Watershed Algorithm Research Articles

A novel fully convolutional network based on marker-controlled watershed segmentation algorithm for industrial soot robot target segmentation

A novel fully convolutional network based on marker-controlled watershed segmentation algorithm for industrial soot robot target segmentation

High-spatial-resolution remote sensing image segmentation using adaptive watershed-driven joint MDEDNet

When segmenting high-spatial-resolution remote sensing images using convolution neural networks, it is difficult to balance the number of segmented regions against the contour accuracy. This article proposes a segmentation method that combines the minimum area adaptive watershed transform based on morphological reconstruction with the modified deep edge detection network (MDEDNet). First, the minimum area adaptive watershed algorithm is adopted to reconstruct the gradient features of the image and perform watershed segmentation, giving the initial segmentation results. A ground object contour optimization method and MDEDNet are then used to further improve the accuracy of the segmentation, resulting in the final segmented image. Experiments on image sets of varying complexity indicate that the algorithm proposed in this article can achieve accurate image segmentation with good segmentation results and fast operation efficiency.

A back‐propagation neural network model using hyperspectral imaging applied to variety nondestructive detection of cereal

AbstractCereal crops play an important role in preventing chronic diseases and regulating human functions due to their rich phytochemicals. However, an increasing number of cases of cereal‐crop adulteration are occurring, affecting the nutrition of food and threatening food safety. Therefore, this study used hyperspectral imaging (HSI) and a back‐propagation neural network (BPNN) to establish the variety identification model of different grain shapes, colors, and sizes. This model can conduct a quantitative and qualitative analysis of samples rapidly and nondestructively, because HSI can collect spectral information and spatial information of samples. Meanwhile, the Watershed algorithm was used to identify cereal varieties by particles, and the model identification performance was verified by the unlabeled test sets. The results show that the data extraction success rate of the new watershed algorithm reached 98%, and the comprehensive identification accuracy of the model reached 90%. In addition, the cereal in the training set can be changed to identify other cereal crops, thereby providing a method of rapid and nondestructive adulteration detection of cereals.Practical ApplicationsCereal crops play an important role in preventing chronic diseases and regulating human functions due to their rich phytochemicals. However, an increasing number of cases of cereal‐crop adulteration are occurring, affecting the nutrition of food and threatening food safety. Therefore, this study used hyperspectral imaging (HSI) and a back‐propagation neural network (BPNN) to establish the variety identification model of different grain shapes, colors, and sizes. This model can conduct a quantitative and qualitative analysis of samples rapidly and nondestructively, because HSI can collect spectral information and spatial information of samples at the same time. The spectral information was used for qualitative analysis, while spatial information was used for quantitative analysis, so this model can realize the rapid and nondestructive detection of different varieties of cereals. And we can also change the training‐set data to realize the variety identification of other varieties of crops, which provides guidance for the method detecting the adulteration of cereal crops.

Detection of Rotten Fresh-Cut Cauliflowers based on Machine Vision Technology and Watershed Segmentation Method

In this study, machine vision technology was used to separate the samples and detect the rotting degrees of fresh-cut cauliflowers. First, the improved watershed algorithm was used for the segmentation of fresh-cut cauliflower samples and the extraction of single-sample. Then, three color models, a gray co-occurrence matrix and two feature extraction algorithms were used to extract the color, texture and spectral feature parameters of the images. At the same time, the Partial Least Squares Discriminant Analysis (PLS-DA) and Extreme Learning Machines (ELM) discriminant models were established. The identification accuracy of PLS-DA and ELM discriminant models for rotting samples was 95 and 90.9%, respectively. Moreover, according to the size of rotten areas, the rotting grades were divided and the contours and feature areas of rotten cauliflower samples were identified by the region growth algorithm and the “Sobel” operator. Finally, the detection and identification of the rotting degree of cauliflower samples were realized. The results showed that machine vision technology can segment the cohesive fresh-cut cauliflower samples and can be used for qualitative and quantitative identification of the intact and rotten cauliflower samples.

