Search

Object detection has become a very prominent subject for research in recent times. This study's main goal is to suggest a technique for video saliency object detection. It seems to sense that using the depth information in photos to detect salient things. Since depth offers abundant information about scene structure, object forms, and other 3D cues. This information is very compatible to distinguish between objects in the foreground and background. As a result of the high object density, small object size, and cluttered background, aerial photos and movies provide results with low precision. In this paper, the proposed SPTM (Super Pixel Transmission Map)-YOLO model, the input RGB image has applied Dark Channel Prior (DCP) method for estimating the transmission map. From the transmission map only, the background probability is estimated with the help of SLIC (simple linear iterative clustering algorithm) superpixel segmentation. That foreground extracted image is further learned with YOLO architecture to detect the objects effectively. For object detection in aerial images, this proposed SPTM-YOLO approach outperforms classic YOLO by up to 6% accuracy. Accurate detection of things that are small in size, partially occluded, and out of view is possible. [ABSTRACT FROM AUTHOR]

Recently, deep learning methods have been widely studied in the field of polarimetric synthetic aperture radar (PolSAR) ship detection. However, extracting polarimetric and spatial features on the whole PolSAR image will result in high computational complexity. In addition, in the massive data ship detection task, the image to be detected contains a large number of invalid areas, such as land and seawater without ships. Therefore, using ship coarse detection methods to quickly locate the potential areas of ships, that is, ship potential area extraction, is an important prerequisite for PolSAR ship detection. Since existing unsupervised PolSAR ship detection methods based on pixel-level features often rely on fine sea–land segmentation pre-processing and have poor applicability to images with complex backgrounds, in order to solve the abovementioned issue, this paper proposes a PolSAR ship potential area extraction method based on the neighborhood semantic differences of an LDA bag-of-words topic model. Specifically, a polarimetric feature suitable for the scattering diversity condition is selected, and a polarimetric feature map is constructed; the superpixel segmentation method is used to generate the bag of words on the feature map, and latent high-level semantic features are extracted and classified with the improved LDA bag-of-words topic model method to obtain the PolSAR ship potential area extraction result, i.e., the PolSAR ship coarse detection result. The experimental results on the self-established PolSAR dataset validate the effectiveness and demonstrate the superiority of our method. [ABSTRACT FROM AUTHOR]

Superpixel segmentation can get the middle features in image processing, effectively reduce the dimensionality of the image, and is widely used in image processing fields. To get the regular and compact superpixels in real-time, a superpixel segmentation algorithm based on image density is proposed in this paper. Firstly, the image is uniformly divided according to the number of superpixels to be obtained. Secondly, to get the clustering ability of the pixels, the density image is produced. Thirdly, the seed is chosen in each sub-region block according to the density and then the superpixels are obtained by clustering. During the clustering process, the pixel around the seed should be added into the superpixel if it meets the conditions, and the small supeipixels are merged into the big superpixels around them. Finally, the result shows that the proposed algorithm has the best segmentation effect, and a good balance in accuracy, regularity, and time cost. [ABSTRACT FROM AUTHOR]

This paper proposes a weakly-supervised structural surface crack detection algorithm that can detect the crack area in an image with low data labeling cost. The algorithm consists of a convolutional neural networks Vgg16-Crack for classification, an improved and optimized class activation map (CAM) algorithm for accurately reflecting the position and distribution of cracks in the image, and a method that combines superpixel segmentation algorithm simple linear iterative clustering (SLIC) with CAM for more accurate semantic segmentation of cracks. In addition, this paper uses Bayesian optimization algorithm to obtain the optimal parameter combination that maximizes the performance of the model. The test results show that the algorithm only requires image-level labeling, which can effectively reduce the labor and material consumption brought by pixel-level labeling while ensuring accuracy. [ABSTRACT FROM AUTHOR]

The density peak clustering (DPC) is one of the most popular algorithms for segmenting images due to its simplicity and efficiency. Since DPC and its variants are not specifically designed for image segmentation, their segmentation results do not necessarily meet both subjective and objective metrics. We propose an expanded relative density-based clustering algorithm as a solution to the above problems, which can automatically determine the cluster number and make the image segmentation results more consistent with subjective criteria. First, the image is pre-segmented into superpixels using the simple linear iterative clustering algorithm, and the superpixels are represented by feature vectors containing color and texture information. Secondly, the expanded relative density of the data point is obtained by comparing the tightness of a mini-cluster with its neighboring mini-clusters. The Sigmoid function is then applied to the data point with small density but large relative distance to further adjust its relative distance so that the distribution of cluster centers matches the characteristics of the image. Next, the optimal cluster number is determined by the rate of change of the sum of squared errors. Finally, the cluster center pairs with smaller distances are merged using the cluster center merging algorithm. The experiments conducted on synthetic and real datasets demonstrate that the performance of the proposed algorithm outperforms the compared algorithms. [ABSTRACT FROM AUTHOR]

