Matches in SemOpenAlex for { <https://semopenalex.org/work/W2907322140> ?p ?o ?g. }
- W2907322140 abstract "Deep approaches to predict monocular depth and ego-motion have grown in recent years due to their ability to produce dense depth from monocular images. The main idea behind them is to optimize the photometric consistency over image sequences by warping one view into another, similar to direct visual odometry methods. One major drawback is that these methods infer depth from a single view, which might not effectively capture the relation between pixels. Moreover, simply minimizing the photometric loss does not ensure proper pixel correspondences, which is a key factor for accurate depth and pose estimations. In contrast, we propose a 2-view depth network to infer the scene depth from consecutive frames, thereby learning inter-pixel relationships. To ensure better correspondences, thereby better geometric understanding, we propose incorporating epipolar constraints to make the learning more geometrically sound. We use the Essential matrix obtained using Nist'er's Five Point Algorithm, to enforce meaningful geometric constraints, rather than using it as training labels. This allows us to use lesser no. of trainable parameters compared to state-of-the-art methods. The proposed method results in better depth images and pose estimates, which capture the scene structure and motion in a better way. Such a geometrically constrained learning performs successfully even in cases where simply minimizing the photometric error would fail." @default.
- W2907322140 created "2019-01-11" @default.
- W2907322140 creator A5021055773 @default.
- W2907322140 creator A5024771523 @default.
- W2907322140 creator A5080835883 @default.
- W2907322140 date "2018-12-23" @default.
- W2907322140 modified "2023-09-27" @default.
- W2907322140 title "Epipolar Geometry based Learning of Multi-view Depth and Ego-Motion from Monocular Sequences" @default.
- W2907322140 cites W1803059841 @default.
- W2907322140 cites W1836465849 @default.
- W2907322140 cites W1970504153 @default.
- W2907322140 cites W2008706659 @default.
- W2907322140 cites W2033819227 @default.
- W2907322140 cites W2085261163 @default.
- W2907322140 cites W2105303354 @default.
- W2907322140 cites W2108134361 @default.
- W2907322140 cites W2109635530 @default.
- W2907322140 cites W2110917409 @default.
- W2907322140 cites W2115579991 @default.
- W2907322140 cites W2117083552 @default.
- W2907322140 cites W2133665775 @default.
- W2907322140 cites W2150066425 @default.
- W2907322140 cites W2151290401 @default.
- W2907322140 cites W2151996626 @default.
- W2907322140 cites W2171740948 @default.
- W2907322140 cites W2175030374 @default.
- W2907322140 cites W2215578878 @default.
- W2907322140 cites W2300779272 @default.
- W2907322140 cites W2474281075 @default.
- W2907322140 cites W2507953016 @default.
- W2907322140 cites W2520707372 @default.
- W2907322140 cites W2561074213 @default.
- W2907322140 cites W2562127556 @default.
- W2907322140 cites W2593414960 @default.
- W2907322140 cites W2593813181 @default.
- W2907322140 cites W2608018946 @default.
- W2907322140 cites W2788857104 @default.
- W2907322140 cites W2789332689 @default.
- W2907322140 cites W2806446538 @default.
- W2907322140 cites W2949634581 @default.
- W2907322140 cites W2951261569 @default.
- W2907322140 cites W2953384591 @default.
- W2907322140 cites W2963149945 @default.
- W2907322140 cites W2963279353 @default.
- W2907322140 cites W2963583471 @default.
- W2907322140 cites W2963654727 @default.
- W2907322140 cites W2963758027 @default.
- W2907322140 cites W2964121744 @default.
- W2907322140 cites W2964270628 @default.
- W2907322140 cites W603908379 @default.
- W2907322140 cites W612478963 @default.
- W2907322140 cites W840954190 @default.
- W2907322140 hasPublicationYear "2018" @default.
- W2907322140 type Work @default.
- W2907322140 sameAs 2907322140 @default.
- W2907322140 citedByCount "1" @default.
- W2907322140 countsByYear W29073221402020 @default.
- W2907322140 crossrefType "posted-content" @default.
- W2907322140 hasAuthorship W2907322140A5021055773 @default.
- W2907322140 hasAuthorship W2907322140A5024771523 @default.
- W2907322140 hasAuthorship W2907322140A5080835883 @default.
- W2907322140 hasConcept C104114177 @default.
- W2907322140 hasConcept C115961682 @default.
- W2907322140 hasConcept C134306372 @default.
- W2907322140 hasConcept C146159030 @default.
- W2907322140 hasConcept C154945302 @default.
- W2907322140 hasConcept C157202957 @default.
- W2907322140 hasConcept C160633673 @default.
- W2907322140 hasConcept C23379248 @default.
- W2907322140 hasConcept C2524010 @default.
- W2907322140 hasConcept C2780427248 @default.
- W2907322140 hasConcept C28719098 @default.
- W2907322140 hasConcept C31972630 @default.
- W2907322140 hasConcept C33923547 @default.
- W2907322140 hasConcept C41008148 @default.
- W2907322140 hasConcept C65909025 @default.
- W2907322140 hasConceptScore W2907322140C104114177 @default.
- W2907322140 hasConceptScore W2907322140C115961682 @default.
- W2907322140 hasConceptScore W2907322140C134306372 @default.
- W2907322140 hasConceptScore W2907322140C146159030 @default.
- W2907322140 hasConceptScore W2907322140C154945302 @default.
- W2907322140 hasConceptScore W2907322140C157202957 @default.
- W2907322140 hasConceptScore W2907322140C160633673 @default.
- W2907322140 hasConceptScore W2907322140C23379248 @default.
- W2907322140 hasConceptScore W2907322140C2524010 @default.
- W2907322140 hasConceptScore W2907322140C2780427248 @default.
- W2907322140 hasConceptScore W2907322140C28719098 @default.
- W2907322140 hasConceptScore W2907322140C31972630 @default.
- W2907322140 hasConceptScore W2907322140C33923547 @default.
- W2907322140 hasConceptScore W2907322140C41008148 @default.
- W2907322140 hasConceptScore W2907322140C65909025 @default.
- W2907322140 hasLocation W29073221401 @default.
- W2907322140 hasOpenAccess W2907322140 @default.
- W2907322140 hasPrimaryLocation W29073221401 @default.
- W2907322140 hasRelatedWork W2890732989 @default.
- W2907322140 hasRelatedWork W2892197942 @default.
- W2907322140 hasRelatedWork W2905561397 @default.
- W2907322140 hasRelatedWork W2962314446 @default.
- W2907322140 hasRelatedWork W2970065341 @default.
- W2907322140 hasRelatedWork W2970369664 @default.