Color-texture based unsupervised segmentation using JSEG with fuzzy connectedness

Color quantization is bound to lose spatial information of color distribution. If too much necessary spatial distribution information of color is lost in JSEG, it is difficult or even impossible for JSEG to segment image correctly. Enlightened from segmentation based on fuzzy theories, soft class-map is constracted to solve that problem. The definitions of values and other related ones are adjusted according to the soft class-map. With more detailed values obtained from soft class map, more color distribution information is preserved. Experiments on a synthetic image and many other color images illustrate that JSEG with soft class-map can solve efficiently the problem that in a region there may exist color gradual variation in a smooth transition. It is a more robust method especially for images which haven＇ t been heavily blurred near boundaries of underlying regions.

Unsupervised Color Segmentation with Reconstructed Spatial Weighted Gaussian Mixture Model and Random Color Histogram

Image classification and unsupervised image segmentation can be achieved using the Gaussian mixture model.Although the Gaussian mixture model enhances the flexibility of image segmentation,it does not reflect spatial information and is sensitive to the segmentation parameter.In this study,we first present an efficient algorithm that incorporates spatial information into the Gaussian mixture model(GMM)without parameter estimation.The proposed model highlights the residual region with considerable information and constructs color saliency.Second,we incorporate the content-based color saliency as spatial information in the Gaussian mixture model.The segmentation is performed by clustering each pixel into an appropriate component according to the expectation maximization and maximum criteria.Finally,the random color histogram assigns a unique color to each cluster and creates an attractive color by default for segmentation.A random color histogram serves as an effective tool for data visualization and is instrumental in the creation of generative art,facilitating both analytical and aesthetic objectives.For experiments,we have used the Berkeley segmentation dataset BSDS-500 and Microsoft Research in Cambridge dataset.In the study,the proposed model showcases notable advancements in unsupervised image segmentation,with probabilistic rand index(PRI)values reaching 0.80,BDE scores as low as 12.25 and 12.02,compactness variations at 0.59 and 0.7,and variation of information(VI)reduced to 2.0 and 1.49 for the BSDS-500 and MSRC datasets,respectively,outperforming current leading-edge methods and yielding more precise segmentations.

Trajectory Time Series Compression Algorithm Based on Unsupervised Segmentation

Aiming at the problem of ignoring the importance of starting point features of trajecory segmentation in existing trajectory compression algorithms,a study was conducted on the preprocessing process of trajectory time series.Firstly,an algorithm improvement was proposed based on the segmentation algorithm GRASP-UTS(Greedy Randomized Adaptive Search Procedure for Unsupervised Trajectory Segmentation).On the basis of considering trajectory coverage,this algorithm designs an adaptive parameter adjustment to segment long-term trajectory data reasonably and the identification of an optimal starting point for segmentation.Then the compression efficiency of typical offline and online algorithms,such as the Douglas-Peucker algorithm,the Sliding Window algorithm and its enhancements,was compared before and after segmentation.The experimental findings highlight that the Adaptive Parameters GRASP-UTS segmentation approach leads to higher fitting precision in trajectory time series compression and improved algorithm efficiency post-segmentation.Additionally,the compression performance of the Improved Sliding Window algorithm post-segmentation showcases its suitability for trajectories of varying scales,providing reasonable compression accuracy.

Unsupervised segmentation of medical image based on difference of mutual information

In the scope of medical image processing, segmentation is important and difficult. There are still two problems which trouble us in this field. One is how to determine the number of clusters in an image and the other is how to segment medical images containing lesions. A new segmentation method called DDC, based on difference of mutual information （dMI） and pixon, is proposed in this paper. Experiments demonstrate that dMI shows one kind of intrinsic relationship between the segmented image and the original one and so it can be used to well determine the number of clusters. Furthermore, multi-modality medical images with lesions can be automatically and successfully segmented by DDC method.

Faster Region Based Convolutional Neural Network for Skin Lesion Segmentation

The diagnostic interpretation of dermoscopic images is a complex task as it is very difficult to identify the skin lesions from the normal.Thus the accurate detection of potential abnormalities is required for patient monitoring and effec-tive treatment.In this work,a Two-Tier Segmentation(TTS)system is designed,which combines the unsupervised and supervised techniques for skin lesion seg-mentation.It comprises preprocessing by the medianfilter,TTS by Colour K-Means Clustering(CKMC)for initial segmentation and Faster Region based Con-volutional Neural Network(FR-CNN)for refined segmentation.The CKMC approach is evaluated using the different number of clusters(k=3,5,7,and 9).An inception network with batch normalization is employed to segment mel-anoma regions effectively.Different loss functions such as Mean Absolute Error(MAE),Cross Entropy Loss(CEL),and Dice Loss(DL)are utilized for perfor-mance evaluation of the TTS system.The anchor box technique is employed to detect the melanoma region effectively.The TTS system is evaluated using 200 dermoscopic images from the PH2 database.The segmentation accuracies are analyzed in terms of Pixel Accuracy(PA)and Jaccard Index(JI).Results show that the TTS system has 90.19%PA with 0.8048 JI for skin lesion segmentation using DL in FR-CNN with seven clusters in CKMC than CEL and MAE.

Evaluating quality of motion for unsupervised video object segmentation

Current mainstream unsupervised video object segmentation(UVOS) approaches typically incorporate optical flow as motion information to locate the primary objects in coherent video frames. However, they fuse appearance and motion information without evaluating the quality of the optical flow. When poor-quality optical flow is used for the interaction with the appearance information, it introduces significant noise and leads to a decline in overall performance. To alleviate this issue, we first employ a quality evaluation module(QEM) to evaluate the optical flow. Then, we select high-quality optical flow as motion cues to fuse with the appearance information, which can prevent poor-quality optical flow from diverting the network's attention. Moreover, we design an appearance-guided fusion module(AGFM) to better integrate appearance and motion information. Extensive experiments on several widely utilized datasets, including DAVIS-16, FBMS-59, and You Tube-Objects, demonstrate that the proposed method outperforms existing methods.

A Chan–Vese Model Based on the Markov Chain for Unsupervised Medical Image Segmentation

The accurate segmentation of medical images is crucial to medical care and research;however, many efficient supervised image segmentation methods require sufficient pixel level labels. Such requirement is difficult to meet in practice and even impossible in some cases, e.g., rare Pathoma images. Inspired by traditional unsupervised methods, we propose a novel Chan–Vese model based on the Markov chain for unsupervised medical image segmentation. It combines local information brought by superpixels with the global difference between the target tissue and the background. Based on the Chan–Vese model, we utilize weight maps generated by the Markov chain to model and solve the segmentation problem iteratively using the min-cut algorithm at the superpixel level.Our method exploits abundant boundary and local region information in segmentation and thus can handle images with intensity inhomogeneity and object sparsity. In our method, users gain the power of fine-tuning parameters to achieve satisfactory results for each segmentation. By contrast, the result from deep learning based methods is rigid.The performance of our method is assessed by using four Computerized Tomography(CT) datasets. Experimental results show that the proposed method outperforms traditional unsupervised segmentation techniques.