AI Paper Atlas

Research map / Detection and segmentation / 3D detection and driving

MatrixVT: Efficient Multi-Camera to BEV Transformation for 3D Perception

Hongyu Zhou, Zheng Ge, Zeming Li, Xiangyu Zhang

ICCV 2023 · Detection and segmentation › 3D detection and driving · approach: Multi-camera BEV detection · 60 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add Prime Extraction (compress image feature and depth) to BEVDepthBEVDepth Res-50 · nuScenes valview transformation memory: 734M -> 316MImproved
add Prime Extraction to BEVDepthBEVDepth Res-50 · nuScenes valNDS: 0.414 -> 0.411No clear effect
add Prime Extraction to BEVDepth V2-99BEVDepth V2-99 · nuScenes valmAP: 0.464 -> 0.459No clear effect
use max pooling instead of CollapseConv for feature reduction (ablation)MatrixVT Prime Feature Extractor · nuScenes val Res-50NDS: 0.404 -> 0.415Improved
add refine sub-net after reduction (ablation)MatrixVT Prime Feature Extractor · nuScenes val Res-50mAP: 0.328 -> 0.332Improved
add front-view position embedding (ablation)MatrixVT Prime Feature Extractor · nuScenes val Res-50mAP: 0.332 -> 0.336Improved
use MatrixVT instead of Lift-Splat view transformers on CPUview transformation · latency S1-S6speed: 4 to 8 times faster than LS-BEVDetImproved
use MatrixVT instead of LS-BEVDepth for memoryview transformation · intermediate variables S1-S6memory footprint: 2 to 15 times less than LS-BEVDepthImproved
use Ring and Ray decomposition with reformulated pipelineMatrixVT · common setting C=80, Nd=112, WI=44calculation and memory: 46 times less calculation, 96% less memoryImproved
replace BEVDepth with MatrixVT on lane segmentationBEV map segmentation · nuScenes valIoU-Lane: 0.464 -> 0.448Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy