{"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/deep-depth-from-defocus-how-can-defocus-blur","title":"Deep Depth from Defocus: how can defocus blur improve 3D estimation using dense neural networks?","arxiv_id":"1809.01567","date":"2018-09-05","proceeding":null,"authors":["Marcela Carvalho","Bertrand Le Saux","Pauline Trouvé-Peloux","Andrés Almansa","Frédéric Champagnat"],"abstract":"Depth estimation is of critical interest for scene understanding and accurate\n3D reconstruction. Most recent approaches in depth estimation with deep\nlearning exploit geometrical structures of standard sharp images to predict\ncorresponding depth maps. However, cameras can also produce images with defocus\nblur depending on the depth of the objects and camera settings. Hence, these\nfeatures may represent an important hint for learning to predict depth. In this\npaper, we propose a full system for single-image depth prediction in the wild\nusing depth-from-defocus and neural networks. We carry out thorough experiments\nto test deep convolutional networks on real and simulated defocused images\nusing a realistic model of blur variation with respect to depth. We also\ninvestigate the influence of blur on depth prediction observing model\nuncertainty with a Bayesian neural network approach. From these studies, we\nshow that out-of-focus blur greatly improves the depth-prediction network\nperformances. Furthermore, we transfer the ability learned on a synthetic,\nindoor dataset to real, indoor and outdoor images. For this purpose, we present\na new dataset containing real all-focus and defocused images from a Digital\nSingle-Lens Reflex (DSLR) camera, paired with ground truth depth maps obtained\nwith an active 3D sensor for indoor scenes. The proposed approach is\nsuccessfully validated on both this new dataset and standard ones as NYUv2 or\nDepth-in-the-Wild. Code and new datasets are available at\nhttps://github.com/marcelampc/d3net_depth_estimation","url_abs":"http://arxiv.org/abs/1809.01567v2","url_pdf":"http://arxiv.org/pdf/1809.01567v2.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":"deep-depth-from-defocus-how-can-defocus-blur","repo_url":"https://github.com/marcelampc/d3net_depth_estimation","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"3d-reconstruction","task_name":"3D Reconstruction"},{"task_slug":"depth-estimation","task_name":"Depth Estimation"},{"task_slug":"depth-prediction","task_name":"Depth Prediction"},{"task_slug":"scene-understanding","task_name":"Scene Understanding"}],"methods":[],"datasets_introduced":[{"slug":"indoor-and-outdoor-dfd-dataset","name":"Indoor and outdoor DFD dataset","full_name":""}],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}