The presence of haze significantly degrades the quality of remote sensing images, resulting in issues such as color distortion, reduced contrast, loss of texture, and blurred image edges, which can ultimately lead to the failure of remote sensing application systems. In this paper, we propose a superpixel-based visible remote sensing image dehazing algorithm, namely SRD. To begin, the remote sensing haze images are divided into content-aware patches using superpixels, which cluster adjacent pixels considering their similarities in color and brightness. We assume that each superpixel region shares the same atmospheric light and transmission properties. Subsequently, methods to estimate local atmospheric light and transmission within each superpixel are proposed. Unlike existing dehazing algorithms that assume a globally constant atmospheric light, our approach considers the global heterogeneous distribution of the atmospheric ambient light, which allows us to model it as a global non-uniform variable. Furthermore, we introduce an effective atmospheric light estimation method inspired by the maximum reflectance prior. Moreover, recognizing the wavelength-dependent nature of light transmission, we independently estimate the transmittance for each RGB channel of the input image. The quantitative and qualitative evaluation results of comprehensive experiments on synthetic datasets and real-world samples demonstrate the superior performance of the proposed algorithm compared to state-of-the-art methods for remote sensing image dehazing. [ABSTRACT FROM AUTHOR]

As a popular research problem in computational aesthetics, image aesthetic assessment has many important applications in image editing, retrieval, and recommendation. However, the existing mainstream CNN-based image aesthetic assessment methods are difficult to obtain the global aesthetic attributes of images well. To this end, we propose a two-stream image aesthetic assessment model that couples Transformer and CNN features. We use the traditional CNN network to extract the image's local aesthetic feature in the first stream, apply the superpixel algorithm to segment the image, and then feed the segmented image region into the Transformer network to learn the image's aesthetic global features in the second stream. Finally, the features learned by Transformer and CNN are fused to achieve the image aesthetic assessment. The experimental results on the AVA dataset show that our proposed method can obtain local and global aesthetic information on images, which enables the model to learn richer aesthetic information, and the combination of whole and part is more in line with human aesthetic characteristics. Our proposed model achieves an accuracy of 84.5% in the classification task, achieving optimal performance compared to existing methods and good performance in the other two tasks (Score Regression and Distribution). [ABSTRACT FROM AUTHOR]

On the verge of superpixel methods exploiting saliency information, the Superpixels through Iterative CLEarcutting (SICLE) framework has reported fast and accurate superpixel delineation. It is composed of three steps: (i) seed oversampling; (ii) superpixel generation; and (iii) seed removal. It starts from (i) and applies several iterations of (ii) and (iii) until reaching the desired superpixel quantity. In this work, we improve SICLE such that it can now generate compact superpixels with accurate delineation. We exploit differential computation and propose several novel functions for steps (ii) and (iii) for proper saliency incorporation, compact superpixel generation, and improvement in speed and delineation. Results show that, with our proposals, SICLE achieves state-of-the-art performance in delineation and speed whenever saliency is absent with on-par compacity. When an accurate saliency map is provided, its performance improves significantly and requires only two iterations for segmentation. [ABSTRACT FROM AUTHOR]

Superpixel become increasingly popular in image segmentation field as it greatly helps image segmentation techniques to segment the region of interest accurately in noisy environment and also reduces the computation effort to a great extent. However, selection of proper superpixel generation techniques and superpixel image segmentation techniques play a very crucial role in the domain of different kinds of image segmentation. Clustering is a well-accepted image segmentation technique and proved their effective performance over various image segmentation field. Therefore, this study presents an up-to-date survey on the employment of superpixel image in combined with clustering techniques for the various image segmentation. The contribution of the survey has four parts namely (i) overview of superpixel image generation techniques, (ii) clustering techniques especially efficient partitional clustering techniques, their issues and overcoming strategies, (iii) Review of superpixel combined with clustering strategies exist in literature for various image segmentation, (iv) lastly, the comparative study among superpixel combined with partitional clustering techniques has been performed over oral pathology and leaf images to find out the efficacy of the combination of superpixel and partitional clustering approaches. Our evaluations and observation provide in-depth understanding of several superpixel generation strategies and how they apply to the partitional clustering method. [ABSTRACT FROM AUTHOR]

Evaluating and tracking the size of a wound is a crucial step in wound assessment. The measurement of various indicators on wounds over time plays a vital role in treating and managing crucial wounds. This article introduces the concept of utilizing mobile device-captured photographs to address this challenge. The research explores the application of digital technologies in the treatment of chronic wounds, offering tools to assist healthcare professionals in enhancing patient care and decision-making. Additionally, it investigates the use of deep learning (DL) algorithms along with the use of computer vision techniques to enhance the validation results of wounds. The proposed method involves tissue classification as well as visual recognition system. The wound's region of interest (RoI) is determined using superpixel techniques, enabling the calculation of its wounded zone. A classification model based on the Region Anchored CNN framework is employed to detect and differentiate wounds and classify their tissues. The outcome demonstrates that the suggested method of DL, with visual methodologies to detect the shape of a wound and measure its size, achieves exceptional results. By utilizing Resnet50, an accuracy of 0.85 percent is obtained, while the Tissue Classification CNN exhibits a Median Deviation Error of 2.91 and a precision range of 0.96%. These outcomes highlight the effectiveness of the methodology in real-world scenarios and its potential to enhance therapeutic treatments for patients with chronic wounds. [ABSTRACT FROM AUTHOR]