Benchmarking conventional and machine learning segmentation techniques for digital rock physics analysis of fractured rocks

Image segmentation remains the most critical step in Digital Rock Physics (DRP) workflows, affecting the analysis of physical rock properties. Conventional segmentation techniques struggle with numerous image artifacts and user bias, which lead to considerable uncertainty. This study evaluates the advantages of using the random forest (RF) algorithm for the segmentation of fractured rocks. The segmentation quality is discussed and compared with two conventional image processing methods (thresholding-based and watershed algorithm) and an encoder–decoder network in the form of convolutional neural networks (CNNs). The segmented images of the RF method were used as the ground truth for CNN training. The images of two fractured rock samples are acquired by X-ray computed tomography scanning (XCT). The skeletonized 3D images are calculated, providing information about the mean mechanical aperture and roughness. The porosity, permeability, flow fields, and preferred flow paths of segmented images are analyzed by the DRP approach. Moreover, the breakthrough curves obtained from tracer injection experiments are used as ground truth to evaluate the segmentation quality of each method. The results show that the conventional methods overestimate the fracture aperture. Both machine learning approaches show promising segmentation results and handle all artifacts and complexities without any prior CT-image filtering. However, the RF implementation has superior inherent advantages over CNN. This method is resource-saving (e.g., quickly trained), does not need an extensive training dataset, and can provide the segmentation uncertainty as a measure for evaluating the segmentation quality. The considerable variation in computed rock properties highlights the importance of choosing an appropriate segmentation method.

A high-throughput semi-automated bone segmentation workflow for murine hindpaw micro-CT datasets

IntroductionMicro-computed tomography (μCT) is a valuable imaging modality for longitudinal quantification of bone volumes to identify disease or treatment effects for a broad range of conditions that affect bone health. Complex structures, such as the hindpaw with up to 31 distinct bones in mice, have considerable analytic potential, but quantification is often limited to a single bone volume metric due to the intensive effort of manual segmentation. Herein, we introduce a high-throughput, user-friendly, and semi-automated method for segmentation of murine hindpaw μCT datasets. MethodsIn vivo μCT was performed on male (n = 4; 2–8-months) and female (n = 4; 2–5-months) C57BL/6 mice longitudinally each month. Additional 9.5-month-old male C57BL/6 hindpaws (n = 6 hindpaws) were imaged by ex vivo μCT to investigate the effects of resolution and integration time on analysis outcomes. The DICOMs were exported to Amira software for the watershed-based segmentation, and watershed markers were generated automatically at approximately 80% accuracy before user correction. The semi-automated segmentation method utilizes the original data, binary mask, and bone-specific markers that expand to the full volume of the bone using watershed algorithms. ResultsCompared to the conventional manual segmentation using Scanco software, the semi-automated approach produced similar raw bone volumes. The semi-automated segmentation also demonstrated a significant reduction in segmentation time for both experienced and novice users compared to standard manual segmentation. ICCs between experienced and novice users were >0.9 (excellent reliability) for all but 4 bones. DiscussionThe described semi-automated segmentation approach provides remarkable reliability and throughput advantages. Adoption of the semi-automated segmentation approach will provide standardization and reliability of bone volume measures across experienced and novice users and between institutions. The application of this model provides a considerable strategic advantage to accelerate various research opportunities in pre-clinical bone and joint analysis towards clinical translation.

Bubble Image Segmentation Based on a Novel Watershed Algorithm With an Optimized Mark and Edge Constraint

Bubble size contains important indication information that is closely related to flotation production conditions and process indicators. However, bubble images often have low contrast, noise, and many other shortcomings, making foam segmentation a difficult problem that the existing segmentation methods cannot solve. In this article, an improved watershed algorithm based on optimal labeling and edge constraints is proposed. Three algorithms are designed to obtain different initial tags, and then the extracted content of different tags is fused to obtain the combined foreground tag. To reduce the offset of the segmentation line, the edge operator is applied to extract the bubble boundary, and the boundary priori condition is used as a constraint to correct the segmentation line. Finally, the optimal segmentation line is obtained by fusing foreground markers and external constraints. Industrial experiments show that this method is effective and has a higher accuracy than the other methods. The average value and variance of rand index (RI) are 92.88% and 0.69, respectively.

The Clinical Diagnostic Value of F-FDG PET/CT Combined with MRI in Pancreatic Cancer.

Pancreatic cancer is a common gastrointestinal tumor with increasing morbidity and mortality, and it is difficult to differentiate it from chronic mass pancreatitis. F-18-fluorodeoxyglucose positron emission tomography/computed tomography (F-FDG PET/CT) can be used for diagnosis of pancreatic cancer, staging, radiotherapy planning, evaluation of efficacy and recurrence, and differentiation from posttreatment fibrosis but the sensitivity and specificity of diagnosis of chronic pancreatitis are poor. Magnetic resonance imaging (MRI) not only shows the morphology, location, and size of organs and lesions but also has the advantages of low radiation, high soft tissue contrast, and multiparameter imaging, which can help determine the nature of occupying pancreatic lesions. In this study, the watershed algorithm (WA) was used to segment the abdominal images to extract images of pancreatic cancer cells and to compare the value of F-FDG PET/CT combined with MRI in the diagnosis of pancreatic occupying lesions.

