AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Hyperpixel Flow: Semantic Correspondence With Multi-Layer Neural Features

Juhong Min, Jongmin Lee, Jean Ponce, Minsu Cho

ICCV 2019 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 92 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
use regularized Hough matching instead of nearest neighbor (ablation)HPF ResNet-101 · PF-PASCALPCK@0.1: 69.0 -> 81.4Improved
use RHM with similarity exponent d=2 instead of d=1 (ablation)HPF ResNet-101 · PF-PASCALPCK@0.1: 81.4 -> 84.4Improved
raise exponent d from 3 to 5 (ablation)HPF ResNet-101 · PF-PASCALPCK@0.1: 84.8 -> 84.5No clear effect
use beam-searched layers instead of all layers (ablation)HPF ResNet-101 · PF-PASCALPCK@0.1: 74.5 -> 84.8Improved
use beam-searched layers instead of all layers (ablation)HPF ResNet-50 · PF-PASCALPCK@0.1: 70.1 -> 83.4Improved
use layer search instead of random 8 layers (ablation)HPF ResNet-101 · PF-PASCALPCK@0.05: 44.5 -> 60.1Improved
use layer search instead of random 8 layers (ablation)HPF ResNet-101 · PF-WILLOWPCK@0.1: 62.4 -> 74.4Improved
search layers with only 1 pair per class instead of all 308 validation pairs (ablation)HPF ResNet-101 · PF-PASCALPCK@0.1: 84.8 -> 83.9No clear effect
use FCN backbone pretrained on PASCAL VOC instead of ResNet-101HPF · PF-PASCALPCK@0.1: 84.8 -> 88.3Improved
use ResNet-101 instead of ResNet-50HPF · PF-PASCALPCK@0.1: 83.4 -> 84.8Improved
use fewer layers (4,7,11,12,13) with ResNet-50 for speedHPF · PF-PASCALPCK@0.1 and time: 83.4 -> 81.1Mixed
finetune NC-Net on SPair-71kNC-Net · SPair-71kPCK@0.1 (bbox): 26.4 -> 20.1Hurt

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