Image Segmentation Method Research Articles

A lie group semi-supervised FCM clustering method for image segmentation

As an unsupervised clustering method with low overhead, Fuzzy C-means (FCM) clustering has been widely used in a variety of image segmentation tasks. However, existing FCM clustering methods are sensitive to image noises and are either suffer from losing of image detail or falling into local optima in identifying cluster centers. Aiming at these problems, this paper proposes a Lie group semi-supervised FCM (LieSSFCM) clustering method for image segmentation. The method maps the input image from Euclidean space to Lie group manifold by representing each image pixel as a matrix Lie group and calculates geodesic distances between group elements and cluster centers on Lie group manifold. Prior information of the image and neighborhood relationships of pixels are used to guide the initialization and constrain the update of cluster centers and the corresponding fuzzy membership matrix. The proposed LieSSFCM has been validated against two medical image datasets and was compared with seven FCM clustering methods. Experimental results along with a systematic evaluation demonstrated that the method was superior in segmentation accuracy both visually and statistically, robustness to noises, adaptability to different tasks, and stability while maintaining a moderate computational complexity.

Cellpose as a reliable method for single-cell segmentation of autofluorescence microscopy images.

Autofluorescence microscopy uses intrinsic sources of molecular contrast to provide cellular-level information without extrinsic labels. However, traditional cell segmentation tools are often optimized for high signal-to-noise ratio (SNR) images, such as fluorescently labeled cells, and unsurprisingly perform poorly on low SNR autofluorescence images. Therefore, new cell segmentation tools are needed for autofluorescence microscopy. Cellpose is a deep learning network that is generalizable across diverse cell microscopy images and automatically segments single cells to improve throughput and reduce inter-human biases. This study aims to validate Cellpose for autofluorescence imaging, specifically from multiphoton intensity images of NAD(P)H. Manually segmented nuclear masks of NAD(P)H images were used to train new Cellpose models. These models were applied to PANC-1 cells treated with metabolic inhibitors and patient-derived cancer organoids (across 9 patients) treated with chemotherapies. These datasets include co-registered fluorescence lifetime imaging microscopy (FLIM) of NAD(P)H and FAD, so fluorescence decay parameters and the optical redox ratio (ORR) were compared between masks generated by the new Cellpose model and manual segmentation. The Dice score between repeated manually segmented masks was significantly lower than that of repeated Cellpose masks (p<0.0001) indicating greater reproducibility between Cellpose masks. There was also a high correlation (R2>0.9) between Cellpose and manually segmented masks for the ORR, mean NAD(P)H lifetime, and mean FAD lifetime across 2D and 3D cell culture treatment conditions. Masks generated from Cellpose and manual segmentation also maintain similar means, variances, and effect sizes between treatments for the ORR and FLIM parameters. Overall, Cellpose provides a fast, reliable, reproducible, and accurate method to segment single cells in autofluorescence microscopy images such that functional changes in cells are accurately captured in both 2D and 3D culture.

Improved Latin hypercube sampling initialization-based whale optimization algorithm for COVID-19 X-ray multi-threshold image segmentation

Image segmentation techniques play a vital role in aiding COVID-19 diagnosis. Multi-threshold image segmentation methods are favored for their computational simplicity and operational efficiency. Existing threshold selection techniques in multi-threshold image segmentation, such as Kapur based on exhaustive enumeration, often hamper efficiency and accuracy. The whale optimization algorithm (WOA) has shown promise in addressing this challenge, but issues persist, including poor stability, low efficiency, and accuracy in COVID-19 threshold image segmentation. To tackle these issues, we introduce a Latin hypercube sampling initialization-based multi-strategy enhanced WOA (CAGWOA). It incorporates a COS sampling initialization strategy (COSI), an adaptive global search approach (GS), and an all-dimensional neighborhood mechanism (ADN). COSI leverages probability density functions created from Latin hypercube sampling, ensuring even solution space coverage to improve the stability of the segmentation model. GS widens the exploration scope to combat stagnation during iterations and improve segmentation efficiency. ADN refines convergence accuracy around optimal individuals to improve segmentation accuracy. CAGWOA's performance is validated through experiments on various benchmark function test sets. Furthermore, we apply CAGWOA alongside similar methods in a multi-threshold image segmentation model for comparative experiments on lung X-ray images of infected patients. The results demonstrate CAGWOA's superiority, including better image detail preservation, clear segmentation boundaries, and adaptability across different threshold levels.

