Prior studies have shown that adults with SCI have body awareness deficits. Mechanisms of neuropathic pain are unclear, which makes finding effective treatments challenging. There are many methods based on level set, which are classified into region-based and edge-based. Background: Neuropathic pain after spinal cord injury (SCI) is notoriously hard to treat. The level set methods are specially used in image with intensity inhomogeneity, such as medical image, SAR image, etc. Extensive experiments on the Automated Cardiac DiagnosisChallenge (ACDC) 2017 dataset and the Brain Tumor Segmentation (BraTS) 2019challenge dataset show that our method outperforms the state-of-the-artsemi-supervised methods, which demonstrates its effectiveness for medical imagesegmentation. Level set is one of active contour models, which is good at handling complex topologies and capturing boundary. Furthermore, at thediscriminator's input, we supplement semantic information constraints onimages, making it simpler for unlabeled data to fit the expected datadistribution. The prediction of unlabeled data learns the pixelstructure and context information in each patch to get enough gradientfeedback, which aids the discriminator in convergent to an optimal state andimproves semi-supervised segmentation performance. Ratherthan single scalar classification results or pixel-level confidence maps, ourproposed discriminator creates patch confidence maps and classifies them at thescale of the patches. Inthis paper, we propose a new semi-supervised adversarial method called PatchConfidence Adversarial Training (PCA) for medical image segmentation. Level set methods over the past decade are summed up and categorized. This article firstly derives the function of curve evolution and original model of level set based on region and edge, respectively. There are many methods based on level set, which are classified into region-based and edge-based. We argue that the current performance restrictions for suchapproaches are the problems of feature extraction and learning preference. The level set methods are specially used in image with intensity inhomogeneity, such as medical image, SAR image, etc. Unlike mostexisting semi-supervised learning methods, adversarial training based methodsdistinguish samples from different sources by learning the data distribution ofthe segmentation map, leading the segmenter to generate more accuratepredictions. Deep learning based semi-supervised learning (SSL) methods have achievedstrong performance in medical image segmentation, which can alleviate doctors'expensive annotation by utilizing a large amount of unlabeled data.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |