Existing Image Segmentation Methods Research Articles

PalScDiff: A diffusion-based framework with progressive augmentation learning and semantic consistency for breast ultrasound tumor segmentation

Background: Breast cancer diagnosis relies on accurate lesion segmentation in medical images. Automated computer-aided diagnosis reduces clinician workload and improves efficiency, but existing image segmentation methods face challenges in model performance and generalization. Objective: This study aims to develop a generative framework using a denoising diffusion model for efficient and accurate breast cancer lesion segmentation in medical images. Methods: We design a novel generative framework, PalScDiff, that leverages a denoising diffusion probabilistic model to reconstruct the label distribution for medical images, thereby enabling the sampling of diverse, plausible segmentation outcomes. Specifically, with the condition of the corresponding image, PalScDiff learns to estimate the masses region probability through denoising step by step. Furthermore, we design a Progressive Augmentation Learning strategy to incrementally handle segmentation challenges of irregular and blurred tumors. Moreover, multi-round sampling is employed to achieve robust breast mass segmentation. Results: Our experimental results show that PalScDiff outperforms established models such as U-Net and transformer-based alternatives, achieving an accuracy of 95.15%, precision of 79.74%, Dice coefficient of 77.61%, and Intersection over Union (IOU) of 81.51% . Conclusion: The proposed model demonstrates promising capabilities for accurate and efficient computer-aided segmentation of breast cancer.

FishSegSSL: A Semi-Supervised Semantic Segmentation Framework for Fish-Eye Images.

The application of large field-of-view (FoV) cameras equipped with fish-eye lenses brings notable advantages to various real-world computer vision applications, including autonomous driving. While deep learning has proven successful in conventional computer vision applications using regular perspective images, its potential in fish-eye camera contexts remains largely unexplored due to limited datasets for fully supervised learning. Semi-supervised learning comes as a potential solution to manage this challenge. In this study, we explore and benchmark two popular semi-supervised methods from the perspective image domain for fish-eye image segmentation. We further introduce FishSegSSL, a novel fish-eye image segmentation framework featuring three semi-supervised components: pseudo-label filtering, dynamic confidence thresholding, and robust strong augmentation. Evaluation on the WoodScape dataset, collected from vehicle-mounted fish-eye cameras, demonstrates that our proposed method enhances the model's performance by up to 10.49% over fully supervised methods using the same amount of labeled data. Our method also improves the existing image segmentation methods by 2.34%. To the best of our knowledge, this is the first work on semi-supervised semantic segmentation on fish-eye images. Additionally, we conduct a comprehensive ablation study and sensitivity analysis to showcase the efficacy of each proposed method in this research.

A multiangle polarised imaging-based method for thin sectionsegmentation.

The most crucial task of petroleum geology is to explore oil and gas reservoirs in the deep underground. As one of the analysis techniques in petroleum geological research, rock thin sectionidentification method includes particle segmentation, which is one of the key steps. A conventional sandstone thin sectionimage typically contains hundreds of mineral particles with blurred boundaries and complex microstructures inside the particles. Moreover, the complex lithology and low porosity of tight sandstone make traditional image segmentation methods unsuitable for solving the complex thin sectionsegmentation problems. This paper combines petrology and image processing technologies. First, polarised sequence images are aligned, and then the images are transformed to the HSV colour space to extract pores. Second, particles are extracted according to their extinction characteristics. Last, a concavity and corner detection matching method is used to process the extracted particles, thereby completing the segmentation of sandstone thin sectionimages. The experimental results show that our proposed method can more accurately fit the boundaries of mineral particles in sandstone images than existing image segmentation methods. Additionally, when applied in actual production scenarios, our method exhibits excellent performance, greatly improving thin sectionidentification efficiency and significantly assisting experts inidentification.

Sequential Active Contour Based on Morphological-Driven Thresholding for Ultrasound Image Segmentation of Ascites.

Paracentesis is a high-demanding and routine operation, which has great potentials and benefits if semi-autonomous procedures can be developed. One of the most important techniques that facilitate semi-autonomous paracentesis is to segment the ascites from ultrasound images accurately and efficiently. The ascites, however, is usually with significantly different shapes and noise among different patients, and its shape/size changes dynamically during the paracentesis. This makes most of existing image segmentation methods either time consuming or inaccurate for segmenting ascites from its background. In this article, we propose a two-stage active contour method to facilitate accurate and efficient segmentation of ascites. First, a morphological-driven thresholding method is developed to locate the initial contour of the ascites automatically. Then, the identified initial contour is fed into a novel sequential active contour algorithm to segment the ascites from background accurately. The proposed method is tested and compared with state-of-the-art active contour methods on over 100 real ultrasound images of ascites, and the results show the superiority of our method in both accuracy and time efficiency.

