DC Field | Value | Language |
---|---|---|
dc.contributor.author | Lee, Seokju | ko |
dc.contributor.author | Im, Sunghoon | ko |
dc.contributor.author | Lin, Stephen | ko |
dc.contributor.author | Kweon, In So | ko |
dc.date.accessioned | 2021-10-27T08:50:21Z | - |
dc.date.available | 2021-10-27T08:50:21Z | - |
dc.date.created | 2021-10-27 | - |
dc.date.created | 2021-10-27 | - |
dc.date.issued | 2021-02 | - |
dc.identifier.citation | 35th AAAI Conference on Artificial Intelligence / 33rd Conference on Innovative Applications of Artificial Intelligence / 11th Symposium on Educational Advances in Artificial Intelligence, pp.1863 - 1872 | - |
dc.identifier.issn | 2159-5399 | - |
dc.identifier.uri | http://hdl.handle.net/10203/288362 | - |
dc.description.abstract | We present an end-to-end joint training framework that explicitly models 6-DoF motion of multiple dynamic objects, ego-motion, and depth in a monocular camera setup without supervision. Our technical contributions are three-fold. First, we highlight the fundamental difference between inverse and forward projection while modeling the individual motion of each rigid object, and propose a geometrically correct projection pipeline using a neural forward projection module. Second, we design a unified instance-aware photometric and geometric consistency loss that holistically imposes self-supervisory signals for every background and object region. Lastly, we introduce a general-purpose auto-annotation scheme using any off-the-shelf instance segmentation and optical flow models to produce video instance segmentation maps that will be utilized as input to our training pipeline. These proposed elements are validated in a detailed ablation study. Through extensive experiments conducted on the KITTI and Cityscapes dataset, our framework is shown to outperform the state-of-the-art depth and motion estimation methods. Our code, dataset, and models are publicly available. | - |
dc.language | English | - |
dc.publisher | ASSOC ADVANCEMENT ARTIFICIAL INTELLIGENCE | - |
dc.title | Learning Monocular Depth in Dynamic Scenes via Instance-Aware Projection Consistency | - |
dc.type | Conference | - |
dc.identifier.wosid | 000680423501106 | - |
dc.identifier.scopusid | 2-s2.0-85130038874 | - |
dc.type.rims | CONF | - |
dc.citation.beginningpage | 1863 | - |
dc.citation.endingpage | 1872 | - |
dc.citation.publicationname | 35th AAAI Conference on Artificial Intelligence / 33rd Conference on Innovative Applications of Artificial Intelligence / 11th Symposium on Educational Advances in Artificial Intelligence | - |
dc.identifier.conferencecountry | US | - |
dc.identifier.conferencelocation | Virtual | - |
dc.contributor.localauthor | Kweon, In So | - |
dc.contributor.nonIdAuthor | Lee, Seokju | - |
dc.contributor.nonIdAuthor | Im, Sunghoon | - |
dc.contributor.nonIdAuthor | Lin, Stephen | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.