{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/normalization-in-training-u-net-for-2d","title":"Normalization in Training U-Net for 2D Biomedical Semantic Segmentation","arxiv_id":"1809.03783","date":"2018-09-11","proceeding":null,"authors":["Xiao-Yun Zhou","Guang-Zhong Yang"],"abstract":"2D biomedical semantic segmentation is important for robotic vision in\nsurgery. Segmentation methods based on Deep Convolutional Neural Network (DCNN)\ncan out-perform conventional methods in terms of both accuracy and levels of\nautomation. One common issue in training a DCNN for biomedical semantic\nsegmentation is the internal covariate shift where the training of\nconvolutional kernels is encumbered by the distribution change of input\nfeatures, hence both the training speed and performance are decreased. Batch\nNormalization (BN) is the first proposed method for addressing internal\ncovariate shift and is widely used. Instance Normalization (IN) and Layer\nNormalization (LN) have also been proposed. Group Normalization (GN) is\nproposed more recently and has not yet been applied to 2D biomedical semantic\nsegmentation, however, no specific validations on GN were given. Most DCNNs for\nbiomedical semantic segmentation adopt BN as the normalization method by\ndefault, without reviewing its performance. In this paper, four normalization\nmethods - BN, IN, LN and GN are compared in details, specifically for 2D\nbiomedical semantic segmentation. U-Net is adopted as the basic DCNN structure.\nThree datasets regarding the Right Ventricle (RV), aorta, and Left Ventricle\n(LV) are used for the validation. The results show that detailed subdivision of\nthe feature map, i.e. GN with a large group number or IN, achieves higher\naccuracy. This accuracy improvement mainly comes from better model\ngeneralization. Codes are uploaded and maintained at Xiao-Yun Zhou's Github.","url_abs":"http://arxiv.org/abs/1809.03783v3","url_pdf":"http://arxiv.org/pdf/1809.03783v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"normalization-in-training-u-net-for-2d","repo_url":"https://github.com/XiaoYunZhou27/UNet-with-BN-GN-IN-LN","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"2d-semantic-segmentation","task_name":"2D Semantic Segmentation"},{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"}],"methods":[{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"concatenated-skip-connection","method_name":"Concatenated Skip Connection"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dcnn","method_name":"DCNN"},{"method_slug":"group-normalization","method_name":"Group Normalization"},{"method_slug":"instance-normalization","method_name":"Instance Normalization"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"speed","method_name":"SPEED"},{"method_slug":"u-net","method_name":"U-Net"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}