, the continuous-time quantum walk model is constructed based on the weighted graph by redefining the new Hamiltonian operator. Finally, continuous-time quantum walk is executed and the image segmentation result can be obtained, which is realized by assigning each superpixel the class label corresponding to the greatest probability. Experiments on the BSD500 dataset show that the proposed algorithm can significantly improve segmentation efficiency and accuracy while the manually selected seeds is reduced by 91%. More importantly, the new algorithm reduce the demision of the quantum walk system by more than 99%, which will yield a huge saving on the quantum resources. [ABSTRACT FROM AUTHOR]

Failures in urban areas' solid waste management lead to clandestine garbage dumping and pollution. This affects sanitation and public human hygiene, deteriorates quality of life, and contributes to deprivation. This study aimed to test a combination of machine learning, high-resolution earth observation and GIS data to detect diverse categories of residual waste on the streets, such as sacks and construction debris. We conceptualised five different classes of solid waste from image interpretation: "Sure", "Half-sure", "Not-sure", "Dispersed", and "Non-garbage". We tested a combination of k-means-based segmentation and supervised random forest to investigate the capabilities of automatic classification of these waste classes. The model can detect the presence of solid waste on the streets and achieved an accuracy of up from 73.95%–95.76% for the class "Sure". Moreover, a building extraction using an EfficientNet deep-learning-based semantic segmentation allowed masking the rooftops. This improved the accuracy of the classes "Sure" and "Non-garbage". The systematic evaluation of all parameters considered in this model provides a robust and reliable method of solid waste detection for decision-makers. These results highlight areas where insufficient waste management affects the citizens of a given city. The best segmentation using simple linear iterative clustering (SLIC) was achieved with the parameter values 8,000 segments and 0.3 compactness. The following supervised classification of the segmented images using Random Forest yielded an average overall accuracy of 80.18%. The model can detect the presence of solid waste on the streets and achieved an accuracy of up from 73.95%–95.76% for the class "Sure". The average reflectance values of the classes "Sure" and "non-Garbage" overlapped. Removing the building rooftops from the orthotiles reduced the overlap of the classes mentioned above. This allowed better identification of the class "Sure". Moreover, rooftop removal helped improve the accuracy of the classifier, from 59.51% to 90.18% to 71.53% to 95.76% in study areas with and without rooftops, respectively. [ABSTRACT FROM AUTHOR]

The superpixel usually serves as a region-level feature in various image processing tasks, and is known for segmentation accuracy, spatial compactness and running efficiency. However, since these properties are intrinsically incompatible, there is still a compromise within the overall performance of existing superpixel algorithms. In this work, the property constraint in superpixels is relaxed by in-depth understanding of the image content, and a novel two-stage superpixel generation framework is proposed to produce content-aware superpixels. In the global processing stage, a diffusion-based online average clustering framework is introduced to efficiently aggregate image pixels into multiple superpixel candidates according to color and spatial information. During this process, a centroid relocation strategy is established to dynamically guide the region updating. According to the area feature in manifold space, several superpixel centroids are then split or merged to optimize the regional representation of image content. Subsequently, local updating is adopted on pixels in those superpixel regions to further improve the performance. As a result, the dynamic centroid relocating strategy offers online averaging clustering the property of content awareness through coarse-to-fine label updating. Extensive experiments verify that the produced superpixels achieve desirable and comprehensive performance on boundary adherence, visual satisfactory and time consumption. The quantitative results are on par with existing state-of-the-art algorithms in terms with several common property metrics.

