AI Paper Atlas

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

GH-ESD: Grounded Hypothesis-Driven Error Slice Discovery for Instance-Level Vision Tasks

Wei Zhang, Chaoqun Wang, Zixuan Guan, Ping Kao, Pengfei Zhao, Peng Wu, Sifeng He

ECCV 2026 · Detection and segmentation › Open-set and few-shot detection · approach: Open-set and unknown object detection · 0 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 tag-based slices (HiBug2) with grounded language hypotheses and statistical verificationGH-ESD · GESD detectionPrecision@10: 0.63 -> 0.73Improved
use same Gemini backend for HiBug2 as GH-ESD (ablation)HiBug2 · GESD detectionPrecision@10: 0.63 -> 0.63No clear effect
add data-driven hypotheses to knowledge-driven ones (ablation)GH-ESD · GESD detectionPrecision@10: 0.62 -> 0.73Improved
use instance-level grounding instead of image-level aggregation (ablation)GH-ESD · faces partially hidden by objectsPrecision@10: 0.60 -> 0.80Improved
use trend-based verification instead of error-rate thresholding (ablation)GH-ESD · GESDF1: average 3.7% F1 gainImproved
use Spearman correlation test instead of max-slope trend (ablation)GH-ESD · GESDF1: 0.77 -> 0.78No clear effect
use yes/no logits instead of direct numeric score (ablation)GH-ESD · GESD subsetmAP: 0.23 -> 0.47Improved
use yes/no logits instead of yes/no/not-sure categorical (ablation)GH-ESD · GESD subsetmAP: 0.18 -> 0.47Improved
use Qwen2.5-VL-7B instead of SigLIP for matching (ablation)GH-ESD · NICO++ (95%)Precision@10: 0.51 -> 0.69Improved
use GPT-4o instead of Qwen2.5-VL-7B for matching (ablation)GH-ESD · NICO++ (95%)Precision@10: 0.69 -> 0.71No clear effect
swap hypothesis LLM from Gemini-2.5-Pro to GPT-4o or Claude-3.5-Sonnet (ablation)GH-ESD · NICO++ (95%)Precision@10: 0.69 vs 0.68 vs 0.67No clear effect
slice-aware GroupDRO repair using GH-ESD sliceFaster R-CNN · COCO bicyclemAP-bicycle: 28.58 -> 33.59Improved

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