AI Paper Atlas

Research map / Detection and segmentation / CNN features

TALISMAN: Targeted Active Learning for Object Detection with Rare Classes and Slices Using Submodular Mutual Information

Suraj Kothawade, Saikat Ghosh, Sumit Shekhar, Yu Xiang, Rishabh Iyer

ECCV 2022 · Detection and segmentation › CNN features · approach: Region-based and query-based detectors · 29 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
select with SMI functions (FLMI/GCMI) instead of uncertainty samplingActive learning for Faster R-CNN · VOC07+12 rare classesavg AP of rare classes: ≈8%−10%Improved
select with SMI functions instead of uncertainty samplingActive learning for Faster R-CNN · VOC07+12mAP: ≈2%−3%Improved
select with SMI functions instead of baselinesActive learning for Faster R-CNN · BDD100K motorcycle/bicycle at nightrare-class AP on rare slice: ≈5%−14%Improved
select with SMI functions instead of baselinesActive learning for Faster R-CNN · BDD100K motorcycle/bicycle at nightfull-test mAP: ≈2%−4%Improved
select with SMI functions instead of baselinesActive learning for Faster R-CNN · BDD100K pedestrian at night/rainy/highwaypedestrian AP on rare slice: ≈5%−10%Improved
use GCMI rather than other methodsActive learning for Faster R-CNN · BDD100K motorcycle/bicycle at nightrare slice objects selectedImproved

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