ABSTRACTFailures in urban areas’ solid waste management lead to clandestine garbage dumping and pollution. This affects sanitation and public human hygiene, deteriorates quality of life, and contributes to deprivation. This study aimed to test a combination of machine learning, high-resolution earth observation and GIS data to detect diverse categories of residual waste on the streets, such as sacks and construction debris. We conceptualised five different classes of solid waste from image interpretation: “Sure”, “Half-sure”, “Not-sure”, “Dispersed”, and “Non-garbage”. We tested a combination of k-means-based segmentation and supervised random forest to investigate the capabilities of automatic classification of these waste classes. The model can detect the presence of solid waste on the streets and achieved an accuracy of up from 73.95%–95.76% for the class “Sure”. Moreover, a building extraction using an EfficientNet deep-learning-based semantic segmentation allowed masking the rooftops. This improved the accuracy of the classes “Sure” and “Non-garbage”. The systematic evaluation of all parameters considered in this model provides a robust and reliable method of solid waste detection for decision-makers. These results highlight areas where insufficient waste management affects the citizens of a given city.

Infrared and visible image fusion aims to highlight the infrared target and preserve valuable texture details as much as possible. However, the infrared target needs to be more apparent in most image fusion methods. A large amount of infrared noise remains in the fusion results, significantly reducing the proportion of valuable texture details in the fusion results. How to highlight the salient of infrared targets, lower noise, and retain more valuable texture details in the fusion results still need to be solved. We propose an adaptive salient region analysis method based on superpixels (SSRA) for infrared and visible fusion to solve this problem. This method uses salient region analysis based on superpixels to highlight the salience region effectively. We design a texture detail fusion method based on brightness analysis of the visible image to suppress noise and keep more meaningful texture detail information. The experimental results show that our proposed method performs better in subjective vision and quantitative evaluation than some advanced methods. In addition, we also demonstrate that SSRA is capable of supporting high-level visual tasks well. Our code is publicly available at: https://github.com/VCMHE/SSRA. [ABSTRACT FROM AUTHOR]

Explaining the predictions of neural networks to comprehend which region of an image influences the most its decision has become an imperative prerequisite when classifying medical images. In the case of convolutional neural networks, gradient-weighted class activation mapping is an explainability scheme that is more than often utilized for the unveiling of connections between stimuli and predictions especially in classification tasks that address the determination of the class between distinct objects in an image. However, certain categories of medical imaging such as confocal and histopathology images contain rich and dense information that differs from the cat versus dog paradigm. To further improve the performance of the gradient-weighted class activation mapping technique and the generated visualizations, we propose a segmentation-based explainability scheme that focuses on the common visual characteristics of each segment in an image to provide enhanced visualizations instead of highlighting rectangular regions. The explainability performance was quantified by applying random noise perturbations on microscopy images. The area over perturbation curve is utilized to demonstrate the improvement of the proposed methodology when utilizing the Slic superpixel algorithm against the Grad-CAM technique by an average of 4% for the confocal dataset and 9% for histopathology dataset. The results show that the generated visualizations are more comprehensible to humans than the initial heatmaps and demonstrate improved performance against the original Grad-CAM technique. [ABSTRACT FROM AUTHOR]

PurposeDevelop a hierarchical longitudinal regression model for estimating local rates of change of macular ganglion cell complex (GCC) measurements with optical coherence tomography (OCT).MethodsWe enrolled 112 eyes with four or more macular OCT images and ≥2 years of follow-up. GCC thickness measurements within central 6 × 6 superpixels were extracted from macular volume scans. We fit data from each superpixel separately with several hierarchical Bayesian random-effects models. Models were compared with the Watanabe-Akaike information criterion. For our preferred model, we estimated population and individual slopes and intercepts (baseline thickness) and their correlation.ResultsMean (SD) follow-up time and median (interquartile range) baseline 24-2 visual field mean deviation were 3.6 (0.4) years and -6.8 (-12.2 to -4.3) dB, respectively. The random intercepts and slopes model with random residual variance was the preferred model. While more individual and population negative slopes were observed in the paracentral and papillomacular superpixels, superpixels in the superotemporal and inferior regions displayed the highest correlation between baseline thickness and rates of change (r = -0.43 to -0.50 for the top five correlations).ConclusionsA Bayesian linear hierarchical model with random intercepts/slopes and random variances is an optimal initial model for estimating GCC slopes at population and individual levels. This novel model is an efficient method for estimating macular rates of change and probability of glaucoma progression locally.Translational relevanceThe proposed Bayesian hierarchical model can be applied to various macular outcomes from different OCT devices and to superpixels of variable sizes to estimate local rates of change and progression probability.

PurposeTo compare local ganglion cell-inner plexiform layer (GCIPL) thickness measurements between 2 OCT devices and to explore factors that may influence the difference in measurements.DesignCross-sectional study.ParticipantsSixty-nine glaucoma eyes (63 patients) with evidence of central damage or mean deviation (MD) of -6.0 dB or worse on a 24-2 visual field (VF).MethodsCirrus and Spectralis OCT macular volume scans were exported, data from the central 20° of both OCT devices were centered and aligned, and 50 × 50 arrays of 0.4° × 0.4° superpixels were created. We estimated nonparametric (Spearman's) correlations and used Bland-Altman plots to compare GCIPL thickness measurements between the two OCTs at the superpixel level. Factors that may have influenced the differences between thickness measurements between the two devices were explored with linear mixed models.Main outcome measuresPooled and individual-eye Spearman's correlation and agreement between thickness measurements from the two devices.ResultsThe median 24-2 VF MD was -6.8 dB (interquartile range [IQR], -4.9 to -12.3 dB). The overall pooled Spearman's correlation between the two devices for all superpixels and eyes was 0.97 (P < 0.001). The median within-eye correlation coefficient was 0.72 (IQR, 0.59-0.79). Bland-Altman plots demonstrated a systematic bias in most individual eyes, with Spectralis GCIPL measurements becoming larger than Cirrus measurements with increasing superpixel thickness. The average superpixel thickness and distance to the fovea influenced the thickness difference between the two devices in multivariate models (P < 0.001).ConclusionsLocal macular thickness measurements from the Spectralis and Cirrus devices are highly correlated, but not interchangeable. Differences in thickness measurements between the two devices are influenced by the location of superpixels and their thickness.

Hyperspectral Images, which are high-dimensional in nature and capture bands over hundreds of wavelengths of the electromagnetic spectrum. These images have piqued researchers' curiosity in the last two decades. The purpose of this paper is to investigate how researchers segmented and classified Hyperspectral Images with unbalanced data and few labelled training examples. For the sake of comprehension, the background of Hyperspectral Images and segmentation techniques is briefly discussed at first. The study is organised around different Hyperspectral Image processing techniques such as thresholding, clustering, watershed, deep learning, and other methods. The recent trends and developments in HSI segmentation have been reviewed and compiled using benchmark datasets such as Indian Pines, Salinas Valley, Pavia University, and others. Finally, it is intended that the readers will gain a thorough understanding of existing segmentation techniques, their performance, and fresh research areas for HSI that need to be studied or explored. [ABSTRACT FROM AUTHOR]

In the present scenario, Alzheimer's Disease (AD) is one of the incurable neuro-degenerative disorders, which accounts for nearly 60% to 70% of dementia cases. Currently, several machine-learning approaches and neuroimaging modalities are utilized for diagnosing AD. Among the available neuroimaging modalities, functional Magnetic Resonance Imaging (fMRI) is extensively utilized for studying brain activities related to AD. However, analyzing complex brain structures in fMRI is a time-consuming and complex task; so, a novel automated model was proposed in this manuscript for early diagnosis of AD using fMRI images. Initially, the fMRI images are acquired from an online dataset: Alzheimer's Disease Neuroimaging Initiative (ADNI). Further, the quality of the acquired fMRI images was improved by implementing a normalization technique. Then, the Segmentation by Aggregating Superpixels (SAS) method was implemented for segmenting the brain regions (AD, Normal Controls (NC), Mild Cognitive Impairment (MCI), Early Mild Cognitive Impairment (EMCI), Late Mild Cognitive Impairment (LMCI), and Significant Memory Concern (SMC)) from the denoised fMRI images. From the segmented brain regions, feature vectors were extracted by employing Gabor and Gray Level Co-Occurrence Matrix (GLCM) techniques. The obtained feature vectors were dimensionally reduced by implementing Honey Badger Optimization Algorithm (HBOA) and fed to the Multi-Layer Perceptron (MLP) model for classifying the fMRI images as AD, NC, MCI, EMCI, LMCI, and SMC. The extensive investigation indicated that the presented model attained 99.44% of classification accuracy, 88.90% of Dice Similarity Coefficient (DSC), 90.82% of Jaccard Coefficient (JC), and 88.43% of Hausdorff Distance (HD). The attained results are better compared with the conventional segmentation and classification models. [ABSTRACT FROM AUTHOR]

