Matches in SemOpenAlex for { <https://semopenalex.org/work/W4281488904> ?p ?o ?g. }
Showing items 1 to 81 of
81
with 100 items per page.
- W4281488904 abstract "In this paper, we formulate a potentially valuable panoramic depth completion (PDC) task as panoramic 3D cameras often produce 360{deg} depth with missing data in complex scenes. Its goal is to recover dense panoramic depths from raw sparse ones and panoramic RGB images. To deal with the PDC task, we train a deep network that takes both depth and image as inputs for the dense panoramic depth recovery. However, it needs to face a challenging optimization problem of the network parameters due to its non-convex objective function. To address this problem, we propose a simple yet effective approach termed M{^3}PT: multi-modal masked pre-training. Specifically, during pre-training, we simultaneously cover up patches of the panoramic RGB image and sparse depth by shared random mask, then reconstruct the sparse depth in the masked regions. To our best knowledge, it is the first time that we show the effectiveness of masked pre-training in a multi-modal vision task, instead of the single-modal task resolved by masked autoencoders (MAE). Different from MAE where fine-tuning completely discards the decoder part of pre-training, there is no architectural difference between the pre-training and fine-tuning stages in our M$^{3}$PT as they only differ in the prediction density, which potentially makes the transfer learning more convenient and effective. Extensive experiments verify the effectiveness of M{^3}PT on three panoramic datasets. Notably, we improve the state-of-the-art baselines by averagely 26.2% in RMSE, 51.7% in MRE, 49.7% in MAE, and 37.5% in RMSElog on three benchmark datasets." @default.
- W4281488904 created "2022-05-26" @default.
- W4281488904 creator A5009232538 @default.
- W4281488904 creator A5017320345 @default.
- W4281488904 creator A5027835055 @default.
- W4281488904 creator A5030691366 @default.
- W4281488904 creator A5045309022 @default.
- W4281488904 creator A5050141622 @default.
- W4281488904 date "2022-03-18" @default.
- W4281488904 modified "2023-10-16" @default.
- W4281488904 title "Multi-Modal Masked Pre-Training for Monocular Panoramic Depth Completion" @default.
- W4281488904 doi "https://doi.org/10.48550/arxiv.2203.09855" @default.
- W4281488904 hasPublicationYear "2022" @default.
- W4281488904 type Work @default.
- W4281488904 citedByCount "0" @default.
- W4281488904 crossrefType "posted-content" @default.
- W4281488904 hasAuthorship W4281488904A5009232538 @default.
- W4281488904 hasAuthorship W4281488904A5017320345 @default.
- W4281488904 hasAuthorship W4281488904A5027835055 @default.
- W4281488904 hasAuthorship W4281488904A5030691366 @default.
- W4281488904 hasAuthorship W4281488904A5045309022 @default.
- W4281488904 hasAuthorship W4281488904A5050141622 @default.
- W4281488904 hasBestOaLocation W42814889041 @default.
- W4281488904 hasConcept C108583219 @default.
- W4281488904 hasConcept C115961682 @default.
- W4281488904 hasConcept C141268832 @default.
- W4281488904 hasConcept C144024400 @default.
- W4281488904 hasConcept C153180895 @default.
- W4281488904 hasConcept C153294291 @default.
- W4281488904 hasConcept C154945302 @default.
- W4281488904 hasConcept C162324750 @default.
- W4281488904 hasConcept C185592680 @default.
- W4281488904 hasConcept C187736073 @default.
- W4281488904 hasConcept C188027245 @default.
- W4281488904 hasConcept C205649164 @default.
- W4281488904 hasConcept C2777211547 @default.
- W4281488904 hasConcept C2779304628 @default.
- W4281488904 hasConcept C2780451532 @default.
- W4281488904 hasConcept C31972630 @default.
- W4281488904 hasConcept C36289849 @default.
- W4281488904 hasConcept C41008148 @default.
- W4281488904 hasConcept C65909025 @default.
- W4281488904 hasConcept C71139939 @default.
- W4281488904 hasConcept C82990744 @default.
- W4281488904 hasConceptScore W4281488904C108583219 @default.
- W4281488904 hasConceptScore W4281488904C115961682 @default.
- W4281488904 hasConceptScore W4281488904C141268832 @default.
- W4281488904 hasConceptScore W4281488904C144024400 @default.
- W4281488904 hasConceptScore W4281488904C153180895 @default.
- W4281488904 hasConceptScore W4281488904C153294291 @default.
- W4281488904 hasConceptScore W4281488904C154945302 @default.
- W4281488904 hasConceptScore W4281488904C162324750 @default.
- W4281488904 hasConceptScore W4281488904C185592680 @default.
- W4281488904 hasConceptScore W4281488904C187736073 @default.
- W4281488904 hasConceptScore W4281488904C188027245 @default.
- W4281488904 hasConceptScore W4281488904C205649164 @default.
- W4281488904 hasConceptScore W4281488904C2777211547 @default.
- W4281488904 hasConceptScore W4281488904C2779304628 @default.
- W4281488904 hasConceptScore W4281488904C2780451532 @default.
- W4281488904 hasConceptScore W4281488904C31972630 @default.
- W4281488904 hasConceptScore W4281488904C36289849 @default.
- W4281488904 hasConceptScore W4281488904C41008148 @default.
- W4281488904 hasConceptScore W4281488904C65909025 @default.
- W4281488904 hasConceptScore W4281488904C71139939 @default.
- W4281488904 hasConceptScore W4281488904C82990744 @default.
- W4281488904 hasLocation W42814889041 @default.
- W4281488904 hasOpenAccess W4281488904 @default.
- W4281488904 hasPrimaryLocation W42814889041 @default.
- W4281488904 hasRelatedWork W1882467697 @default.
- W4281488904 hasRelatedWork W2138569648 @default.
- W4281488904 hasRelatedWork W2283162247 @default.
- W4281488904 hasRelatedWork W2378609488 @default.
- W4281488904 hasRelatedWork W2547650990 @default.
- W4281488904 hasRelatedWork W2794771666 @default.
- W4281488904 hasRelatedWork W2921666650 @default.
- W4281488904 hasRelatedWork W2980953096 @default.
- W4281488904 hasRelatedWork W3010888991 @default.
- W4281488904 hasRelatedWork W3201632203 @default.
- W4281488904 isParatext "false" @default.
- W4281488904 isRetracted "false" @default.
- W4281488904 workType "article" @default.