A Review of Semantic Medical Image Segmentation Based on Different Paradigms

In recent years, with the widespread application of medical images, the rapid and accurate identification of these regions of interest in a large number of medical images has received widespread attention. This article provides a review of medical image segmentation methods based on deep learning. Firstly, an overview of medical image segmentation methods was provided in the relevant knowledge, segmentation types, segmentation processes, and image processing applications. Secondly, the applications of supervised, semi supervised, and unsupervised methods in medical image segmentation were discussed, and their advantages, disadvantages, and applicable scenarios were revealed through the application of a large number of specific segmentation examples in practical scenarios. Finally, the commonly used medical image segmentation datasets and evaluation indicators were introduced, and the current medical image segmentation methods were summarized and prospected. This review provides a comprehensive and in-depth understanding for researchers in the field of medical image segmentation, and provides valuable references for the design and implementation of future related work.

Cosine-enhanced tuna swarm optimized exponential entropy segmentation method for sand grain microscopic images

The feature extraction of sand grain size, color and texture is a necessary step to identify clastic components. The sand grain features are complex and various, which brings difficulties to geological identification. Aiming at these images, a cosine-enhanced tuna swarm optimized exponential entropy segmentation method is proposed, which can effectively preserve the texture features of various sand grains. Firstly, for the tuna swarm optimization (TSO) algorithm, three improvement strategies are proposed: cosine spiral movement, cosine parabolic movement and Gauss-Cauchy mutation, which improve the TSO's global and local search. This algorithm is called cosine-enhanced TSO (CETSO). Benchmark function experiments showed that the convergence accuracy and stability of CETSO are greatly improved, and the convergence speed is also slightly increased. Secondly, CETSO optimized the exponential entropy to automatically determine the segmentation thresholds, and the feasibility of the method is verified by taking the information content of the segmented image as the standard. Finally, segmentation experiments were carried out on the Yarlung Zangbo River sand microscopic image dataset, and the results show that the method has high segmentation accuracy and stability for images with high contrast, rich texture, or significant differences in the size of sand debris. Compared with TSO, the CETSO optimized exponential entropy segmented images achieved an improvement of 21% and 93% in the evaluation of the average and standard deviation of the peak signal-to-noise ratio on thirty experiments. And this method has a fast processing speed, and it only takes about 0.85s to divide an image on average, which meets the needs of engineering applications.

Segmentation-based BI-RADS ensemble classification of breast tumours in ultrasound images

BackgroundThe development of computer-aided diagnosis systems in breast cancer imaging is exponential. Since 2016, 81 papers have described the automated segmentation of breast lesions in ultrasound images using artificial intelligence. However, only two papers have dealt with complex BI-RADS classifications. PurposeThis study addresses the automatic classification of breast lesions into binary classes (benign vs. malignant) and multiple BI-RADS classes based on a single ultrasonographic image. Achieving this task should reduce the subjectivity of an individual operator’s assessment. Materials and MethodsAutomatic image segmentation methods (PraNet, CaraNet and FCBFormer) adapted to the specific segmentation task were investigated using the U-Net model as a reference. A new classification method was developed using an ensemble of selected segmentation approaches. All experiments were performed on publicly available BUS B, OASBUD, BUSI and private datasets. ResultsFCBFormer achieved the best outcomes for the segmentation task with intersection over union metric values of 0.81, 0.80 and 0.73 and Dice values of 0.89, 0.87 and 0.82, respectively, for the BUS B, BUSI and OASBUD datasets. Through a series of experiments, we determined that adding an extra 30-pixel margin to the segmentation mask counteracts the potential errors introduced by the segmentation algorithm. An assembly of the full image classifier, bounding box classifier and masked image classifier was the most accurate for binary classification and had the best accuracy (ACC; 0.908), F1 (0.846) and area under the receiver operating characteristics curve (AUROC; 0.871) in the BUS B and ACC (0.982), F1 (0.984) and AUROC (0.998) in the UCC BUS datasets, outperforming each classifier used separately. It was also the most effective for BI-RADS classification, with ACC of 0.953, F1 of 0.920 and AUROC of 0.986 in UCC BUS. Hard voting was the most effective method for dichotomous classification. For the multi-class BI-RADS classification, the soft voting approach was employed. ConclusionsThe proposed new classification approach with an ensemble of segmentation and classification approaches proved more accurate than most published results for binary and multi-class BI-RADS classifications.

How many specimens make a sufficient training set for automated three-dimensional feature extraction?