Abstract The use of uncrewed aerial vehicle to map the environment increased significantly in the last decade enabling a finer assessment of the land cover. However, creating accurate maps of the environment is still a complex and costly task. Deep learning (DL) is a new generation of artificial neural network research that, combined with remote sensing techniques, allows a refined understanding of our environment and can help to solve challenging land cover mapping issues. This research focuses on the vegetation segmentation of kettle holes. Kettle holes are small, pond‐like, depressional wetlands. Quantifying the vegetation present in this environment is essential to assess the biodiversity and the health of the ecosystem. A machine learning workflow has been developed, integrating a superpixel segmentation algorithm to build a robust dataset, which is followed by a set of DL architectures to classify 10 plant classes present in kettle holes. The best architecture for this task was Xception, which achieved an average F1‐score of 85% in the segmentation of the species. The application of solely 318 samples per class enabled a successful mapping in the complex wetland environment, indicating an important direction for future health assessments in such landscapes.

Although deep learning techniques are known to achieve outstanding classification accuracies, remote sensing datasets often present limited labeled data and class imbalances, two challenges to attaining high levels of accuracy. In recent years, the GAN architecture has achieved great success as a data augmentation method, driving research toward further enhancements. This work presents ResBaGAN, a GAN-based method for the classification of remote sensing images, designed to overcome the challenges of data scarcity and class imbalances by constructing an advanced data augmentation framework. This framework builds upon a GAN architecture enhanced with an autoencoder initialization and class balancing properties, a superpixel-based sample extraction procedure with traditional augmentation techniques, and an improved residual network as classifier. Experiments were conducted on large, very high-resolution multispectral images of riparian forests in Galicia, Spain, with limited training data and strong class imbalances, comparing ResBaGAN to other machine learning methods such as simpler GANs. ResBaGAN achieved higher overall classification accuracies, particularly improving the accuracy of minority classes with F1-score enhancements reaching up to 22%.

Although supervised classification of hyperspectral images (HSI) has achieved success in remote sensing, its applications in real scenarios are often constrained, mainly due to the insufficiently available or lack of labelled data. As a result, unsupervised HSI classification based on data clustering is highly desired, yet it generally suffers from high computational cost and low classification accuracy, especially in large datasets. To tackle these challenges, a novel unsupervised spatial-spectral HSI classification method is proposed. By combining the entropy rate superpixel segmentation (ERS), superpixel-based principal component analysis (PCA), and PCA-domain 2D singular spectral analysis (SSA), both the efficacy and efficiency of feature extraction are improved, followed by the anchor-based graph clustering (AGC) for effective classification. Experiments on three publicly available and five self-collected aerial HSI datasets have fully demonstrated the efficacy of the proposed PCA-domain superpixelwise SSA (PSSA) method, with a gain of 15–20% in terms of the overall accuracy, in comparison to a few state-of-the-art methods. In addition, as an extra outcome, the HSI dataset we acquired is provided freely online. [ABSTRACT FROM AUTHOR]

