AI Paper Atlas

Research map / Detection and segmentation / Tracking

Detect or Track: Towards Cost-Effective Video Object Detection/Tracking

AAAI 2019 · Detection and segmentation › Tracking · approach: Unified transformer tracking · 57 citations (OpenAlex)

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's title only (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
Use a learned scheduler to choose between detection and tracking instead of fixed frame skippingDorT · ImageNet VID video object detection/trackingmAPHelped
Use RoI convolution instead of crop-and-resize SiamFC for multi-box tracking (ablation)DorT · ImageNet VIDspeed: over 10x; 3 fps -> 54 fpsHelped
Use RoI convolution instead of crop-and-resize SiamFC for multi-box tracking (ablation)DorT · ImageNet VIDmAP: 56.5\%No clear effect
Use DorT rather than deep feature flow with Hungarian associationVideo object detection/tracking · ImageNet VIDmAPHelped
Use DorT rather than deep feature flowVideo object detection · ImageNet VIDspeedHelped
Use DorT rather than deep feature flowVideo object detection · ImageNet VIDmAP: no more than 1% mAP lossNo clear effect
Set the scheduler track-decision threshold to 0.97 for conservative tracking decisionsDorT scheduler · ImageNet VIDfalse positive rate: 1.9%Helped

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