AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

Scaling Open-Vocabulary Image Segmentation with Image-Level Labels

Golnaz Ghiasi, Xiuye Gu, Yin Cui, Tsung-Yi Lin

ECCV 2022 · Detection and segmentation › Open-set and few-shot detection · approach: Open-vocabulary semantic segmentation · 473 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 predicted mask proposals at inference (ablation)OpenSeg (COCO) · PASCAL-Context-59 / ADE20k-150mIoU: 32.1 -> 42.1 (PC-59); 16.4 -> 21.1 (A-150)Improved
use ground-truth masks as proposals instead of predicted masksOpenSeg (COCO) · PASCAL-Context-59 / ADE20k-150mIoU: 42.1 -> 49.3 (PC-59); 21.1 -> 27.5 (A-150)Improved
add class-agnostic segmentation loss to grounding loss (ablation)OpenSeg (COCO) · PASCAL-Context-59 / ADE20k-847mIoU: 15.6 -> 42.1 (PC-59); 2.5 -> 6.3 (A-847)Improved
scale training data from COCO to COCO+Localized Narratives with pseudo mask labelsOpenSeg · ADE20k-150 / PASCAL-Context-459mIoU: 21.1 -> 24.8 (A-150); 9.0 -> 11.2 (PC-459)Improved
compute segmentation loss on pseudo-labeled Localized Narratives examples (ablation)OpenSeg (COCO+Loc. Narr.) · PASCAL-Context-59 / ADE20k-150mIoU: 47.3 -> 45.9 (PC-59); 22.4 -> 24.8 (A-150)Mixed
add ensembling and prompt engineering of text queriesOpenSeg (EfficientNet-B7) · ADE20k-150mIoU: 26.4 -> 28.6Improved
replace per-pixel class-label training (LSeg+) with caption supervision (OpenSeg), ResNet101 COCOLSeg+ · ADE20k-150mIoU: 13.0 -> 15.3Improved
fine-tune ALIGN with per-pixel COCO labels (LSeg+)ALIGN · ADE20k-847mIoU: 1.0 mIoU worse than ALIGNHurt
initialize backbone from ALIGN instead of random (ablation)OpenSeg (COCO+Loc. Narr.) · ADE20k-150mIoU: 18.6 -> 24.8Improved
initialize backbone from ALIGN instead of NoisyStudent (ablation)OpenSeg (COCO+Loc. Narr.) · ADE20k-150mIoU: 24.4 -> 24.8No clear effect
filter captions to nouns and adjectives instead of all words (ablation)OpenSeg (COCO) · ADE20k-150 / PASCAL-Context-59mIoU: 20.0 -> 21.1 (A-150); 41.3 -> 42.1 (PC-59)Improved
randomly drop caption words with keep probability 0.75 instead of keeping all (ablation)OpenSeg (COCO+Loc. Narr.) · ADE20k-150mIoU: 22.7 -> 24.8Improved

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