The dimension reduction (DR) technique plays an important role in hyperspectral image (HSI) processing. Among various DR methods, superpixel-based approaches offer flexibility in capturing spectral–spatial information and have shown great potential in HSI tasks. The superpixel-based methods divide the samples into groups and apply the DR technique to the small groups. Nevertheless, we find these methods would increase the intra-class disparity by neglecting the fact the samples from the same class may reside on different superpixels, resulting in performance decay. To address this problem, a novel unsupervised DR named the Collaborative superpixelwise Auto-Encoder (ColAE) is proposed in this paper. The ColAE begins by segmenting the HSI into different homogeneous regions using a superpixel-based method. Then, a set of Auto-Encoders (AEs) is applied to the samples within each superpixel. To reduce the intra-class disparity, a manifold loss is introduced to restrict the samples from the same class, even if located in different superpixels, to have similar representations in the code space. In this way, the compact and discriminative spectral–spatial feature is obtained. Experimental results on three HSI data sets demonstrate the promising performance of ColAE compared to existing state-of-the-art methods. [ABSTRACT FROM AUTHOR]

In modern society, digital images can be far-reaching, and the images are manipulated by various software and hardware technologies. The image forgery activities are undertaken by the attackers mainly for damaging the reputation of people or receiving fiscal gain, etc. Taking this into consideration, many techniques are developed to detect the forged images. In this paper, a new deep learning-based approach is introduced for copy-move forgery detection. The input images are segmented into non-overlapping patches using superpixel-based modified dense peak clustering and the features are extracted from the segmented patches by applying deep learning structure of attention-based DenseNet 121 model. Besides, to compare every block, the depth of each pixel is reconstructed, and eventually matching process is carried out using the adaptive chimp patch matching approach, which detects the suspicious forged regions in an image. Finally, the matched keypoints are merged with the segmented patches using the merged keypoint matching algorithm. As a result, the new deep learning approach has detected the forged regions efficiently from the tampered image with less time even the image is compressed, rotated, or scaled. The performance is evaluated in terms of recall, precision, accuracy, F1-score, computational time, and False Positive Rate (FPR). Moreover, the performance is compared with the other existing approaches, and the outcomes showed that the proposed method has achieved higher accuracy of 97%, recall of 99%, precision of 97.84%, F1-score of 98.81%, FPR of 0.022 and less computational time of 2.5 s. [ABSTRACT FROM AUTHOR]

Moving object detection is still a challenging task in complex scenes. The existing methods based on deep learning mainly use U-Nets and have achieved amazing results. However, they ignore the local continuity between pixels. In order to solve this problem, a method based on a superpixel fusion network (SF-Net) is proposed in this article. First, the median filter is used to extract the candidate foreground (called pixel features) and the image sequence is segmented by superpixel. Then, the histogram features (called superpixel features) of the candidate foreground superpixels are extracted. Next, the pixel features and the superpixel features are the inputs of SF-Net, respectively. Experiments show the effectiveness of SF-Net on 34 image sequences and the average F-measure reaches 0.84. SF-Net can remove more background noise and has stronger expression ability than a network with the same depth. [ABSTRACT FROM AUTHOR]

Aiming at the problems of misjudgment of focus region, unclear textures and artifacts in multi-focus image fusion, a non-subsampled shearlet transform (NSST) fusion method combining superpixels and guided filter is proposed. Firstly, in the high-frequency fusion, a three-fold guided filter is used to identify the focal region and smooth the focal boundary, and the noise immunity of the image is also enhanced by median denoising. Secondly, gamma correction is performed on the low-frequency subbands based on superpixel to improve the contrast of the image and ensure the continuity of textures. Super-pixel fusion can also maintain the consistency of regional chroma. Finally, the fusion result is composed of brightness and chromaticity subgraphs by inverse Lab transformation. In seven groups of experiments, the spatial frequency (SF), standard deviation (SD), edge intensity (EI) and visual information fidelity(VIFF) values of the proposed algorithm are much higher than those of the contrast algorithms, indicating that the proposed algorithm can preserve the information of textures and contours well, and improve the contrast and sharpness of images. [ABSTRACT FROM AUTHOR]

