One-Class SVM Based Segmentation for SAR Image

This paper proposes a new texture image retrieval scheme based on contourlet transform and support vector machines (SVMs). In the scheme, the energies and the generalized Gaussian distribution (GGD) parameters are used to represent the contourlet subband features. Using the representations, a two-run SVM retrieval algorithm which employs an one-class SVM followed by a two-class SVM is proposed to carry out the perceptual similarity measurement. For the query image, the one-class SVM is used to obtain the effective initial training set with positive and negative samples. Using these initial samples, the two-class SVM is applied to refine on the image classification subject to the user’s relevance feedback. Compared with existing texture image retrieval methods, the proposed retrieval scheme is demonstrated respectively to be effective on the VisTex database of 640 texture images and the Brodatz database of 1760 texture images. Experimental results have shown that the proposed retrieval scheme can attain 99.38% and 98.07% of the average rates respectively for the two databases.KeywordsContourlet transformgeneralized Gaussian distributionsupport vector machinerelevance feedbacktexture image retrieval

The traditional template matching strategy of optical and synthetic aperture radar (SAR) is sensitive to the nonlinear transformation between two images. In some cases, the optical and SAR image pairs do not conform to the affine transformation condition. To address this issue, this study presents a novel template matching strategy which uses the One-Class Support Vector Machine (SVM) to remove outliers. First, we propose a method to construct the similarity map dataset using the SEN1-2 dataset for training the One-Class SVM. Second, a four-step strategy for optical and SAR image registration is presented in this paper. In the first step, the optical image is divided into some grids. In the second step, the strongest Harris response point is selected as the feature point in each grid. In the third step, we use Gaussian pyramid features of oriented gradients (GPOG) descriptor to calculate the similarity map in the search region. The trained One-Class SVM is used to remove outliers through similarity maps in the fourth step. Furthermore, the number of improve matches (NIM) and the rate of improve matches (RIM) are designed to measure the effect of image registration. Finally, this paper designs two experiments to prove that the proposed strategy can correctly select the matching points through similarity maps. The experimental results of the One-Class SVM in dataset show that the One-Class SVM can select the correct points in different datasets. The image registration results obtained on the second experiment show that the proposed strategy is robust to the nonlinear transformation between optical and SAR images.

Image segmentation is a key step in the application of Synthetic Aperture Radar (SAR) images, but because of the existing of speckles in SAR images, image can not be divided well by using traditional methods. According to the remarkable result of wavelet transform on texture feature extraction, image filtration and the advantages of support vector machine (SVM) classification, an efficient method of SAR image segmentation is proposed. First, extracting texture feature of sample points by wavelet transform. Second, image preprocessing is performed by using wavelet filtering method. The vector composed of wavelet energy feature, weighted mean value of wavelet energy feature and gray values of eight-neighborhood of filtered SAR image. Then a SVM classifier is designed and trained by using normalized feature vectors. Finally, the testing sets of SAR images are divided by trained SVM. In the experiments of SAR image segmentation, better results have been obtained with this new method.

The SAR image semantic segmentation using computer vision techniques has gained much popularity in the research community due to their wide applications. Despite the advancements in Deep Learning for image analysis, these models still struggle to segment SAR images due to the existence of speckle noise and a poor feature extractor. Moreover, deep learning models are challenging to train on small datasets and the performance of the model is significantly impacted by the quality of the data. This calls for the development of an effective network that can draw out critical information from the low resolution SAR images. In this regard, the present work proposes a unique Self attention module in U-Net for the semantic segmentation of low resolution SAR images.. The Self Attention Model makes use of Laplacian kernel to highlight the sharp discontinuities in the features that define the boundaries of the objects. The proposed model, employs dilated convolution layers at the initial layers, enabling the model to more effectively capture larger contextual information. With an accuracy of 0.84 and an F1-score of 0.83, the proposed model outperforms the state-of-the-art techniques in semantic segmentation of low resolution SAR images. The results clearly demonstrate the importance of the self attention module and the consideration of dilated convolution layers in the initial layers in semantic segmentation of low resolution SAR images.

Oceanic internal waves play a crucial role in ocean activities. Currently, the approach to detecting oceanic internal waves from synthetic aperture radar (SAR) images is becoming robust. To efficiently identify the stripes of oceanic internal waves from SAR images, we propose an integrated algorithm for the detection and recognition of oceanic internal waves. First, the Gamma Map filtering method was adopted to reduce speckle noise in the SAR images. Then, histogram of orientated gradients (HOG), Grey level co-occurrence matrix (GLCM), and fractal dimension (FD) were utilized to extract the image features. Subsequently, support vector machine (SVM) was adopted to classify the SAR images and obtain images that contain oceanic internal waves. Next, the Canny edge detection method was used to detect and recognize the stripes of oceanic internal waves in the SAR images, and these stripes were screened by three parameters, namely their lengths, area ratios, and directions. Finally, the positions of the stripes of the oceanic internal waves were obtained. The experimental results verify that the proposed method can identify whether SAR images contain oceanic internal waves, and also determine the locations of their stripes in the SAR images. Meanwhile, the algorithm exhibits reasonable robustness and recognition rate. In addition, the optimal accuracy and kappa coefficient () are 94.2% and 0.878, respectively.