Deep learning has emerged as a robust tool for automating feature extraction from three-dimensional images, offering an efficient alternative to labour-intensive and potentially biased manual image segmentation methods. However, there has been limited exploration into the optimal training set sizes, including assessing whether artficial expansion by data augmentation can achieve consistent results in less time and how consistent these benefits are across different types of traits. In this study, we manually segmented 50 planktonic foraminifera specimens from the genus Menardella to determine the minimum number of training images required to produce accurate volumetric and shape data from internal and external structures. The results reveal unsurprisingly that deep learning models improve with a larger number of training images with eight specimens being required to achieve 95% accuracy. Furthermore, data augmentation can enhance network accuracy by up to 8.0%. Notably, predicting both volumetric and shape measurements for the internal structure poses a greater challenge compared with the external structure, owing to low contrast differences between different materials and increased geometric complexity. These results provide novel insight into optimal training set sizes for precise image segmentation of diverse traits and highlight the potential of data augmentation for enhancing multivariate feature extraction from three-dimensional images.

A Medical Image Segmentation Method Combining Knowledge Distillation and Contrastive Learning

A Medical Image Segmentation Method Combining Knowledge Distillation and Contrastive Learning

NKUT: Dataset and Benchmark for Pediatric Mandibular Wisdom Teeth Segmentation.

Germectomy is a common surgery in pediatric dentistry to prevent the potential dangers caused by impacted mandibular wisdom teeth. Segmentation of mandibular wisdom teeth is a crucial step in surgery planning. However, manually segmenting teeth and bones from 3D volumes is time-consuming and may cause delays in treatment. Deep learning based medical image segmentation methods have demonstrated the potential to reduce the burden of manual annotations, but they still require a lot of well-annotated data for training. In this paper, we initially curated a Cone Beam Computed Tomography (CBCT) dataset, NKUT, for the segmentation of pediatric mandibular wisdom teeth. This marks the first publicly available dataset in this domain. Second, we propose a semantic separation scale-specific feature fusion network named WTNet, which introduces two branches to address the teeth and bones segmentation tasks. In WTNet, We design a Input Enhancement (IE) block and a Teeth-Bones Feature Separation (TBFS) block to solve the feature confusions and semantic-blur problems in our task. Experimental results suggest that WTNet performs better on NKUT compared to previous state-of-the-art segmentation methods (such as TransUnet), with a maximum DSC lead of nearly 16%.

A Novel Regional‐Minima Image Segmentation Method for Fluid Transport Simulations in Unresolved Rock Images

AbstractUnresolved digital rock images are often used to avoid high computational costs and limited field of views associated with processing fine‐resolution rock images. However, segmentation of unresolved images using classical methods is suboptimal due to the presence of the partial‐volume effect. Suboptimal segmentations can significantly influence the geometry and effective properties of the reconstructed models. This study reveals that partial‐volume pixels with high pore fractions remain as regional minima in intensity levels in unresolved images. By identifying these regional‐minima pixels, we can effectively extract pore space obscured by the partial‐volume effect. Based on this observation, we propose a novel segmentation method capable of identifying these regional‐minima partial‐volume pixels and converting them to pure pore pixels, thereby binarizing the digital rock images. The method is validated on sandstone and carbonate rock samples. Our method demonstrates a notable improvement in modeled permeability accuracy, surpassing 50% compared to the thresholding method and over 30% compared to the watershed method. Moreover, models segmented by this approach exhibit smaller pore and throat sizes compared to the substantially overestimated results obtained by classical methods. These findings suggest that the regional‐minima segmentation method effectively corrects for the partial‐volume effect and preserves more detailed pore structures. Consequently, it enhances the quality of binarized rock geometries, leading to improved accuracy in fluid‐flow simulations.

An encoder-decoder-based image segmentation method for abrasive height detection of diamond wire

An encoder-decoder-based image segmentation method for abrasive height detection of diamond wire

A RGB-Thermal Image Segmentation Method Based on Parameter Sharing and Attention Fusion for Safe Autonomous Driving

A RGB-Thermal Image Segmentation Method Based on Parameter Sharing and Attention Fusion for Safe Autonomous Driving

Optimization-driven artificial intelligence-enhanced municipal waste classification system for disaster waste management

This research addresses the critical challenge of disaster waste management, a growing concern exacerbated by the increasing frequency and intensity of natural disasters like flooding. Traditional waste systems often struggle with the volume and heterogeneity of disaster waste, highlighting the need for innovative solutions. In this study, we present a novel disaster waste classification model integrating advanced artificial intelligence (AI) and optimization techniques to streamline waste categorization in post-disaster environments. Our approach leverages a dual ensemble deep learning framework. The first ensemble combines various image-segmentation methods, while the second integrates outputs from diverse convolutional neural network architectures. A modified artificial multiple intelligence system serves as a decision fusion strategy, enhancing accuracy at both ensemble points. We rigorously evaluated our model using three datasets: the “TrashNet” dataset for benchmarking against existing methods, as well as two meticulously curated, real-world datasets collected from flood-affected areas in Thailand. The results demonstrate that our method outperforms existing algorithms like VGG19, YoloV5, and InceptionV3 in general solid waste classification, achieving an average improvement of 11.18%. Regarding disaster waste specifically, our model achieves 96.48% and 96.49% accuracy on the curated datasets, consistently outperforming ResNet-101, DenseNet-121, and InceptionV3 by an average of 3.47%. These findings demonstrate the potential of our AI-enhanced model to revolutionize disaster waste management practices. Thus, we advocate integrating such technologies into municipal waste management policies to enhance resilience and optimize disaster responses. Future research will explore scaling the model to diverse disaster types and incorporating real-time data for adaptable waste management strategies.