Data augmentation effectively alleviates the over-fitting problem in convolutional neural network-based (CNN-based) models, especially in the limited dataset. However, the inconsistency problem between the augmented sample and its original label is still a critical challenge during the augmentation operation. In this paper, we propose a novel data augmentation scheme named Interpretability-Mask (IM), which exploits the interpretability of the classifier to obtain the most discriminative regions and preserve label invariance. Concretely, we first construct a set-based representation for a sample and its label by superpixel segmentation and the local interpretable model-agnostic explanations (LIME) operator. Secondly, the sample represented by the superpixel set is utilized to synthesize the region-level disturbance augmentation sample with a random removal strategy. Then, the label constructed by the most interpretive superpixel set is applied to maintain the consistency between the augmented sample and its original label. Lastly, the augmentation scheme will be randomly used to the training stage. Extensive experiments are conducted on challenging datasets. A significant improvement in classification performance has achieved with the IM scheme. On the CIFAR-10 dataset, the Top-1 error rate drops by 2.15% at most. On the CIFAR-100 dataset, the Top-1 error rate decreases by up to 3.69%. And the maximum decline of the Top-1 error rate is 3.35% on the Mini-ImageNet. Experimental results manifest the effectiveness and generality of the proposed method. [ABSTRACT FROM AUTHOR]

Synthetic Aperture Radar (SAR) image ship target detection has attracted considerable attention. As a state-of-the-art method, the Constant False Alarm Rate (CFAR) detection algorithm is often used in SAR image ship target detection. However, the detection performance of the classical CFAR is easily affected by speckle noise. Moreover, the detection results based on the sliding window are sensitive to the size of the sliding window. Thus, ensuring that there are no target pixels in the cluttered background is difficult, which easily leads to a high computational load. This study proposes a new ship target detection method for SAR images based on fast superpixel-based non-window CFAR to solve these problems. The superpixel generation method of Density Based Spatial Clustering of Applications with Noise is used to generate superpixels for SAR images. Under the assumption that SAR data obey the Rayleigh mixture distribution, we define a superpixel dissimilarity measure. Then, the clutter parameters of each pixel are accurately estimated using superpixels, which can avoid the shortcomings of the traditional CFAR sliding window even in the case of multiple targets. A local contrast based on the Coefficient of Variation (CoV) of the SAR image is proposed to optimize the CFAR detection result, which can eliminate a large number of false alarms from man-made targets in urban areas. The experimental results of five real SAR images show that the proposed method for ship target detection in SAR images with different scenes is robust compared with other state-of-the-art methods.

Land use recognition from multispectral satellite images is fundamentally critical for geological applications, but the results are not satisfied. The scale dimension of current multiscale learning is too coarse to account for rich scales in multispectral images, and pixel-wise classification tends to produce “salt-and-pepper” labels due to possible misclassification in heterogeneous regions. In this article, these issues are addressed by proposing a new pixel-wise classification model with finer scales for convolutional neural networks. The model is designed to extract multiscale contextual information using multiscale networks at a fine-grained level, addressing the issue of insufficient multiscale learning for classification. Furthermore, a small-scale segmentation-combination method is introduced as a postprocessing solution to smooth fragmented classification results. The proposed method is tested on GF-1, GF-2, DEIMOS-2, GeoEye-1, and Sentinel-2 satellite images, and compared with six neural-network-based algorithms. The results demonstrate the effectiveness of the proposed model in finding objects of large scale difference, improving classification accuracy, and reducing classified fragments. The discussion also illustrates that convolutional neural networks and pixel-wise inference are more practical than transformer and patch-wise recognition.

To enhance spectral unmixing performance, a large number of algorithms have simultaneously investigated spatial and spectral information in hyperspectral images. However, sophisticated algorithms with high computational complexity can be very time-consuming when a large amount of data are involved in processing hyperspectral images. In this article, we first introduce a group sparse graph-regularized unmixing method with superpixel structure, to promote piecewise consistency of abundances and reduce computational burden. Segmenting the image into several nonoverlapped superpixels also enables to decompose the unmixing problem into uncoupled subproblems that can be processed in parallel. An implementation for the proposed algorithm on graphics processing units (GPUs) is then developed based on the NVIDIA compute unified device architecture (CUDA) framework. The proposed scheme achieves parallelism at both the intrasuperpixel and intersuperpixel levels, where multiple concurrent streams have been used to enable multiple kernels to execute on the device simultaneously. Simulation results with a series of experiments demonstrate advantages of the proposed algorithm. The performance of the GPU implementation also illustrates that parallel scheme largely expedites the implementation.

Each of the three satellites constituting the RADARSAT Constellation Mission (RCM) provides compact polarimetric synthetic aperture radar (CP SAR) data. The complex CP data have similar properties to the complex quad polarimetric (QP) data provided by prior RADARSAT missions. In this article, a land cover classification method using spatial information is designed based on the statistical characteristics of the complex CP and QP SAR data. First, the local spatial dependency among pixels is captured by superpixels. Second, a graph is constructed on the superpixels to model the global spatial dependency among superpixels. The land cover classification image with land cover type labels is then estimated by propagating labels from the few labeled superpixels to the unlabeled superpixels. Classification of two RCM complex CP and QP scenes demonstrates that the proposed method, with few labeled pixels, provides much higher classification accuracy than methods that do not exploit global spatial dependency.