Edge Detection With Direction Guided Postprocessing for Farmland Parcel Extraction

Farmland is a significant resource for human survival and development. Rapid acquisition of farmland information is the basis for dynamic crop detection and sustainable land development. The continuous development of high resolution remote sensing imagery makes it possible to make a wide range of refined earth observation. With better image interpretation ability, image segmentation method based on deep learning can bring specific results from high resolution imagery and is widely used in remote sensing. However, existing image segmentation methods based on semantic segmentation have difficulties to extracting refined farmland parcels. Deep neural network is used to detect farmland edge. We use HRNet to achieve feature extraction that retains high resolution features, strengthens the feature representation of network context information based on Object-Contextual Representations module, and carries out more complete interpretation of farmland and its boundary. Finally, we design a farmland edge post-processing method to connect the disconnected boundary based on the direction information generated by the connectivity attention module, and finally obtained the farmland boundary which is complete enough to be closed for generating farmland parcels. To verify our method, we used Google Earth image to label farmland boundaries and conduct experimental verification. The results show that our proposed model has a higher precision for farmland edge detection, and the postprocessing method of boundary connection can effectively close the boundary lines and achieve more detailed and complete farmland parcels.

A Double Random Matrix Design Model for Fractal Art Patterns Based on Visual Characteristics

This paper adopts the method of visual characteristics of the double random matrix to conduct in-depth research and analysis on the design of fractal art patterns. For the practical application needs in fractal graphic design, a method is proposed to automatically extract the core base pattern based on fractal graphic content and generate a four-sided continuous pattern. The method first uses the Canny operator for edge detection to analyze the area of the main pattern. Then, it uses a grayscale cogeneration matrix to extract and analyze the graphic texture features, based on which the best splicing method is selected to splice the extracted pattern, and then, it achieves two splicing methods of flat row and staggered four-sided continuous pattern. The method has good practicality with low complexity and high versatility under the premise of ensuring the beauty of the generated four-sided continuous pattern. It can assist designers to design patterns, improve efficiency, and save design costs. In this paper, we improve the existing image segmentation methods, adopt two segmentation methods, namely quadratic tree segmentation and HV segmentation, propose a new local codebook selection strategy, and study the degree of self-adaptation of different methods to images in terms of segmentation methods and local codebook selection strategies. It makes the network unable to train efficiently on long text prediction problems. Finally, the improved algorithm of this paper is tested on the standard and living image libraries, and the experimental results show that the use of local codebooks makes the image coding speed significantly improved compared with the fixed fractal. In the face of the image to be retrieved, it is only necessary to perform the coding operation to obtain the fractal code to perform similarity matching, which can meet the requirement of real-time retrieval. Applying the improved distance formula, the search accuracy obtained on the test gallery is significantly better than that of the grayscale histogram algorithm.

A Dilated Segmentation Network with the Morphological Correction Method in Farming Area Image Series

Farming areas are made up of diverse land use types, such as arable lands, grasslands, woodlands, water bodies, and other surrounding agricultural architectures. They possess imperative economic value, and are considerably valued in terms of farmers’ livelihoods and society’s flourishment. Meanwhile, detecting crops in farming areas, such as wheat and corn, allows for more direct monitoring of farming area production and is significant for practical production and management. However, existing image segmentation methods are relatively homogeneous, with insufficient ability to segment multiple objects around the agricultural environment and small-scale objects such as corn and wheat. Motivated by these issues, this paper proposed a global-transformer segmentation network based on the morphological correction method. In addition, we applied the dilated convolution technique to the backbone of the model and the transformer technique to the branches. This innovation of integrating the above-mentioned techniques has an active impact on the segmentation of small-scale objects. Subsequently, the backbone improved by this method was applied to an object detection network based on a corn and wheat ears dataset. Experimental results reveal that our model can effectively detect wheat ears in a complicated environment. For two particular segmentation objects in farming areas, namely water bodies and roads, we notably proposed a morphological correction method, which effectively reduces the number of connected domains in the segmentation results with different parameters of dilation and erosion operations. The segmentation results of water bodies and roads were thereby improved. The proposed method achieved 0.903 and 13 for mIoU and continuity. This result reveals a remarkable improvement compared with the comparison model, and the continuity has risen by 408%. These comparative results demonstrate that the proposed method is eminent and robust enough to provide preliminary preparations and viable strategies for managing farming area resources and detecting crops.

