AI Paper Atlas

Research map / Detection and segmentation / Weakly supervised detection and segmentation

APL: Anchor-based Prompt Learning for One-stage Weakly Supervised Referring Expression Comprehension

Yaxin Luo, Jiayi Ji, Xiaofu Chen, Yuxin Zhang, Tianhe Ren, Gen Luo

ECCV 2024 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Weakly supervised semantic segmentation · 3 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 anchor-based prompt encoder (position, color, category prompts) (ablation)RefCLIP-style baseline · RefCOCO valaccuracy (%): 60.36 -> 62.89Improved
add text reconstruction loss (ablation)APL · RefCOCO+ valaccuracy (%): 41.54 -> 42.31Improved
add visual alignment loss (ablation)APL · RefCOCO valaccuracy (%): 63.01 -> 64.51Improved
use APL instead of RefCLIPone-stage weakly supervised REC · RefCOCO valaccuracy (%): 60.36 -> 64.51Improved
use APL instead of RefCLIPinference speed · inference speedfps: 31.3fps -> 26.7fpsHurt
use APL pseudo-labels instead of RefCLIP pseudo-labels to train SimRECSimREC student · ReferIt testaccuracy (%): 42.33 -> 45.22Improved
use APL pseudo-labels instead of RefCLIP pseudo-labels to train TransVGTransVG student · RefCOCO valaccuracy (%): 64.08 -> 64.86No clear effect
use text description instead of positional function for spatial prompt (ablation)APE · RefCOCOaccuracy (%): -6.63%Hurt
use BLIP-generated region captions as text prompts (ablation)APE · RefCOCOaccuracy (%)Hurt
extend APL to weakly supervised referring segmentation with YOLOv5-SegAPL · RefCOCO val (RES)mask accuracy (%): 31.17 -> 55.92Improved

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