AI Paper Atlas

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

Polysemy Deciphering Network for Robust Human–Object Interaction Detection

Xubin Zhong, Changxing Ding, Xian Qu, Dacheng Tao

IJCV 2021 · Detection and segmentation › Open-set and few-shot detection · approach: Human-object interaction detection · 60 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 Polysemy-Aware Modal Fusion (PAMF) (ablation)baseline with SH classifiers · HICO-DET Full DTmAP: 17.57 -> 18.86Improved
add Language Prior-based Feature Augmentation (LPFA) (ablation)baseline + PAMF · HICO-DET Full DTmAP: 18.86 -> 19.38Improved
add Language Prior-guided Channel Attention (LPCA) (ablation)baseline + PAMF + LPFA · HICO-DET Full DTmAP: 19.38 -> 20.71Improved
replace object-shared classifiers with clustering-based object-specific (CSP) classifiers (ablation)PD-Net without INet · HICO-DET Full DTmAP: 20.71 -> 21.77Improved
add INet (ablation)PD-Net · HICO-DET Full DTmAP: 21.77 -> 22.37Improved
add INet (ablation)PD-Net · HICO-DET Full KOmAP: 26.98 -> 26.86No clear effect
add object word embedding to the language prior (ablation)PD-Net · HICO-DET Full DTmAP: 19.98 -> 22.37Improved
add auxiliary supervision S_au to LPCA (ablation)PD-Net · HICO-DET Full DTmAP: 19.41 -> 22.37Improved
use LPCA instead of FiLM (ablation)PD-Net · HICO-DET Full DTmAP: 20.90 -> 22.37Improved
add plain channel attention refinement C_att in LPCA (ablation)PD-Net · HICO-DET Full DTmAP: 21.99 -> 22.37Improved
use CSP classifiers instead of SH classifiers (ablation)PD-Net · HICO-DET Full DTmAP: 21.06 -> 22.37Improved
use object-shared instead of object-specific classifiers on rare HOIs (ablation)PD-Net · HICO-DET Rare DTmAP: 15.03 -> 16.45Mixed

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