DeepProfile: Accurate Under-the-Clothes Body Profile Estimation

Accurate human body profiles have many potential applications. Image-based human body profile estimation can be regarded as a fine-grained semantic segmentation problem, which is typically used to locate objects and boundaries in images. However, existing image segmentation methods, such as human parsing, require significant amounts of annotation and their datasets consider clothes as part of the human body profile. Therefore, the results they generate are not accurate when the human subject is dressed in loose-fitting clothing. In this paper, we created and labeled an under-the-clothes human body contour keypoint dataset; we utilized a convolutional neural network (CNN) to extract the contour keypoints, then combined them with a body profile database to generate under-the-clothes profiles. In order to improve the precision of keypoint detection, we propose a short-skip multi-scale dense (SMSD) block in the CNN to keep the details of the image and increase the information flow among different layers. Extensive experiments were conducted to show the effectiveness of our method. We demonstrate that our method achieved better results—especially when the person was dressed in loose-fitting clothes—than and competitive quantitative performance compared to state-of-the-art methods, while requiring less annotation effort. We also extended our method to the applications of 3D human model reconstruction and body size measurement.

A Flexible Region of Interest Extraction Algorithm with Adaptive Threshold for 3-D Synthetic Aperture Radar Images

Most of the existing image segmentation methods have a strong anti-noise ability but are susceptible to the interference in the background, so they are not suitable for 3-D synthetic aperture radar (SAR) image target extraction. Region of interest (ROI) extraction can improve the anti-interference ability of the image segmentation methods. However, the existing ROI extraction method uses the same threshold to process all the images in the data set. This threshold is not optimal for each image. Designed for 3-D SAR image target extraction, we propose an ROI extraction algorithm with adaptive threshold (REAT) to enhance the anti-interference ability of the existing image segmentation methods. The required thresholds in the proposed algorithm are adaptively obtained by the mapping of the image features. Moreover, the proposed algorithm can easily be applied to existing image segmentation methods. The experiments demonstrate that the proposed algorithm significantly enhances the anti-interference ability and computational efficiency of the image segmentation methods. Compared with the existing ROI extraction algorithm, the proposed algorithm improves the dice similarity coefficient by 6.4%.

Efficient image segmentation based on deep learning for mineral image classification

Mineral image segmentation plays a vital role in the realization of machine vision based intelligent ore sorting equipment. However, the existing image segmentation methods still cannot effectively solve the problem of adhesion and overlap between mineral particles, and the segmentation performance of small and irregular particles still needs to be improved. To overcome these bottlenecks, we propose a deep learning based image segmentation method to segment the key areas in mineral images using morphological transformation to process mineral image masks. This investigation explores four aspects of the deep learning-based mineral image segmentation model, including backbone selection, module configuration, loss function construction, and its application in mineral image classification. Specifically, referring to the designs of U-Net, FCN, Seg Net, PSP Net, and DeepLab Net, this experiment uses different backbones as Encoder to building ten mineral image segmentation models with different layers, structures, and sampling methods. Simultaneously, we propose a new loss function suitable for mineral image segmentation and compare CNNs-based segmentation models' training performance under different loss functions. The experiment results show that the proposed mineral image segmentation has excellent segmentation performance, effectively solves adhesion and overlap between adjacent particles without affecting the classification accuracy. By using the Mobile Net as backbone, the PSP Net and DeepLab can achieve a high segmentation performance in mineral image segmentation tasks, and the 15 × 15 is the most suitable size for erosion element structure to process the mask images of the segmentation models.

Machine vision on the positioning accuracy evaluation of poultry viscera in the automatic evisceration robot system

ABSTRACT In the poultry slaughtering, accurate viscera positioning is essential to reduce the damage of internal organs. The introduction of machine vision technique can help to locate the viscera and can provide new direction for the poultry evisceration. After a midline abdominal incision of the poultry, the internal organs are taken out from the poultry that placed on the conveyor line by the parallel robot. Based on machine vision, the recognition accuracy of opened poultry viscera directly affects the level of visceral damage and residue caused by gripping manipulator. However, visceral positioning is often influenced by different noise in the abdominal cavity, such as mucous membranes and blood stains. Thus, the image segmentation of poultry viscera is a complex process, and it is challenging to remove the noise. In general, existing image segmentation methods can hardly segment visceral regions well. To strengthen the anti-noise ability, we proposed an improved region-based active contour method with the level-set formulation. This method combined with several operations of color space conversion and top-low-hat transformation, which could extract the viscera contour and precisely removed the noise. The results showed that recognition accuracy of the heart-liver area and fat area in the viscera are 98.98% and 99.75%, respectively, while the overall viscera for poultry was 98.96%. The results of this experiment suggested that the proposed image segmentation algorithm could achieve the required accuracy for poultry viscera detection. Thus, the proposed visceral contour recognition method could be applied in poultry processing, providing critical information to guide the robot for automated evisceration.