An Improved YOLOv7-Tiny Method for the Segmentation of Images of Vegetable Fields

In response to the limitations of existing methods in differentiating between vegetables and all types of weeds in farmlands, a new image segmentation method is proposed based on the improved YOLOv7-tiny. Building on the original YOLOv7-tiny framework, we replace the CIoU loss function with the WIoU loss function, substitute the Leaky ReLU loss function with the SiLU activation function, introduce the SimAM attention mechanism in the neck network, and integrate the PConv convolution module into the backbone network. The improved YOLOv7-tiny is used for vegetable target detection, while the ExG index, in combination with the OTSU method, is utilized to obtain a foreground image that includes both vegetables and weeds. By integrating the vegetable detection results with the foreground image, a vegetable distribution map is generated. Subsequently, by excluding the vegetable targets from the foreground image using the vegetable distribution map, a single weed target is obtained, thereby achieving accurate segmentation between vegetables and weeds. The experimental results show that the improved YOLOv7-tiny achieves an average precision of 96.5% for vegetable detection, with a frame rate of 89.3 fps, Params of 8.2 M, and FLOPs of 10.9 G, surpassing the original YOLOv7-tiny in both detection accuracy and speed. The image segmentation algorithm achieves a mIoU of 84.8% and an mPA of 97.8%. This method can effectively segment vegetables and a variety of weeds, reduce the complexity of segmentation with good feasibility, and provide a reference for the development of intelligent plant protection robots.

High-Accuracy Image Segmentation Based on Hybrid Attention Mechanism for Sandstone Analysis

Mineral image segmentation based on computer vision is vital to realize automatic mineral analysis. However, current image segmentation methods still cannot effectively solve the problem of sandstone grains that are adjoined and concealed by leaching processes, and the segmentation performance of small and irregular grains still needs to be improved. This investigation explores and designs a Mask R-CNN-based sandstone image segmentation model, including a hybrid attention mechanism, loss function construction, and receptive field enlargement. Simultaneously, we propose a high-quality sandstone dataset with abundant labels named SMISD to facilitate comprehensive training of the model. The experimental results show that the proposed segmentation model has excellent segmentation performance, effectively solving adhesion and overlap between adjacent grains without affecting the classification accuracy. The model has comparable performance to other models on the COCO dataset, and performs better on SMISD than others.

Semi-supervised CT image segmentation via contrastive learning based on entropy constraints.

Deep learning-based methods for fast target segmentation of computed tomography (CT) imaging have become increasingly popular. The success of current deep learning methods usually depends on a large amount of labeled data. Labeling medical data is a time-consuming and laborious task. Therefore, this paper aims to enhance the segmentation of CT images by using a semi-supervised learning method. In order to utilize the valid information in unlabeled data, we design a semi-supervised network model for contrastive learning based on entropy constraints. We use CNN and Transformer to capture the image's local and global feature information, respectively. In addition, the pseudo-labels generated by the teacher networks are unreliable and will lead to degradation of the model performance if they are directly added to the training. Therefore, unreliable samples with high entropy values are discarded to avoid the model extracting the wrong features. In the student network, we also introduce the residual squeeze and excitation module to learn the connection between different channels of each layer feature to obtain better segmentation performance. We demonstrate the effectiveness of the proposed method on the COVID-19 CT public dataset. We mainly considered three evaluation metrics: DSC, HD95, and JC. Compared with several existing state-of-the-art semi-supervised methods, our method improves DSC by 2.3%, JC by 2.5%, and reduces HD95 by 1.9mm.In this paper, a semi-supervised medical image segmentation method is designed by fusing CNN and Transformer and utilizing entropy-constrained contrastive learning loss, which improves the utilization of unlabeled medical images.