Star trackers are currently the most accurate sensors for determining the attitude of a spacecraft. These sensors comprise not only highly capable optical detectors and processor units but also complicated software solvers. One of the main solvers employed in star trackers is image segmentation. In this study, the aim is to develop a hybrid image segmentation method which is a combination of both global thresholding and density-based spatial clustering of applications with noise (DBSCAN) method to increase detection probability of the stars in heavy noise. Secondly, a sorting algorithm is added to list the detected stars in terms of their brightness to increase the efficiency of the star tracking algorithm. Then, this new approach and two different conventional segmentation methods are applied to the Orion star constellation image polluted with Gaussian, salt and pepper, and uneven background noises. The resulting images of these segmentation methods are compared in terms of denoising capabilities. Although computationally more expensive, the proposed DBSCAN-based hybrid method displays a background pixel recovery performance of 99.5%, compared to Otsu global thresholding and adaptive thresholding methods’ 73.5% and 79.9% recovery values, respectively. Additionally, it has been demonstrated that the sorting algorithm successfully listed the detected stars in accordance with their brightness. Cite this article as: N. Sengil, “Implementation of DBSCAN method in star trackers to improve image segmentation in heavy noise conditions,” Electrica., 23(1), 3-10, 2023.

A novel hybrid undersampling method for mining unbalanced datasets in banking and insurance

A two-class support vector machine (SVM)-based image segmentation approach has been developed for the extraction of nasopharyngeal carcinoma (NPC) lesion from magnetic resonance (MR) images. By exploring two-class SVM, the developed method can learn the actual distribution of image data without prior knowledge and draw an optimal hyperplane for class separation, via an SVM parameters training procedure and an implicit kernel mapping. After learning, segmentation task is performed by the trained SVM classifier. The proposed technique is evaluated by 39 MR images with NPC and the results suggest that the proposed query-based approach provides an effective method for NPC extraction from MR images with high accuracy.

Image segmentation of brain low-grade glioma (LGG) magnetic resonance imaging (MRI) contributes tremendously to diagnosis, classification and treatment of the disease. A tangible, accurate, reliable and fast image segmentation technique is demanded in clinical diagnosis and research. The emerging machine learning technique has been demonstrated its unique capability in the field of medical image processing, including medical image segmentation. Support vector machine (SVM) and convolutional neural network (CNN) are two widely used machine learning methods. In this work, image segmentation tools based on SVM and CNN are developed and evaluated for brain LGG MR image segmentation studies. The segmentation performance in terms of accuracy and cost is quantitatively analyzed and compared between the SVM and CNN techniques developed. Computed on the Google CoLab, each of the 109 SVM models represents an individual patient, is trained using a single image of that patient and takes a few seconds to complete. The CNN model is trained on a drastically larger dataset of 19,760 data augmented images and takes approximately 2 hours to obtain the most optimal result. The SVM models achieved an average and median accuracy of 0.937 and 0.976 respectively, precision of 0.456 and 0.535 respectively, recall of 0.878 and 0.906 respectively, and F1 score of 0.546 and 0.662 respectively. Although the CNN model required a significantly longer calculation time, it surpassed the SVM models in performance in LGG MR image segmentation, achieving an accuracy of 0.998, a precision of 0.999, a recall of 0.999 and an F1 score of 0.999. This study shows that SVM with appropriate filtering techniques is capable of obtaining reliable and fast segmentation of brain LGG MR images with sufficient accuracy and limited image data. CNN technique outperforms SVM in the accuracy of segmentation with requirements of significantly enlarged data set, long computation time and high-performance computer.

Synthetic aperture radar (SAR) image segmentation usually involves two crucial issues: suitable speckle noise removing technique and effective image segmentation methodology. Here, an efficient SAR image segmentation method considering both of the two aspects is presented. As for the first issue, the famous nonlocal mean (NLM) filter is introduced in this study to suppress the multiplicative speckle noise in SAR image. Furthermore, to achieve a higher denoising accuracy, the local neighboring pixels in the searching window are projected into a lower dimensional subspace by principal component analysis (PCA). Thus, the nonlocal mean filter is implemented in the subspace. Afterwards, a multi-objective clustering algorithm is proposed using the principals of artificial immune system (AIS) and kernel-induced distance measures. The multi-objective clustering has been shown to discover the data distribution with different characteristics and the kernel methods can improve its robustness to noise and outliers. Experiments demonstrate that the proposed method is able to partition the SAR image robustly and accurately than the conventional approaches.