Mineral segmentation is an equally important and difficult task in the quantification of mineral composition. Difficulties come from the process of determining boundaries of distinctive mineral grains necessary for further analysis and mineral identification. Done by hand, the task is very time-consuming and higher accuracies are burdened with the possible human fatigue factor. The presented method is a fully automated solution to the problem that uses a superpixel approach and feature-based merging. The method is validated by comparison with the manual approach. Analyzed data consist of photos taken by a Nikon Eclipse LV100N POL polarizing microscope at 200× magnification, in transmitted light, with crossed polarizers. Images are first prepared by Gaussian filter and meanshift operations, then the initial segmentation is provided by the superpixel algorithm. Oversegmentation is resolved by feature-based merging. The last step consists of counting the individual grain boundaries and preparing the results as easily readable visual data. [ABSTRACT FROM AUTHOR]

The use of uncrewed aerial vehicle to map the environment increased significantly in the last decade enabling a finer assessment of the land cover. However, creating accurate maps of the environment is still a complex and costly task. Deep learning (DL) is a new generation of artificial neural network research that, combined with remote sensing techniques, allows a refined understanding of our environment and can help to solve challenging land cover mapping issues. This research focuses on the vegetation segmentation of kettle holes. Kettle holes are small, pond‐like, depressional wetlands. Quantifying the vegetation present in this environment is essential to assess the biodiversity and the health of the ecosystem. A machine learning workflow has been developed, integrating a superpixel segmentation algorithm to build a robust dataset, which is followed by a set of DL architectures to classify 10 plant classes present in kettle holes. The best architecture for this task was Xception, which achieved an average F1‐score of 85% in the segmentation of the species. The application of solely 318 samples per class enabled a successful mapping in the complex wetland environment, indicating an important direction for future health assessments in such landscapes. [ABSTRACT FROM AUTHOR]

Depth estimation of pixels having low intensity gradient is a difficult task. Since they contain very less information with its neighborhood (i.e., similar in intensity due to lack of texture), their tracking for depth becomes difficult in consecutive temporal images. In recent times, dense-mapping methods based on mid-level features such as Superpixels or planes are being proposed to estimate depth of the low-gradient pixels. These methods have advantage of less computation resource requirements, with the possibility of computing full Monocular-Simultanuous Localization And Mapping (SLAM) processing on a CPU, i.e., these do not require GPU. In the sperpixel based existing approaches, the superpixels are formed by extracting similar intensity pixels and their depth map is estimated from their high gradient border pixels. Whereas the drawbacks are higher computational time is required for the superpixel segmentation, semi-dense mapping and planar mapping. This paper proposes a method to estimate the depth map of planar regions using binary images instead of full intensity range images (gray/color image) and a novel solution to differentiate differently oriented planes grouped as a single superpixel due to similarity in the intensities. The computational time is reduced while improving the efficiency of the segmentation, which is demonstrated in real-time experiments. Efficacy of the proposed method is compared with recent techniques based on learning approach and model-based approach. The proposed method is compared with the existing methods using publicly available datasets. The results show, higher density and lesser computational time for the depth map estimation in the proposed method when compared with other model-based approaches. When compared with learning-based approaches, the proposed method shows better accuracy in the depth estimation. [ABSTRACT FROM AUTHOR]

