{"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/robust-optimization-for-deep-regression","title":"Robust Optimization for Deep Regression","arxiv_id":"1505.06606","date":"2015-05-25","proceeding":"ICCV 2015 12","authors":["Vasileios Belagiannis","Christian Rupprecht","Gustavo Carneiro","Nassir Navab"],"abstract":"Convolutional Neural Networks (ConvNets) have successfully contributed to\nimprove the accuracy of regression-based methods for computer vision tasks such\nas human pose estimation, landmark localization, and object detection. The\nnetwork optimization has been usually performed with L2 loss and without\nconsidering the impact of outliers on the training process, where an outlier in\nthis context is defined by a sample estimation that lies at an abnormal\ndistance from the other training sample estimations in the objective space. In\nthis work, we propose a regression model with ConvNets that achieves robustness\nto such outliers by minimizing Tukey's biweight function, an M-estimator robust\nto outliers, as the loss function for the ConvNet. In addition to the robust\nloss, we introduce a coarse-to-fine model, which processes input images of\nprogressively higher resolutions for improving the accuracy of the regressed\nvalues. In our experiments, we demonstrate faster convergence and better\ngeneralization of our robust loss function for the tasks of human pose\nestimation and age estimation from face images. We also show that the\ncombination of the robust loss function with the coarse-to-fine model produces\ncomparable or better results than current state-of-the-art approaches in four\npublicly available human pose estimation datasets.","url_abs":"http://arxiv.org/abs/1505.06606v2","url_pdf":"http://arxiv.org/pdf/1505.06606v2.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":"robust-optimization-for-deep-regression","repo_url":"https://github.com/bazilas/matconvnet-deepReg","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}}],"tasks":[{"task_slug":"age-estimation","task_name":"Age Estimation"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"pose-estimation","task_name":"Pose Estimation"},{"task_slug":"object-detection-1","task_name":"object-detection"},{"task_slug":"regression-1","task_name":"regression"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1505.06606","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}