Bridge crack segmentation and measurement based on SOLOv2 segmentation model

With the continuous increase of vehicular traffic, the safety caused by bridge crack damage is becoming increasingly prominent. Bridge crack analysis and measurement are of great significance for promoting road traffic safety. However, existing bridge crack image segmentation methods have shortcomings in processing image detail features, resulting in the inability to better measure the actual size of bridge cracks. Therefore, to further optimize the calculation method, a bridge crack image segmentation method based on improved SOLOv2 is designed to achieve more accurate bridge image segmentation. Based on the image segmentation results and combined with the skeleton data extraction method, a bridge crack calculation method is designed. From the results, the segmentation accuracy for crack images was 92.05 % and 93.57 %, respectively. The average mIoU of AM-SOLOv2 method was 0.75, significantly lower than commonly used crack image segmentation methods. In addition, the mIoU value variation amplitude of the AM-SOLOv2 method was relatively smaller. The crack length and width errors were within 0.05 mm and 0.06 mm, significantly lower than the comparison method. It indicates that this method can achieve more accurate crack image segmentation and calculation. This is beneficial for a deeper understanding of the performance degradation and crack damage evolution of bridge structures, thereby improving bridge design and construction technology.

A bi-directional segmentation method for prostate ultrasound images under semantic constraints.

Due to the lack of sufficient labeled data for the prostate and the extensive and complex semantic information in ultrasound images, accurately and quickly segmenting the prostate in transrectal ultrasound (TRUS) images remains a challenging task. In this context, this paper proposes a solution for TRUS image segmentation using an end-to-end bidirectional semantic constraint method, namely the BiSeC model. The experimental results show that compared with classic or popular deep learning methods, this method has better segmentation performance, with the Dice Similarity Coefficient (DSC) of 96.74% and the Intersection over Union (IoU) of 93.71%. Our model achieves a good balance between actual boundaries and noise areas, reducing costs while ensuring the accuracy and speed of segmentation.

Non-destructive weight prediction model of spherical fruits and vegetables using U-Net image segmentation and machine learning methods

Artificial intelligence has become increasingly prominent in agriculture and other fields. Prediction of body weight in animals and plants has been done by humans using many different methods and observations from the past to the present. Although there has been extensive research on predicting the live body weight of animals, weight prediction of vegetables and fruits is not widely. As spherical or round-shaped fruits and vegetables are sold by weighing in the fields, markets and greengrocers, it is important to make weight predictions. Based on this, a model was developed to predict the weight of fruits and vegetables such as watermelons, melons, apples, oranges and tomatoes with the data obtained from their images. The fruit and vegetable weights were predicted by regression models using data obtained from images segmented by the U-net architecture. Machine learning models such as Multi-Layer Perceptron (MLP), Random Forest (RF), Decision Trees (DT), Support Vector Machines (SVM), Linear and Stochastic Gradient Descent (SDG) regression models were used for weight predictions. The most effective regression models are the Random Forest and Decision Trees models. For regression training, the best success rates were calculated as 0.9112 for watermelon, 0.9944 for apple, 0.9989 for tomato and 0.9996 for orange. In addition, the results were evaluated by comparing to the studies of weight prediction. The weight prediction model will help to sell round-shaped fruits and vegetables in the fields, markets and gardens using the weights predictions from the images. It is also a guideline for studies that follow the growth of fruit and vegetables according to their weight.

A Hybrid CNN-TransXNet Approach for Advanced Glomerular Segmentation in Renal Histology Imaging

In the specialized field of renal histology, precise segmentation of glomeruli in microscopic images is crucial for accurate clinical diagnosis and pathological analysis. Facing the challenge of discerning complex visual features, such as shape, texture, and size within these images, we introduce a novel segmentation model that innovatively combines convolutional neural networks (CNNs) with the advanced TransXNet block, specifically tailored for glomerular segmentation. This innovative model is designed to capture the intricate details and broader contextual features within the images, ensuring a comprehensive and precise segmentation process. The model's architecture unfolds in two primary phases: the down-sampling phase, which utilizes CNNs structures within the TransXNet block for meticulous extraction of detailed features, and the up-sampling phase, which employs CNNs deconvolution techniques to restore spatial resolution and enhance macroscopic feature representation. A critical innovation in our model is the implementation of residual connections between these two phases, which facilitate the seamless integration of features and minimize loss of precision during image reconstruction. Experimental results demonstrate a significant improvement in our model’s performance compared to existing medical image segmentation methods. We report enhancements in mean Pixel Accuracy (mPA) and mean Intersection over Union (mIoU), with increases of approximately 3–5% and 3–8%, respectively. Additionally, the segmented outputs exhibit higher subjective visual quality with fewer noise artifacts. These findings suggest that our model offers promising applications in the segmentation of medical microscopic images, marking a significant contribution to the domain.