Nowadays, the application of Fiber Reinforced Polymer (FRP) materials in concrete structures is more and more extended because of their advantages. However, premature and brittle failure caused by debonding of FRP materials from the concrete surface may be critical and needs to be detected at its earliest stages to avoid possible catastrophic events. Digital image correlation (DIC) has been used in the past for the detection of abnormalities from surface digital images. In this work, an efficient computer-aided diagnosis (CAD) system for damage identification for this type of FRP strengthening is developed. It is based on the combination of cells of strain maps built with a DIC system and an automated clustering classifier optimized with the use of superpixels and principal component analysis. Additionally, preprocessing of the images is also done with some filters to enhance the accuracy of the methodology. The procedure has been validated with experimental results on a progressively damaged FRP externally strengthened beam. Results show that the proposed detection scheme achieves an effective automated classification method for the damage detection of this kind of strengthened specimens. Furthermore, the optimal choice of the subset size, number of superpixels and principal components contributed to the success of the method. Simultaneously, an attempt to quantify damage from the proposed method and its validation with electromechanical impedance method has been developed. • A computer vision system for automated assessment and classification of damage in FRP strengthened RC beams is proposed. • The proposed methodology combines superpixels, principal component analysis and k-means clustering. • Superpixels and principal component analysis contribute to reduce the dimensions and complexity of the DIC images. • A relative damage quantification has been proposed from the superpixel segmented patterns. • The method achieves a very high accuracy compared to other detection schemes. [ABSTRACT FROM AUTHOR]

Automatic and accurate segmentation of lesions from high-resolution computed tomography (HRCT) images play a critical role in diagnosis of interstitial lung disease (ILD). Fully supervised methods require considerable expertise annotations, resulting in time-consuming and labor-intensive cost. Contrastively, weakly supervised segmentation (WSS) algorithms could alleviate greatly this issue of medical image annotation. Multiple-instance learning (MIL), a class of weakly supervised learning (WSL), has proven to be effective in image segmentation tasks. However, most existing MIL methods based on patches or pixels lack meaningful edge information of the region of interests (ROIs) and leave out of considering the dependence relations of inter-instances. In this study, we propose a novel Transformer based multiple superpixel-instance learning (MSIL) for pixel-level lesions segmentation of ILD. This method introduces superpixels into a MIL framework, which solves the problem of inconsistency in lesion boundary segmentation caused by using patches as instances. In addition, the proposed method introduces Transformer into the MIL framework to capture global or long-range dependencies and establish the relationship between superpixel instances based on the multi-head self-attention, which solves the shortcoming that instances are independent of each other in MIL. Experimental results on private HRCT images confirm that our proposed method can achieve state-of-the-art (SOTA) performance compared to other WSS algorithms, MIL-based WSS algorithms, CAMs-based classification algorithms (CA) and fully supervised segmentation algorithms (FSS). • We proposed a novel weakly supervised segmentation (WSS) algorithm (ViT-MSIL). • We explored the combination of Transformer, superpixels and MIL for WSS on HRCT images. • Our experimental results show that ViT-MSIL achieves state-of-the-art results. • The ViT-MSIL segments the lesion edge more accurately than the patch-based method. [ABSTRACT FROM AUTHOR]

Recently, convolutional neural networks (CNNs) have been successfully utilized in polarimetric synthetic aperture radar (PolSAR) image classification and obtained promising results. However, most CNN-based classification methods require a large number of labeled samples and it is difficult to obtain sufficient labeled samples. For this reason, an ensemble dual-branch CNN (EDb-CNN) is proposed for PolSAR image classification with small samples. First, to solve the problem of the small sample in PolSAR image classification, a new data enhancement method based on the superpixel algorithm is proposed to expand the number of labeled samples. Second, to obtain different scales of features from PolSAR images, a Db-CNN model is proposed. This model contains two parallel CNN structures. One CNN branch is used to extract the polarization features from the complex coherency matrix. The other branch is utilized to extract the spatial features based on weighted spatial neighborhood. On the top of these two branches, a feature fusion model is adopted to combine these two deep features, and a weighted loss function is employed to improve the learning procedure. Then, the ensemble learning algorithm is used for each CNN branch and Db-CNN network to obtain the better classification results. Finally, a postprocess algorithm based on the superpixel algorithm is proposed to improve the consistency of classification results. Experiments on two PolSAR datasets show that the proposed method achieves a much better performance than other classification methods, especially when only a few labeled samples are available.

Wetland is of significant ecological value, which is very important and challenging for large-scale mapping. Sentinel-1 can continuously record wetland changes with its all-day, all-weather working capability. How to fully utilize multidimensional information, such as time and statistical texture of synthetic aperture radar (SAR) image, to classify wetlands more accurately has become a research focus. Thus, this article constructed time series similarity parameters to describe the temporal change information of targets in wetland and introduced G0 statistical texture parameters to describe texture characteristics of SAR images. Combining multifeatures and superpixels, we proposed a classification method based on random forest (RF) to map Dongting Lake wetland. The overall classification accuracy was 97.57%, and the Kappa coefficient was 0.97. The classification accuracy of reed beach and grass marsh was above 95%. The results showed the proposed SAR time series similarity features could effectively utilize dynamical information among classes and was helpful to identify mudflat, grass marsh, and reed beach with high dynamics in wetland. The introduced statistical texture features expressed the heterogeneity of targets, and enhanced the recognition and extraction ability of forest beach, mudflat, and reed beach. Compared with support vector machine, decision tree, and RF classification methods, RF with superpixels optimized not only got high precision but also could effectively reduce the pepper-salt error in classification, because of the consideration of superpixels context information.

With the increased reliance on medical imaging, Deep convolutional neural networks (CNNs) have become an essential tool in the medical imaging-based computer-aided diagnostic pipelines. However, training accurate and reliable classification models often require large fine-grained annotated datasets. To alleviate this, weakly-supervised methods can be used to obtain local information such as region of interest from global labels. This work proposes a weakly-supervised pipeline to extract Relevance Maps of medical images from pre-trained 3D classification models using localized perturbations. The extracted Relevance Map describes a given region’s importance to the classification model and produces the segmentation for the region. Furthermore, we propose a novel optimal perturbation generation method that exploits 3D superpixels to find the most relevant area for a given classification using U-net architecture. This model is trained with perturbation loss, which maximizes the difference between unperturbed and perturbed predictions. We validated the effectiveness of our methodology by applying it to the segmentation of Glioma brain tumours in MRI scans using only classification labels for glioma type. The proposed method outperforms existing methods in both Dice Similarity Coefficient for segmentation and resolution for visualizations.

Reality capture technologies such as Structure-from-Motion (SfM) photogrammetry have become a state-of-the-art practice within landslide research workflows in recent years. Such technology has been predominantly utilized to provide detailed digital products in landslide assessment where often, for thorough mapping, significant accessibility restrictions must be overcome. UAV photogrammetry produces a set of multi-dimensional digital models to support landslide management, including orthomosaic, digital surface model (DSM), and 3D point cloud. At the same time, the recognition of objects depicted in images has become increasingly possible with the development of various methodologies. Among those, Geographic Object-Based Image Analysis (GEOBIA) has been established as a new paradigm in the geospatial data domain and has also recently found applications in landslide research. However, most of the landslide-related GEOBIA applications focus on large scales based on satellite imagery. In this work, we examine the potential of different UAV photogrammetry product combinations to be used as inputs to image segmentation techniques for the automated extraction of landslide elements at site-specific scales. Image segmentation is the core process within GEOBIA workflows. The objective of this work is to investigate the incorporation of fully 3D data into GEOBIA workflows for the delineation of landslide elements that are often challenging to be identified within typical rasterized models due to the steepness of the terrain. Here, we apply a common unsupervised image segmentation pipeline to 3D grids based on the superpixel/supervoxel and graph cut algorithms. The products of UAV photogrammetry for two landslide cases in Greece are combined and used as 2D (orthomosaic), 2.5D (orthomosaic + DSM), and 3D (point cloud) terrain representations in this research. We provide a detailed quantitative comparative analysis of the different models based on expert-based annotations of the landscapes and conclude that using fully 3D terrain representations as inputs to segmentation algorithms provides consistently better landslide segments. [ABSTRACT FROM AUTHOR]