Object tracking is a crucial research area within the field of intelligent transportation, providing a vital foundation for anomalous behavior analysis and traffic statistics. Although pedestrian detectors have shown impressive results, leading to the advancement of detection-based tracking methods, target association in complex scenarios remains a difficult and less efficient task due to the lack of feature robustness in the presence of partial occlusions. In the proposed tracking method, we extract convolutional features on each entire object and its local blocks, segmented by the superpixel algorithm. Aiming to emphasize the global and local information respectively, the global features for each entire object are extracted from the last layer of the backbone network, while local features are derived from a specific intermediate layer of the backbone network. The association between tracked targets and detected pedestrian candidates relies on fused similarity degrees. Furthermore, we use the transformer’s self-attention mechanism to predict features for the current frame based on the information within past frames, aiming to eliminate the effects of target appearance variations. Additionally, we remove redundant background pixels in the detected rectangles of pedestrian candidates by using a background modeling algorithm. Experimental results demonstrate that the tracker proposed in this paper outperforms other trackers across five publicly available datasets, indicating its effectiveness and potential for further development.

The accurate and reliable derivation of the pap smear cell, which contains cytoplasm and nucleus regions, depends on the segmentation process employed in the cervical cancer detection mechanism. In this paper, an Improved Boykov’s Graph Cut-based Conditional Random Fields and Superpixel imposed Semantic Segmentation Technique (IBGC-CRF-SPSST) is proposed for efficient cervical cancer detection. This proposed IBGC-CRF-SPSST embeds the complete benefits of constraint association among pixels and superpixel edge data for accurate determination of the nuclei and cytoplasmic boundaries so as to ensure efficient differentiation of the healthy and unhealthy cancer cells. Finally, the pixel-level forecasting potential of Conditional Random Fields is included for enhancing the degree of semantic-based segmentation accuracy to a predominant level. The experimental evaluated results of the proposed IBGC-CRF-SPSST aim to produce an accuracy of 99.78%, a mean processing time of 2.18sec, a precision of 96%, a sensitivity of 98.92%, and a specificity of 99.32% value which is determined to be excellent and on par with the existing detection techniques used for investigating cervical cancer.

Forests are the most important carbon reservoirs on land, and forest carbon sinks can effectively reduce atmospheric CO 2 concentrations and mitigate climate change. In recent years, various satellites have been launched that provide opportunities for identifying forest types with low cost and high time efficiency. Using multi-temporal remote sensing images and combining them with vegetation indices takes into account the vegetation growth pattern and substantially improves the identification accuracy, but it has high requirements for imaging, such as registration, multiple times, etc. Sometimes, it is difficult to satisfy, the plateau area is severely limited by the influence of clouds and rain, and Gaofen (GF) data require more control points for orthophoto correction. The study area was chosen to be Huize County, situated in Qujing City of Yunnan Province, China. The analysis was using the GF and Landsat images. According to deep learning and remote sensing image feature extraction methods, the semantic segmentation method of F-Pix2Pix was proposed, and the domain adaptation method according to transfer learning effectively solved the class imbalance in needleleaf/broadleaf forest identification. The results showed that (1) this method had the best performance and a higher accuracy than the existing products, 21.48% in non-forest/forest and 29.44% in needleleaf/broadleaf forest for MIoU improvement. (2) Applying transfer learning domain adaptation to semantic segmentation showed significant benefits, and this approach utilized satellite images of different resolutions to solve the class imbalance problem. (3) It can be used for long-term monitoring of multiple images and has strong generalization. The identification of needleleaf and broadleaf forests combined with the actual geographical characteristics of the forest provides a foundation for the accurate estimation of regional carbon sources/sinks. [ABSTRACT FROM AUTHOR]

Superpixel segmentation has been widely used in computer vision as a preprocessing to reduce the number of image primitives for subsequent image processing tasks. Simple Linear Iterative Clustering (SLIC) is the most popular superpixel segmentation algorithm in the literature, which is an iterative algorithm. The basic SLIC technique includes two main steps named Assignment and Update steps, and their data dependency is the main problem to speed up execution of the basic SLIC algorithm. It is also well known that this algorithm is not memory efficient, which is the main bottleneck in applying basic SLIC in low-power applications. In this paper, we have proposed a new structure for SLIC implementation. This new structure replaces the cluster-based search operation with pixel-based process in the Assignment step that gains considerable speed improvement. This new technique simply removes the data dependency in the SLIC structure. In the proposed method, the distance index, the main parameter computed in the Assignment step, no longer is needed to be stored in memory. This leads to a significant improvement in the memory efficiency of the proposed structure. The proposed structure is implemented on the Virtex-5 FPGA family, and the results show about 26 % in memory improvement compared to basic SLIC. According to the implementation results, the proposed technique process 143 frames per second for an input image of 300 × 400 pixel resolution which is far faster than real-time process requirements. [ABSTRACT FROM AUTHOR]

Multispectral satellite imagery offers a new perspective for spatial modelling, change detection and land cover classification. The increased demand for accurate classification of geographically diverse regions led to advances in object-based methods. A novel spatiotemporal method is presented for object-based land cover classification of satellite imagery using a Graph Neural Network. This paper introduces innovative representation of sequential satellite images as a directed graph by connecting segmented land region through time. The method's novel modular node classification pipeline utilises the Convolutional Neural Network as a multispectral image feature extraction network, and the Graph Neural Network as a node classification model. To evaluate the performance of the proposed method, we utilised EfficientNetV2-S for feature extraction and the GraphSAGE algorithm with Long Short-Term Memory aggregation for node classification. This innovative application on Sentinel-2 L2A imagery produced complete 4-year intermonthly land cover classification maps for two regions: Graz in Austria, and the region of Portorož, Izola and Koper in Slovenia. The regions were classified with Corine Land Cover classes. In the level 2 classification of the Graz region, the method outperformed the state-of-the-art UNet model, achieving an average F1-score of 0.841 and an accuracy of 0.831, as opposed to UNet's 0.824 and 0.818, respectively. Similarly, the method demonstrated superior performance over UNet in both regions under the level 1 classification, which contains fewer classes. Individual classes have been classified with accuracies up to 99.17%. [ABSTRACT FROM AUTHOR]

4D Light Field (LF) imaging, since it conveys both spatial and angular scene information, can facilitate computer vision tasks and generate immersive experiences for end-users. A key challenge in 4D LF imaging is to flexibly and adaptively represent the included spatio-angular information to facilitate subsequent computer vision applications. Recently, image over-segmentation into homogenous regions with perceptually meaningful information has been exploited to represent 4D LFs. However, existing methods assume densely sampled LFs and do not adequately deal with sparse LFs with large occlusions. Furthermore, the spatio-angular LF cues are not fully exploited in the existing methods. In this paper, the concept of hyperpixels is defined and a flexible, automatic, and adaptive representation for both dense and sparse 4D LFs is proposed. Initially, disparity maps are estimated for all views to enhance over-segmentation accuracy and consistency. Afterwards, a modified weighted $K$ -means clustering using robust spatio-angular features is performed in 4D Euclidean space. Experimental results on several dense and sparse 4D LF datasets show competitive and outperforming performance in terms of over-segmentation accuracy, shape regularity and view consistency against state-of-the-art methods. [ABSTRACT FROM AUTHOR]

The paper presents a low-cost and LiDAR-free approach to efficiently detect 3D objects from stereo camera images, towards autonomous driving applications. It is first proposed to exploit the simple linear iterative clustering algorithm to segment stereo images into superpixel feature maps. The segmented superpixel maps are then used to estimate a depth map. By utilizing the depth map and stereo images, a 3D point cloud can be generated; and the 3D data is considered as pseudo-LiDAR representation as it is similar to measurements collected by a LiDAR sensor. The generated pseudo-LiDAR point cloud can ultimately be fed into any the state-of-the-art LiDAR based 3D object detection techniques to localize objects. By doing this, the proposed approach can effectively detect 3D objects by only employing low-cost stereo cameras, which can save tens of thousands of dollars on LiDAR costs from the existing LiDAR based methods. Effectiveness of the proposed algorithm was evaluated in the real-world KITTI dataset where the obtained results are about 1.33% better than those obtained by the benchmarking pseudo-LiDAR++ method (You et al. 2020). [ABSTRACT FROM AUTHOR]

High-Level Structure (HLS) recognition locates elements on human-made surfaces (objects, buildings, ground, etc.). There are several approaches to HLS recognition, however, most of these approaches are based on processing 3D data in the form of point clouds extracted from the camera images. In general, 3D point cloud approaches have good performance for certain scenes with video sequences or image sequences, but they need sufficient parallax in order to guarantee accuracy. To address this problem, an alternative is to process a single RGB image seeking to interpret areas of the images where the human-made structure may be observed, thus removing parallax dependency, but adding the challenge of having to interpret image ambiguities correctly. Motivated by the latter, this work presents the results of a novel methodology for HLS recognition using a CNN-Superpixel approach from a single image. For that, our approach has three steps. First, the superpixel and centroid analysis obtains the RGB section and the superpixel to analyze. This section is a portion of the input image that our CNN uses to provide a label. Second, the structure recognition step provides a segmentation, location, and delimitation of the urbanized structures in the scene. For that, we propose a CNN-superpixel configuration, this configuration combines the abstraction power of deep learning and fast computational processing using superpixel segmentation. Third, the connectivity analysis replaces the superpixel label considering the connection between the neighbor superpixels. On the other hand, experimental results are encouraging, our approach has high performance under real-world scenarios. Also, the proposed methodology is 6.53 to 12.18 faster than previous work. [ABSTRACT FROM AUTHOR]

