{"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/unsupervised-learning-of-monocular-depth-1","title":"Unsupervised Learning of Monocular Depth Estimation and Visual Odometry with Deep Feature Reconstruction","arxiv_id":"1803.03893","date":"2018-03-11","proceeding":"CVPR 2018 6","authors":["Huangying Zhan","Ravi Garg","Chamara Saroj Weerasekera","Kejie Li","Harsh Agarwal","Ian Reid"],"abstract":"Despite learning based methods showing promising results in single view depth\nestimation and visual odometry, most existing approaches treat the tasks in a\nsupervised manner. Recent approaches to single view depth estimation explore\nthe possibility of learning without full supervision via minimizing photometric\nerror. In this paper, we explore the use of stereo sequences for learning depth\nand visual odometry. The use of stereo sequences enables the use of both\nspatial (between left-right pairs) and temporal (forward backward) photometric\nwarp error, and constrains the scene depth and camera motion to be in a common,\nreal-world scale. At test time our framework is able to estimate single view\ndepth and two-view odometry from a monocular sequence. We also show how we can\nimprove on a standard photometric warp loss by considering a warp of deep\nfeatures. We show through extensive experiments that: (i) jointly training for\nsingle view depth and visual odometry improves depth prediction because of the\nadditional constraint imposed on depths and achieves competitive results for\nvisual odometry; (ii) deep feature-based warping loss improves upon simple\nphotometric warp loss for both single view depth estimation and visual\nodometry. Our method outperforms existing learning based methods on the KITTI\ndriving dataset in both tasks. The source code is available at\nhttps://github.com/Huangying-Zhan/Depth-VO-Feat","url_abs":"http://arxiv.org/abs/1803.03893v3","url_pdf":"http://arxiv.org/pdf/1803.03893v3.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":"unsupervised-learning-of-monocular-depth-1","repo_url":"https://github.com/Huangying-Zhan/Depth-VO-Feat","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"caffe2","reach":{"status":"ok","spdx":"NOASSERTION"}}],"tasks":[{"task_slug":"depth-and-camera-motion","task_name":"Depth And Camera Motion"},{"task_slug":"depth-estimation","task_name":"Depth Estimation"},{"task_slug":"depth-prediction","task_name":"Depth Prediction"},{"task_slug":"monocular-depth-estimation","task_name":"Monocular Depth Estimation"},{"task_slug":"visual-odometry","task_name":"Visual Odometry"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1803.03893","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}