FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4312410080> ?p ?o ?g. }

• W4312410080 endingPage "561" @default.
• W4312410080 startingPage "544" @default.
• W4312410080 abstract "Since Intersection-over-Union (IoU) based optimization maintains the consistency of the final IoU prediction metric and losses, it has been widely used in both regression and classification branches of single-stage 2D object detectors. Recently, several 3D object detection methods adopt IoU-based optimization and directly replace the 2D IoU with 3D IoU. However, such a direct computation in 3D is very costly due to the complex implementation and inefficient backward operations. Moreover, 3D IoU-based optimization is sub-optimal as it is sensitive to rotation and thus can cause training instability and detection performance deterioration. In this paper, we propose a novel Rotation-Decoupled IoU (RDIoU) method that can mitigate the rotation-sensitivity issue, and produce more efficient optimization objectives compared with 3D IoU during the training stage. Specifically, our RDIoU simplifies the complex interactions of regression parameters by decoupling the rotation variable as an independent term, yet preserving the geometry of 3D IoU. By incorporating RDIoU into both the regression and classification branches, the network is encouraged to learn more precise bounding boxes and concurrently overcome the misalignment issue between classification and regression. Extensive experiments on the benchmark KITTI and Waymo Open Dataset validate that our RDIoU method can bring substantial improvement for the single-stage 3D object detection. The code is available at https://github.com/hlsheng1/RDIoU ." @default.
• W4312410080 created "2023-01-04" @default.
• W4312410080 creator A5024965898 @default.
• W4312410080 creator A5044276866 @default.
• W4312410080 creator A5059927420 @default.
• W4312410080 creator A5069455195 @default.
• W4312410080 creator A5071387519 @default.
• W4312410080 creator A5071967339 @default.
• W4312410080 creator A5074924505 @default.
• W4312410080 creator A5079865276 @default.
• W4312410080 date "2022-01-01" @default.
• W4312410080 modified "2023-10-01" @default.
• W4312410080 title "Rethinking IoU-based Optimization for Single-stage 3D Object Detection" @default.
• W4312410080 cites W1536680647 @default.
• W4312410080 cites W2115579991 @default.
• W4312410080 cites W2131396337 @default.
• W4312410080 cites W2555618208 @default.
• W4312410080 cites W2897529137 @default.
• W4312410080 cites W2949708697 @default.
• W4312410080 cites W2962766617 @default.
• W4312410080 cites W2963182550 @default.
• W4312410080 cites W2963351448 @default.
• W4312410080 cites W2963727135 @default.
• W4312410080 cites W2964062501 @default.
• W4312410080 cites W2968296999 @default.
• W4312410080 cites W2981949127 @default.
• W4312410080 cites W2982363097 @default.
• W4312410080 cites W2991359031 @default.
• W4312410080 cites W2997747012 @default.
• W4312410080 cites W3034314779 @default.
• W4312410080 cites W3034602892 @default.
• W4312410080 cites W3034681945 @default.
• W4312410080 cites W3035172746 @default.
• W4312410080 cites W3035346742 @default.
• W4312410080 cites W3097779785 @default.
• W4312410080 cites W3100234280 @default.
• W4312410080 cites W3107212734 @default.
• W4312410080 cites W3118341329 @default.
• W4312410080 cites W3138516171 @default.
• W4312410080 cites W3166089996 @default.
• W4312410080 cites W3167095230 @default.
• W4312410080 cites W3167732492 @default.
• W4312410080 cites W3176187859 @default.
• W4312410080 cites W3205005447 @default.
• W4312410080 cites W4200632008 @default.
• W4312410080 cites W4214777292 @default.
• W4312410080 doi "https://doi.org/10.1007/978-3-031-20077-9_32" @default.
• W4312410080 hasPublicationYear "2022" @default.
• W4312410080 type Work @default.
• W4312410080 citedByCount "5" @default.
• W4312410080 countsByYear W43124100802023 @default.
• W4312410080 crossrefType "book-chapter" @default.
• W4312410080 hasAuthorship W4312410080A5024965898 @default.
• W4312410080 hasAuthorship W4312410080A5044276866 @default.
• W4312410080 hasAuthorship W4312410080A5059927420 @default.
• W4312410080 hasAuthorship W4312410080A5069455195 @default.
• W4312410080 hasAuthorship W4312410080A5071387519 @default.
• W4312410080 hasAuthorship W4312410080A5071967339 @default.
• W4312410080 hasAuthorship W4312410080A5074924505 @default.
• W4312410080 hasAuthorship W4312410080A5079865276 @default.
• W4312410080 hasBestOaLocation W43124100802 @default.
• W4312410080 hasConcept C124101348 @default.
• W4312410080 hasConcept C13280743 @default.
• W4312410080 hasConcept C153180895 @default.
• W4312410080 hasConcept C154945302 @default.
• W4312410080 hasConcept C162324750 @default.
• W4312410080 hasConcept C176217482 @default.
• W4312410080 hasConcept C185798385 @default.
• W4312410080 hasConcept C205649164 @default.
• W4312410080 hasConcept C21547014 @default.
• W4312410080 hasConcept C2776151529 @default.
• W4312410080 hasConcept C41008148 @default.
• W4312410080 hasConcept C63584917 @default.
• W4312410080 hasConceptScore W4312410080C124101348 @default.
• W4312410080 hasConceptScore W4312410080C13280743 @default.
• W4312410080 hasConceptScore W4312410080C153180895 @default.
• W4312410080 hasConceptScore W4312410080C154945302 @default.
• W4312410080 hasConceptScore W4312410080C162324750 @default.
• W4312410080 hasConceptScore W4312410080C176217482 @default.
• W4312410080 hasConceptScore W4312410080C185798385 @default.
• W4312410080 hasConceptScore W4312410080C205649164 @default.
• W4312410080 hasConceptScore W4312410080C21547014 @default.
• W4312410080 hasConceptScore W4312410080C2776151529 @default.
• W4312410080 hasConceptScore W4312410080C41008148 @default.
• W4312410080 hasConceptScore W4312410080C63584917 @default.
• W4312410080 hasLocation W43124100801 @default.
• W4312410080 hasLocation W43124100802 @default.
• W4312410080 hasOpenAccess W4312410080 @default.
• W4312410080 hasPrimaryLocation W43124100801 @default.
• W4312410080 hasRelatedWork W2593809839 @default.
• W4312410080 hasRelatedWork W2801801420 @default.
• W4312410080 hasRelatedWork W2804574147 @default.
• W4312410080 hasRelatedWork W3090145263 @default.
• W4312410080 hasRelatedWork W3130180997 @default.
• W4312410080 hasRelatedWork W3133978618 @default.
• W4312410080 hasRelatedWork W3175230847 @default.
• W4312410080 hasRelatedWork W4225418833 @default.
• W4312410080 hasRelatedWork W4300931582 @default.

• next