Image Segmentation With Language Referring Expression and Comprehension

Image segmentation with language referring expression can complete object segmentation based on expression text. Existing image segmentation methods show good results on high-performance computers, but most robot systems need real-time and high accuracy. At present, most methods cannot meet these requirements well. Therefore, we propose a high-precision and real-time deep learning network that integrates the two tasks of image segmentation with language referring expression and referring expression comprehension and then treats them as two branches. Specifically, the proposed network first merges the two tasks. The feature maps of different scales extracted by each branch are fed back to the two branches to obtain prediction results. These two tasks promote and restrict each other. Experiments show that our method has better real-time performance and higher accuracy than existing methods.

A Software System for Processing Images with Parallel Computing

In this paper, a software system for image processing with parallel computing based on the geometrized histograms method developed for the concise description and segmentation of color images and for designing real-time image understanding systems is described. The parallel processing leans on the fact that, in contrast to the majority of the existing image segmentation methods, the proposed method is designed so that the most labor-consuming operations with the pixel array can be executed using n independent threads. The principles of designing programs for data processing in separate threads in which the program produces a substantial, compressed description of an image (a frame of a video sequence) that preserves the geometrical relations of the source image but has a dimension by several orders of magnitude less than the original image are described. The main operations of the segmentation and image understanding systems are executed without using the image pixel array—only using the designed concise description. These operations require a short execution time (on the average less than 10 ms for the whole set of tasks) on standard modern personal computers, even for HD video. In this paper, a multithreaded implementation of constructing a concise description of an image (a frame) is considered that allows one to enhance the operation speed, which is already fairly high, up to the record productivity figures. The application to systems for understanding road scenes, such as systems for finding the road region, its roadsides, the sky region, to systems for detecting and understanding road markings (permanent white and temporary colored), as well as to finding signal lamps of helicopters, are also described. Examples of processing results for particular road scenes are presented and discussed, and estimates of the operation speed for video sequences of real road scenes are given.

Deep-reinforcement-learning-based images segmentation for quantitative analysis of gold immunochromatographic strip

Gold immunochromatographic strip (GICS) is a widely used lateral flow immunoassay technique. A novel image segmentation method is developed in this paper for quantitative analysis of GICS based on the deep reinforcement learning (DRL), which can accurately distinguish the test line and the control line in the GICS images. The deep belief network (DBN) is employed in the deep Q network in our DRL algorithm. Meanwhile, the multi-factor learning curve is introduced in the DRL algorithm to dynamically adjust the capacity of the replay buffer and the sampling size, which leads to enhanced learning efficiency. It is worth mentioning that the states, actions, and rewards in the developed DRL algorithm are determined based on the characteristics of GICS images. Experiment results demonstrate the feasibility and reliability of the proposed DRL-based image segmentation method and show that the proposed new image segmentation method outperforms some existing image segmentation methods for quantitative analysis of GICS images.

A variational image segmentation method exploring both intensity means and texture patterns

Existing image segmentation methods usually consider only intensity or texture features, and thus cannot work well on images with both varying intensity and salient texture patterns. This paper aims to solve this problem by exploring both intensity and texture features for better image segmentation. To this end, we propose a novel variational image segmentation model, which uses constant vectors to represent intensity means in different image subregions and structured dictionaries to encode local texture patterns in images. We develop an effective algorithm to implement the proposed variational model by seeking for intensity means and a specific dictionary for the input image and meanwhile computing the representation coefficients of image patches with respect to the learned dictionary. We derive closed-form solutions for all these components, which result in the high efficiency of our proposed method. Extensive evaluation experiments with comparison to existing methods have been done on both synthetic and real-world images. The results demonstrate the superiority of our proposed method in effectively segmenting images with both unsmooth intensity variations and salient texture patterns.

Segmentation of Sonar Image on Seafloor Sediments Based on Multiclass SVM

ABSTRACT Chen, J. and Zhang, S., 2018. Segmentation of sonar image on seafloor sediments based on multiclass SVM. In: Liu, Z.L. and Mi, C. (eds.), Advances in Sustainable Port and Ocean Engineering. Journal of Coastal Research, Special Issue No. 83, pp. 597–602. Coconut Creek (Florida), ISSN 0749-0208. Because the seabed sediment sonar image has the complex background and low noise-signal ratio, the existing image segmentation methods are difficult to extract different types of sonar image features from the complex background of seabed sediment sonar image. The image segmentation accuracy is low. Thus, this paper presents a method for segmenting seabed sediment sonar image based on multiclass SVM. Firstly, the seabed sediment sonar image was denoised to extract the statistical feature, texture feature and gray feature with phase consistency. Then they were composed into characteristic sample vectors of multidimensional seabed sediment sonar image. Finally, SVM classification method was used to segment the...

Co-optimization of ethnic-pattern segmentation based on hierarchical patch matching

The segmentation of ethnic patterns is vital for digital analyses of ethnic cultures. Although existing image segmentation methods can properly segment natural images, they cannot preserve the main structure of the image. They also require many user interactions to segment ethnic patterns. This paper presents a co-optimization method that segments ethnic patterns using hierarchical patch matching. Exploiting the repetitive characteristics of ethnic patterns, the method first automatically detects all similar patterns using a global patch match. Second, the relative orientation between similar patterns is estimated by a local patch match, and an accurate dense correspondence is constructed by a constrained patch match. Finally, the pre-segmentation of ethnic patterns is co-optimized to preserve their main structures. Our method can segment all similar ethnic patterns into separate elements with dense correspondence. Besides reducing the user interaction and improving the segmentation accuracy, the proposed method improves the quality of ethnic pattern digital analysis such as vectorization. Experiments demonstrated the validity of our method.

Improved Morphological Band-Pass Filtering Algorithm and Its Application in Circle Detection

Existing image segmentation and image enhancement methods are deficient in complex industrial environments. Therefore, an improved morphological band-pass filter algorithm is presented. The first step of the algorithm is obtaining two marker images by erosion operations for a test image with two kinds of structuring elements: one slightly larger and one smaller than the feature but similar in shape. The second step is obtaining an image only, including the background, and an image including the feature with the background, excluding noise. The final step is realizing the feature segmentation by carrying out a difference operation on the two images. Selection of the structuring elements in the algorithm and the computational cost reduction are also discussed for engineering applications. Experimental results show that the proposed algorithm achieves the accurate segmentation of the circle at a specific scale through the velocity-optimized morphological operation and features good real-time performance and high accuracy in complex industrial environments, which could meet the requirements of industrial online monitoring.

Wavelet based transition region extraction for image segmentation

Transition region based approaches are recent hybrid segmentation techniques well known for its simplicity and effectiveness. Here, the segmentation effectiveness depends on robust extraction of transition regions. So, we have proposed a transition region method which initially decomposes the gray image in wavelet domain. Two existing transition region approaches are applied on approximate coefficients to extract transition region feature matrix. Using this feature matrix the corresponding prominent wavelet coefficients of different bands are found. Inverse wavelet transform are then applied on the modified coefficients to get edge image with more than one pixel width. Otsu thresholding is applied on it to get transition regions. Further, morphological operations are applied to extract the object regions. Finally, the objects are extracted using the object regions. The wavelet domain approach extracts robust transition regions resulting in better segmentation. The proposed method is compared with different existing image segmentation methods. Experimental results reveal that the proposed method achieve 0.95 overall segmentation accuracy. It also works well for extraction of single as well as multiple objects from images.

Efficient numerical schemes for Chan-Vese active contour models in image segmentation

In this paper, we introduce multi-symplectic Lagrangian variational integrators for solving Chan-Vese active contour models in image segmentation. Energy functionals are discretized firstly, and numerical schemes are derived from discrete Euler-Lagrange equations based on discrete variational principle. Lagrangian variational integrators preserve native differential structure-multi-symplecticity, that makes the numerical methods have a satisfied behavior. Experiments are performed on the benchmark images from literature. We further evaluated the methods in a segmentation database containing 1023 images. It shows that the proposed numerical schemes attain relatively faster convergence rates and better segmentation accuracy. Comparisons with the standard explicit Euler method of the original Chan-Vese model and other fast numerical optimization methods show that the proposed methods have better stability, higher accuracy, and are more robust when dealing with a large number of pictures. This study provides an example for further research to improve the performance of other existing image segmentation methods based on active contour models.