Research on the Extraction Method of Urban Built-Up Areas With an Improved Night Light Index

Nighttime light (NTL) data has been widely used in the research of urban built-up areas extraction. Due to the low resolution and the overflow characteristics of NTL data, it is difficult to extract urban built-up areas in small cities with undeveloped economy and small areas. How to extract them accurately becomes an important problem. Based on Normalized Difference Vegetation Index (NDVI), Normalized Difference Built-up Index (NDBI), and Normalized Land Surface Temperature NLST), we proposed an improved night light index, named Vegetation, Building and Temperature Enhanced NTL Urban Index (VBTANUI) to accurately extract built-up areas in small cities by the marker-based watershed segmentation algorithm. Taking Jiayuguan, Ezhou and Wuzhishan, China as three examples, we extracted their urban built-up areas. Comparing with the four traditional NTL indexes, such as VANUI, EANI, BANUI65 and BANUI43, the results showed that: (1) the VBTANUI could effectively enhance the light intensity of the built-up boundaries; (2) F1 scores of VBTANUI in three cities are of 89.60%, 95.70% and 82.75% respectively, which are 6.04%, 7.12%, 8.10% and 8.47% higher in average than that of traditional four indexes; (3) VBTANUI is feasible and effective in the extraction of built-up areas in small cities with undeveloped economy and small areas.

A novel pre-processing approach based on colour space assessment for digestive neuroendocrine tumour grading in immunohistochemical tissue images.

The complexity of histopathological images remains achallenging issue in cancer diagnosis. Apathologist analyses immunohistochemical images to detect acolour-based stain, which is brown for positive nuclei with different intensities and blue for negative nuclei. Several issues emerge during the eyeballing tissue slide analysis, such as colour variations caused by stain inhomogeneity, non-uniform illumination, irregular cell shapes, and overlapping cell nuclei. To overcome those problems, an automated computer-aided diagnosis system is proposed to segment and quantify digestive neuroendocrine tumours. We present anovel pre-processing approach based on colour space assessment. Acriterion called pertinence degree is introduced to select the appropriate colour channel, followed by contrast enhancement. Subsequently, the adaptive local threshold technique that uses the modified Laplacian filter is applied to minimize the implementation complexity, highlight edges, and emphasize intensity variation between cells across the slide. Finally, the improved watershed algorithm based on the concave vertex graph is applied for cell separation. The performance of the algorithms for nucleus segmentation is evaluated according to both the object-level and pixel-level criteria. Our approach increases segmentation accuracy, with the F1-score equal to 0.986. There is significant agreement between the applied approach and the expert's ground truth segmentation. The proposed method outperformed the state-of-the-art techniques based on recall, precision, the F1-score, and the Dice coefficient.

Application of an Improved Watershed Algorithm Based on Distance Map Reconstruction in Bean Image Segmentation

Application of an Improved Watershed Algorithm Based on Distance Map Reconstruction in Bean Image Segmentation

Application of an Improved Watershed Algorithm Based on Distance Map Reconstruction in Bean Image Segmentation

Application of an Improved Watershed Algorithm Based on Distance Map Reconstruction in Bean Image Segmentation

Retracted] Key Frame Extraction Method for Minors’ Participation in Online Short‐Form Video from the Perspective of Government Administration

With the progress of the Internet era, the audience scope of online short‐form video and live broadcast platform is rapidly expanding. In this situation, the physical and mental health of minors is affected and harmed, and serious social problems have been caused. In view of the random images of minors participating in short online videos, the monitoring department should strengthen supervision and control and establish a good network atmosphere. The focus of surveillance is to identify short‐form videos in which minors participate, monitor the content of short‐form videos, and effectively identify minors. On the basis of analyzing and studying the existing methods of moving object extraction, the method of spatio‐temporal information combined with geometric curve evolution is used to improve the effect of moving object extraction. Firstly, the time dynamic information of video frames is fully utilized and the initial contour of the target is segmented by high‐order statistical algorithm. Based on the improved watershed algorithm, the spatial frames of the original image are segmented and the secondary contour of the target is obtained by combining the spatial and temporal information. The level set evolution algorithm is used to remove the noise and information void in the secondary target, and the final moving target is obtained. Through the extraction of key frames, it can effectively control the appearance of minors in short‐form videos, which can actively guide the benign development of the platform, and put forward governance suggestions based on the perspective of government management.

