AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Effectiveness of Vision Transformer for Fast and Accurate Single-Stage Pedestrian Detection

Jing Yuan, Barmpoutis Panagiotis, Tania Stathaki

NeurIPS 2022 · Architectures and efficiency › Attention and transformers · approach: Efficient and robust vision transformers · 0 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 enhancement module (ablation)CSP · Caltech testReasonable MR-2: 6.8 -> 3.7Improved
add enhancement module (ablation)CSP · Caltech testHeavy MR-2: 50.7 -> 42.8Improved
add enhancement module (ablation)CSP · Citypersons valHeavy MR-2: 41.8 -> 38.6Improved
add enhancement module (ablation)CSP · Citypersons valReasonable MR-2: 11.7 -> 10.9Improved
add enhancement module on three levels (ablation)SSD300 · Caltech test, IoU 0.5Reasonable MR-2: 28.6 -> 22.6Improved
add enhancement module on three levels (ablation)Progressive refinement detector · Caltech testReasonable MR-2: 13.4 -> 12.3Improved
add enhancement module (inference cost) (ablation)CSP · Caltech testFPS: 39.2 -> 29.5Hurt
add enhancement module (parameters) (ablation)CSP · Model sizetotal parameters: 39.96 M -> 32.24 MImproved
use 2 encoders instead of 1 (ablation)CSP + deformable-encoder feature enhancement · Caltech testReasonable MR-2: 3.7 -> 4.9Hurt
use 4 encoders instead of 1 (ablation)CSP + deformable-encoder feature enhancement · Caltech testHeavy MR-2: 42.8 -> 45.1Hurt
input scales 1/4,1/8,1/16 instead of 1/4,1/8,1/8 (ablation)CSP + deformable-encoder feature enhancement · Caltech testReasonable MR-2: 5.9 -> 4.1Improved
concatenate enhanced maps with L2Norm instead of add or separate (ablation)CSP + deformable-encoder feature enhancement · Caltech testReasonable and Heavy MR-2: 3.7 and 42.8 bestImproved

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