AI Paper Atlas

Research map / Detection and segmentation / CNN features

Active Learning for Deep Object Detection via Probabilistic Modeling

Jiwoong Choi, Ismail Elezi, Hyuk-Jae Lee, Clement Farabet, Jose M. Alvarez

ICCV 2021 · Detection and segmentation › CNN features · approach: Region-based and query-based detectors · 144 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
replace SSD with a mixture density model on localization and classification heads (ablation)SSD · VOC07mAP IoU>0.75: 43.36 -> 46.11Improved
replace SSD with a mixture density model on localization and classification heads (ablation)SSD · MS-COCOmAP IoU>0.75: 11.93 -> 13.57Improved
use max of four uncertainties instead of random sampling (ablation)active learning on SSD · VOC07mAP at 4k images: 68.47 -> 69.43Improved
combine localization and classification uncertainties rather than one task alone (ablation)active learning on SSD · VOC07image selection overlap: 48% -> 14%Improved
use max instead of sum of the four uncertainties (ablation)active learning on SSD · VOC07mAP at 4k images: 69.09 -> 69.43Improved
use Ours_gmm sampling instead of MC-dropoutactive learning · VOC07mAP at 4k images: 69.39 -> 69.43Improved
add Ours_eff scoring instead of LLAL on MS-COCOactive learning · MS-COCOmAP at 7k images: 29.53 -> 30.02Improved
increase mixture components K from 1 to 4 (ablation)Ours_gmm · VOC07mAP IoU>0.75: 45.18 -> 46.11Improved
increase mixture components K from 4 to 8 (ablation)Ours_gmm · VOC07mAP IoU>0.75: 46.11 -> 45.69Hurt
raise input resolution from 300 to 512 (ablation)Ours_gmm · VOC07mAP IoU>0.75: 46.11 -> 48.23Improved
use Type-2 instead of Type-1 classification loss (ablation)Ours_gmm · MS-COCOmAP at 7k images: 30.51 -> 29.55Hurt
use Type-2 instead of Type-1 classification loss (ablation)Ours_gmm · VOC07mAP at 4k images: 69.43 -> 69.40No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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