Application of Video Processing Technology Based on Diffusion Equation Model in Basketball Analysis

Video event detection and annotation work is an important content of video analysis and the basis of video content retrieval. Basketball is one of the most popular types of sports. Event detection and labeling of basketball videos can help viewers quickly locate events of interest and meet retrieval needs. This paper studies the application of anisotropic diffusion in video image smoothing, denoising, and enhancement. An improved form of anisotropic diffusion that can be used for video image enhancement is analyzed. This paper studies the anisotropic diffusion method for coherent speckle noise removal and proposes a video image denoising method that combines anisotropic diffusion and stationary wavelet transform. This paper proposes an anisotropic diffusion method based on visual characteristics, which adds a factor of video image detail while smoothing, and improves the visual effect of diffusion. This article discusses how to apply anisotropic diffusion methods and ideas to video image segmentation. We introduced the classic watershed segmentation algorithm and used forward-backward diffusion to process video images to reduce oversegmentation, introduced the active contour model and its improved GVF Snake, and analyzed the idea of how to use anisotropic diffusion and improve the GVF Snake model to get a new GGVF Snake model. In the study of basketball segmentation of close-up shots, we propose an improved Hough transform method based on a variable direction filter, which can effectively extract the center and radius of the basketball. The algorithm has good robustness to basketball partial occlusion and motion blur. In the basketball segmentation research of the perspective shot, the commonly used object segmentation method based on the change area detection is very sensitive to noise and requires the object not to move too fast. In order to correct the basketball segmentation deviation caused by the video noise and the fast basketball movement, we make corrections based on the peak characteristics of the edge gradient. At the same time, the internal and external energy calculation methods of the traditional active contour model are improved, and the judgment standard of the regional optimal solution and segmentation validity is further established. In the basketball tracking research, an improved block matching method is proposed. On the one hand, in order to overcome the influence of basketball’s own rotation, this article establishes a matching criterion that has nothing to do with the location of the area. On the other hand, this article improves the diamond motion search path based on the basketball’s motion correlation and center offset characteristics to reduce the number of searches and improve the tracking speed.

基于改进分水岭算法和U-net神经网络模型的复合材料CT图像分割方法

基于改进分水岭算法和U-net神经网络模型的复合材料CT图像分割方法

Automated delineation of agricultural field boundaries from Sentinel-2 images using recurrent residual U-Net

Delineation of agricultural fields is desirable for operational monitoring of agricultural production and is essential to support food security. Due to large within-class variance of pixel values and small inter-class difference, automated field delineation remains to be a challenging task. In this study, a strategy is proposed to effectively address this issue. Firstly, a framework was developed using the Canny operator connected with the Watershed segmentation algorithm (CW) to quickly label the training dataset, which minimizes the workload of dataset generation in comparison with the commonly used manual vectorization. Secondly, a CW-trained deep semantic segmentation network, recurrent residual U-Net, was selected to mine the low level and deep semantic features. Finally, a boundary connecting method (to integrate fragmented boundaries) was used to generate the agricultural field boundary. The proposed methods are tested over smallholder agricultural landscape in Heilongjiang province, China, using Sentinel-2 imagery. Compared with the U-Net (overall accuracy (OA) 82.18%), the residual U-Net (ResU-Net, with OA 85.78%), traditional object-based image analysis (OBIA, with OA about 82%), and the existing 10-m resolution global land cover map (FROM-GLC10), the proposed method shows an improved performance (OA 89.28%, and Kappa 0.85). The successful application of the proposed method suggests that the recurrent residual U-Net model has great universality in agricultural field boundary extraction, and the automated technique has the potential of being applied to other regions.