Synthetic aperture radar (SAR) image segmentation is the key to SAR image automatic interpretation. However, speckle noise, intensity inhomogeneity, and irregular shaped objects with changing edge often make the SAR image segmentation very difficult, and existing algorithms have high computational complexity. We propose a region-based level set method using the local image intensity information. To represent the statistical characteristics of speckle noise, we first use a gamma statistical distribution to model every segmented SAR image. We then apply a modified region mean estimation formula to efficiently segment SAR images with inhomogeneity. Finally, Gaussian filtering is employed to regularize the level set function, which can avoid reinitialization. The experimental results on synthetic and real-world SAR images demonstrate that the proposed method has less computation cost, faster convergence rate, and more accurate segmentation results.

Prostate cancer is one of the most frequent cancers in men and is a major cause of mortality in the most of countries. In many diagnostic and treatment procedures for prostate disease accurate detection of prostate boundaries in transrectal ultrasound(TRUS) images is required. This is a challenging and difficult task due to weak prostate boundaries, speckle noise and the short range of gray levels. In this paper a method for automatic prostate segmentation in TRUS images using Gabor feature extraction and snake-like contour is presented. This method involves preprocessing, extracting Gabor feature, training, and prostate segmentation. The speckle reduction for preprocessing step has been achieved by using stick filter and top-hat transform has been implemented for smoothing the contour. A Gabor filter bank for extraction of rotation- invariant texture features has been implemented. A support vector machine(SVM) for training step has been used to get each feature of prostate and nonprostate. Finally, the boundary of prostate is extracted by the snake-like contour algorithm. A number of experiments are conducted to validate this method and results showed that this new algorithm extracted the prostate boundary with less than 10.2% of the accuracy which is relative to boundary provided manually by experts Keywords—Gabor Filter Bank, Support Vector Machines, Prostate Segmentation

Cyber-security threats are a growing concern in networked environments. The development of Intrusion Detection Systems (IDSs) is fundamental in order to provide extra level of security. We have developed an unsupervised anomaly-based IDS that uses statistical techniques to conduct the detection process. Despite providing many advantages, anomaly-based IDSs tend to generate a high number of false alarms. Machine Learning (ML) techniques have gained wide interest in tasks of intrusion detection. In this work, Support Vector Machine (SVM) is deemed as an ML technique that could complement the performance of our IDS, providing a second line of detection to reduce the number of false alarms, or as an alternative detection technique. We assess the performance of our IDS against one-class and two-class SVMs, using linear and non- linear forms. The results that we present show that linear two-class SVM generates highly accurate results, and the accuracy of the linear one-class SVM is very comparable, and it does not need training datasets associated with malicious data. Similarly, the results evidence that our IDS could benefit from the use of ML techniques to increase its accuracy when analysing datasets comprising of non- homogeneous features.

SAR image segmentation is a key procedure for target extraction, classification, recognition and surveillance. In this paper, firstly, we propose a multiscale likelihood ratio test (MLRT) for SAR image segmentation. MLRT can enhance the distinction between different signals by fusing their all kinds of features from different scales. And in our application, the multiscale features are from SAR multiscale sequence images. Secondly, in order to satisfy the sample dependence condition of maximum likelihood estimation, and improve the computation efficiency greatly, we employed Bootstrap sampling technique to make all the statistical parameters be estimated in the best independent conditions. Thirdly, a precise segmentation results of SAR image can be obtained based on MLRT, where the parameter can be estimated by maximum likelihood estimation and MLRT histogram of multiscale sequence images base on Bootstrap sample. Finally, correct rates of simulated images segmentation are used to test our method, and true SAR images are segmented to demonstrate the validity.

Resting state functional magnetic resonance imaging (fMRI) aims to measure baseline neuronal connectivity independent of functional tasks. Most existing network detection methods rely on a fixed threshold to identify functionally connected voxels under the resting state. Due to fMRI non-stationarity, a fixed threshold cannot adapt to inter-session and inter-subject variation. In this work, a new method is proposed for resting state fMRI data analysis. Specifically, the resting state network mapping is formulated as an outlier detection process that is implemented using one-class support vector machine (SVM). The results are refined by using a spatial-feature domain prototype selection method and two-class SVM reclassification. The final decision on each voxel is made by comparing its probabilities of functionally connected and unconnected. The proposed method was evaluated using synthetic and experimental fMRI data. A comparison study was also performed with independent component analysis (ICA) and correlation analysis. The experimental results show that the proposed method can provide comparable or better network detection performance than ICA and correlation analysis. The method is potentially applicable to various resting state quantitative fMRI studies.