AI Paper Atlas

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

The Functional Correspondence Problem

Zihang Lai, Senthil Purushwalkam, Abhinav Gupta

ICCV 2021 · Detection and segmentation › Open-set and few-shot detection · approach: Visual relationship and scene graph detection · 13 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 task-embedding-conditioned gating (vs uniform gating) (ablation)modular network · FunKPoint functional correspondencePCK: 52.8 -> 58.4Improved
add task-embedding-conditioned gating (vs single learned gating) (ablation)modular network · FunKPoint functional correspondencePCK: 52.5 -> 58.4Improved
train on FunKPoint with modular network vs ImageNet featuresResNet50 conv4_x · FunKPoint functional correspondencePCK: 22.0 -> 58.4Improved
initialize trunk from ImageNet instead of DINO (ablation)task-driven model · FunKPoint functional correspondencePCK: 43.5 -> 58.4Improved
DINO-initialized model vs ImageNet baselinetask-driven model · FunKPoint functional correspondencePCK: 22.0 -> 43.5Improved
use functional correspondence representation vs ImageNetlinear probe · Few-shot classification, 1-shotaccuracy: 44.68 -> 47.46Improved
use functional correspondence representation vs ImageNetlinear probe · Few-shot classification, 2-shotaccuracy: 52.52 -> 55.68Improved
use functional correspondence representation vs ImageNetlinear probe · Few-shot classification, 5-shotaccuracy: 54.63 -> 56.32Improved
use functional correspondence vs finetuning on FunKPoint object classificationlinear probe · Few-shot classification, 1-shotaccuracy: 45.03 -> 47.46Improved
use functional representation vs ImageNetgrasp network · Grasp datasetaccuracy: 88.17 -> 89.85Improved
use functional representation vs ImageNet finetuned on FunKPointgrasp network · Grasp datasetaccuracy: 88.64 -> 89.85Improved

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