Automatic segmentation of foveal avascular zone based on adaptive watershed algorithm in retinal optical coherence tomography angiography images

The size and shape of the foveal avascular zone (FAZ) have a strong positive correlation with several vision-threatening retinovascular diseases. The identification, segmentation and analysis of FAZ are of great significance to clinical diagnosis and treatment. We presented an adaptive watershed algorithm to automatically extract FAZ from retinal optical coherence tomography angiography (OCTA) images. For the traditional watershed algorithm, “over-segmentation” is the most common problem. FAZ is often incorrectly divided into multiple regions by redundant “dams”. This paper analyzed the relationship between the “dams” length and the maximum inscribed circle radius of FAZ, and proposed an adaptive watershed algorithm to solve the problem of “over-segmentation”. Here, 132 healthy retinal images and 50 diabetic retinopathy (DR) images were used to verify the accuracy and stability of the algorithm. Three ophthalmologists were invited to make quantitative and qualitative evaluations on the segmentation results of this algorithm. The quantitative evaluation results show that the correlation coefficients between the automatic and manual segmentation results are 0.945 (in healthy subjects) and 0.927 (in DR patients), respectively. For qualitative evaluation, the percentages of “perfect segmentation” (score of 3) and “good segmentation” (score of 2) are 99.4% (in healthy subjects) and 98.7% (in DR patients), respectively. This work promotes the application of watershed algorithm in FAZ segmentation, making it a useful tool for analyzing and diagnosing eye diseases.

AUTOMATED SEGMENTATION OF COVID-19 REGIONS FROM LUNG CT IMAGES USING WATERSHED ALGORITHM AND CLASSIFICATION USING MACHINE LEARNING CLASSIFIERS

The COVID-19 pandemic originated in Wuhan, China in December 2019 and has since affected over 200 countries worldwide. The highly contagious Coronavirus primarily affects the respiratory system, causing pulmonary inflammation that can be visualized through medical imaging such as CT and X-rays. Conventional testing methods include PCR and antibody tests. Shortage of test kits in hospitals as well as time taken for results to be received can be compensated through medical imaging. Therefore, there is a need for an automated system, which is accurate and robust in detection of Covid-19 from medical radiographs for clinical practice. The objectives of our study are as follows: (i) To segment the lung CT images using a hybrid watershed and fuzzy c-means algorithm. (2) To extract various textural features using the GLCM algorithm. (iii) To implement machine learning classifiers for classification of COVID and non-COVID image classes. Our dataset consisting of 60 chest CT images of COVID-19 and non-COVID-19 patients was pre-processed and segmented using a hybrid watershed and fuzzy c-means algorithm. Then, textural features were extracted from the segmented ROI using the GLCM algorithm. Finally, the images were classified into COVID and non-COVID classes using three machine learning classifiers namely Naïve Bayes, SVM and K-star. Naïve Bayes classifier achieved the highest accuracy of 95%, while SVM achieved 93% accuracy. The ROC curves were also obtained, with AUC of 0.98. Thus, our proposed system has shown promising results in the classification of lung CT images into the two classes namely COVID and non-COVID.

Diffusion Tensor Imaging Features of Watershed Segmentation Algorithm for Analysis of the Relationship between Depression and Brain Nerve Function of Patients with End-Stage Renal Disease.

The aim of this research was to explore the relationship between depression and brain nerve function in patients with end-stage renal disease (ESRD) and long-term maintenance hemodialysis (MHD) based on watershed segmentation algorithm using diffusion tensor imaging (DTI) technology. A total of 29 ESRD patients with depression who received MHD treatment in the hemodialysis center of hospital were included as the research subjects (case group). A total of 29 healthy volunteers were recruited as the control group, and a total of 29 ESRD patients with depression and brain lesions were recruited as the control group (HC group). Within 24 h after hemodialysis, the blood biochemical indexes were collected before this DTI examination. All participants completed the neuropsychological scale (MoCA, TMT A, DST, SAS, and SDS) test. The original DTI data of all subjects were collected and processed based on watershed segmentation algorithm, and the results of automatic segmentation according to the image were evaluated as DSC = 0.9446, MPA = 0.9352, and IOU = 0.8911. Finally, the average value of imaging brain neuropathy in patients with depression in the department of nephrology was obtained. The differences in neuropsychological scale scores (PSQI, MoCA, TMTA, DST, SAS, and SDS) between the two groups were statistically significant (P < 0.05). The differences of FA values in all the white matter partitions of Fu organs, except the cingulum of hippocampus (CgH) between the two groups, were statistically significant (P < 0.05). ESRD and DTI quantitative detection under the guidance of watershed segmentation algorithm in MHD patients showed that ESRD patients can be early identified, so as to carry out psychological nursing as soon as possible to reduce the occurrence of depression, and then protect the brain nerve to reduce brain neuropathy.

A New Watershed Algorithm for Pothole Image Segmentation

A New Watershed Algorithm for Pothole Image Segmentation