{"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/a-large-dataset-to-train-convolutional","title":"A Large Dataset to Train Convolutional Networks for Disparity, Optical Flow, and Scene Flow Estimation","arxiv_id":"1512.02134","date":"2015-12-07","proceeding":"CVPR 2016 6","authors":["Nikolaus Mayer","Eddy Ilg","Philip Häusser","Philipp Fischer","Daniel Cremers","Alexey Dosovitskiy","Thomas Brox"],"abstract":"Recent work has shown that optical flow estimation can be formulated as a\nsupervised learning task and can be successfully solved with convolutional\nnetworks. Training of the so-called FlowNet was enabled by a large\nsynthetically generated dataset. The present paper extends the concept of\noptical flow estimation via convolutional networks to disparity and scene flow\nestimation. To this end, we propose three synthetic stereo video datasets with\nsufficient realism, variation, and size to successfully train large networks.\nOur datasets are the first large-scale datasets to enable training and\nevaluating scene flow methods. Besides the datasets, we present a convolutional\nnetwork for real-time disparity estimation that provides state-of-the-art\nresults. By combining a flow and disparity estimation network and training it\njointly, we demonstrate the first scene flow estimation with a convolutional\nnetwork.","url_abs":"http://arxiv.org/abs/1512.02134v1","url_pdf":"http://arxiv.org/pdf/1512.02134v1.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":"a-large-dataset-to-train-convolutional","repo_url":"https://github.com/HKBU-HPML/FADNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"a-large-dataset-to-train-convolutional","repo_url":"https://github.com/HKBU-HPML/IRS","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"a-large-dataset-to-train-convolutional","repo_url":"https://github.com/arashk7/DispNet_Keras","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"disparity-estimation","task_name":"Disparity Estimation"},{"task_slug":"optical-flow-estimation","task_name":"Optical Flow Estimation"},{"task_slug":"scene-flow-estimation","task_name":"Scene Flow Estimation"}],"methods":[],"datasets_introduced":[{"slug":"flyingthings3d","name":"FlyingThings3D","full_name":""}],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1512.02134","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}