Recently, due to the ability to exploit perceptual information, superpixel-based methods have received attention for hyperspectral image (HSI) denoising. However, existing superpixel-based denoising methods unfold the irregular 3-D superpixels into the matrices along the spectral mode, which inevitably destroys the intrinsic structure of the irregular 3-D superpixels. To tackle the irregular 3-D superpixels, we introduce the irregular tensor representation for superpixel-guided HSI denoising. More concretely, by introducing the weighted tensor, we suggest a tensor representation of each irregular 3-D superpixel, which can preserve the intrinsic structure of the irregular 3-D superpixel. Equipped with the irregular tensor representation, we establish a superpixel-guided tensor optimization model for HSI denoising, which simultaneously exploits the perceptual information and low-rank structure within 3-D superpixels. To solve the resulting tensor optimization problem, we develop an inexact augmented Lagrange multiplier (IALM) algorithm. Experimental results show that the proposed method outperforms other state-of-the-art HSI denoising methods, particularly matrix-based methods, on both simulated and real data. [ABSTRACT FROM AUTHOR]

Ship detection in synthetic aperture radar (SAR) images is critical to ocean surveillance and rescue. Although many deep learning SAR ship detection methods have been proposed, the performance of these methods depends on the size and quality of the training samples. To resolve these issues, this letter presents an unsupervised ship detection method in SAR images using superpixel segmentation and cross stage partial network (CSPNet). First, the SAR image is over-segmented into superpixels based on our previously proposed superpixel generation algorithm. Then, the complex signal kurtosis (CSK) and a local superpixel contrast are integrated as a statistical indicator for the automatic identification of ship superpixels and background superpixels, thus leading to generation of training samples. Finally, the segmented superpixels are input to the CSPNet, which can learn a representative feature set with high discrimination ability between ships and backgrounds. Our method can achieve pixel-level detection map rather than the bounding box result. Experiments based on the Gaofen-3 and TerraSAR SAR data demonstrate that our method can achieve above 90% actual detection rate. [ABSTRACT FROM AUTHOR]

Since fully automatic image segmentation methods often fail for most complex images, researchers turn to the interactive segmentation paradigm to achieve better segmentation performance. However, many interactive image segmentation algorithms are highly dependent on user interactive information. This paper presents a novel interactive image segmentation algorithm based on multi-layer random forests. Given a small amount of user input markers, region merging is done according to the merging rule, in which both the color histogram and gradient orientation histogram of the region are included to avoid the merging error. To speed up the calculation of gradient orientation histogram, breadth-first search is used to determine the intersection of two adjacent regions. Then, we relabel the training samples with k-means algorithm and Silhouette index and further perform the first layer random forest classification. Next, we reconstruct the training samples with the adjacent superpixel pairs and use the second layer random forest classifiers to classify the superpixels whose prediction confidence is lower than the threshold after the first layer random forest classification. Experiments on real natural images are conducted to demonstrate the performance of the proposed algorithm. [ABSTRACT FROM AUTHOR]

Given the escalating frequency and severity of global forest fires, it is imperative to develop advanced detection and segmentation technologies to mitigate their impact. To address the challenges of these technologies, the development of deep learning-based forest fire surveillance has significantly accelerated. Nevertheless, the integration of graph convolutional networks (GCNs) in forest fire detection remains relatively underexplored. In this context, we introduce a novel superpixel-based graph convolutional network (SCGCN) for forest fire image segmentation. Our proposed method utilizes superpixels to transform images into a graph structure, thereby reinterpreting the image segmentation challenge as a node classification task. Additionally, we transition the spatial graph convolution operation to a GraphSAGE graph convolution mechanism, mitigating the class imbalance issue and enhancing the network’s versatility. We incorporate an innovative loss function to contend with the inconsistencies in pixel dimensions within superpixel clusters. The efficacy of our technique is validated on two different forest fire datasets, demonstrating superior performance compared to four alternative segmentation methodologies.

Superpixel segmentation is a fundamental computer vision technique that finds application in a multitude of high level computer vision tasks. Most state-of-the-art superpixel segmentation methods are unsupervised in nature and thus cannot fully utilize frequently occurring texture patterns or incorporate multiscale context. In this paper, we show that superpixel segmentation can be improved by leveraging the superior modeling power of deep convolutional autoencoders in a fully unsupervised manner. We pose the superpixel segmentation problem as one of manifold learning where pixels that belong to similar texture patterns are assigned near identical embedding vectors. The proposed deep network is able to learn image-wide and dataset-wide feature patterns and the relationships between them. This knowledge is used to segment and group pixels in a way that is consistent with a more global definition of pattern coherence. Experiments demonstrate that the superpixels obtained from the embeddings learned by the proposed method outperform the state-of-theart superpixel segmentation methods for boundary precision and recall values. Additionally, we find that semantic edges obtained from the superpixel embeddings to be significantly better than the contemporary unsupervised approaches.

In order to avoid the phenomenon that edges were easily blurred during filtering, a multi-scale guided filtering integrated with superpixel and patch shift was proposed.Firstly, the bilateral filtering was applied to an input image to get more accurate superpixel regions.Then, the overlapping was used as the final filtering region, which was formed by the local window and the superpixel.The local window was selected by a metric function so that it did not contain edge information as much as possible.Finally, a small-scale window was used to preserve the edges, and then the window scale was increased for iterative guided filtering.The final filtering output was calculated by fusing the results of different window scales to smooth details while preserving the structural edges.Additionally, the evaluation index of filtering results quality was proposed and compared in different algorithms.The proposed algorithm has stronger edge preservation ability and can obtain better filtering effectiveness than other algorithms.

This paper presents a semi-supervised learning method based on superpixels and convolutional neural networks (CNNs) to assign and improve the identification of weeds in soybean crops. Despite its promising results, CNNs require massive amounts of labeled training data to learn, so we intend to improve the manual labeling phase with an automated pseudo-labeling process. We propose a method that uses an additional phase of mini-batch processing to fine-tune and assign pseudo labels to the images based on previously annotated SLIC segmentation during the algorithm training phase. This research paper shows that the proposed method improves the soybean monitoring accuracy compared with the traditionally trained methods using a tiny amount of labeled superpixels. There was an increase in the training time, but this is an expected result and even preferable to doing manual label annotation..

To address key issues with traditional landform classification methods that impact the integrity and continuity of the analysis area, as well as neglecting neighboring features, We presents a novel approach called superpixel-based graph convolutional network (SGCN) for automatic classification of digital elevation model (DEM) landform. The SGCN uses superpixel segmentation and graph convolutional networks (GCNs). Specifically, we employ the Simple Linear Iterative Clustering (SLIC) algorithm to perform superpixel segmentation on the DEM image and convert the resulting distribution of superpixels in the 2D image into a graph structure, where nodes are labeled. We then analyze the geographic features and external neighborhood relationships within each DEM superpixel to obtain corresponding node information for the graph structure. Finally, the GCN model is trained according to the graph structure, node feature information, and node label, enabling automatic recognition of node landform types. Importantly, SGCN accounts for the multi-order neighborhood features of the basic analysis unit and expresses the DEM micro-features relatively macroscopically. Experimental results demonstrate that compared with pixel-level semantic segmentation methods like U-Net, FCN, and SegNet, the SGCN significantly improves DEM landform classification accuracy by 17.84%, 33.22%, and 30.60%, respectively. Moreover, SGCN better preserves the integrity and continuity of features inside and outside the analysis area, overcoming the limitation of traditional methods that use regular grids as the basic analysis unit, which can compromise landform edge continuity.

Recently, the deep learning algorithms have been increasingly utilized in remote sensing change detection. However, incomplete buildings and the blurred edges caused by the complex scenes in change detection applications make the detection results fail to describe the real land cover changes. Superpixels can be used to alleviate edge blurring, but the existing superpixel methods cannot be trained jointly with the models in change detection. In this work, we investigated an innovative double-head method using deep learning, called double U-Net (W-Net), which consists of a superpixel module and a change detection module. Due to the superpixel module, W-Net can handle building edges very well. In order to solve problem that multiple subtasks fail to achieve the optimal results, a two-branch multi-task coupling framework of change detection and superpixels is designed for W-Net, which enables the model to achieve a globally optimal detection performance. The advancement of the W-Net was demonstrated using three public datasets. The F1score on LEVIR-CD dataset was 0.9031 and kappa coefficient was 0.8969. The F1-score on WHU building dataset was 0.9172 and kappa coefficient was 0.9142. The F1-score on SYSU-CD dataset was 0.8167and and kappa coefficient was 0.7724. The experiments confirmed that the W-Net is capable to detect the edges of changed area better and outperforms the other advanced change detection methods.

Image security is becoming more and more important in recently years. To improve the efficiency and security, this paper defines the concepts of the superpixel and super image, and proposes a multiple-image encryption (MIE) algorithm based on the bit plane and superpixel. The superpixel is an integer formed by connecting the binary values of multiple pixels head to tail and then converting them into a decimal number. The proposed algorithm adopts the classical scrambling-diffusion structure. At the scrambling stage, our algorithm uses the scrambling operation among bit planes and the extended Zigzag transformation in the bit plane. At the diffusion stage, our algorithm performs the exclusive OR operation on the scrambled super image. Different from other MIE algorithms, our algorithm can encrypt multiple images with the workload of processing one image. The experiments and comparative analysis, i.e., the results such as key space, differential attack, etc., show that the proposed algorithm has excellent encryption efficiency and high security. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. 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.)