{"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/numerical-coordinate-regression-with","title":"Numerical Coordinate Regression with Convolutional Neural Networks","arxiv_id":"1801.07372","date":"2018-01-23","proceeding":null,"authors":["Aiden Nibali","Zhen He","Stuart Morgan","Luke Prendergast"],"abstract":"We study deep learning approaches to inferring numerical coordinates for\npoints of interest in an input image. Existing convolutional neural\nnetwork-based solutions to this problem either take a heatmap matching approach\nor regress to coordinates with a fully connected output layer. Neither of these\napproaches is ideal, since the former is not entirely differentiable, and the\nlatter lacks inherent spatial generalization. We propose our differentiable\nspatial to numerical transform (DSNT) to fill this gap. The DSNT layer adds no\ntrainable parameters, is fully differentiable, and exhibits good spatial\ngeneralization. Unlike heatmap matching, DSNT works well with low heatmap\nresolutions, so it can be dropped in as an output layer for a wide range of\nexisting fully convolutional architectures. Consequently, DSNT offers a better\ntrade-off between inference speed and prediction accuracy compared to existing\ntechniques. When used to replace the popular heatmap matching approach used in\nalmost all state-of-the-art methods for pose estimation, DSNT gives better\nprediction accuracy for all model architectures tested.","url_abs":"http://arxiv.org/abs/1801.07372v2","url_pdf":"http://arxiv.org/pdf/1801.07372v2.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":"numerical-coordinate-regression-with","repo_url":"https://github.com/anibali/dsntnn","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"numerical-coordinate-regression-with","repo_url":"https://github.com/mansimane/WormML","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"pose-estimation","task_name":"Pose Estimation"},{"task_slug":"regression-1","task_name":"regression"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"bottleneck-residual-block","method_name":"Bottleneck Residual Block"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"global-average-pooling","method_name":"Global Average Pooling"},{"method_slug":"heatmap","method_name":"Heatmap"},{"method_slug":"kaiming-initialization","method_name":"Kaiming Initialization"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"residual-block","method_name":"Residual Block"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"speed","method_name":"SPEED"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/pose-estimation-on-mpii-human-pose","task":"Pose Estimation","dataset":"MPII Human Pose","model":"DSNTr ResNet-50@28px","rank_in_archive_order":29,"of":46,"metrics":{"PCKh-0.5":"89.5"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1801.07372","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}