AI Paper Atlas

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

Labels4Free: Unsupervised Segmentation Using StyleGAN

Rameen Abdal, Peihao Zhu, Niloy J. Mitra, Peter Wonka

ICCV 2021 · Detection and segmentation › Open-set and few-shot detection · approach: Amodal and open-world instance segmentation · 67 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 StyleGAN-generated in-domain backgrounds instead of MIT Places backgrounds (ablation)Labels4Free · FFHQ samplesmIOU: 0.34 -> 0.90Improved
use in-domain backgrounds instead of MIT Places backgrounds (ablation)Labels4Free · LSUN-Horse samplesmIOU: 0.14 -> 0.75Improved
use in-domain backgrounds instead of MIT Places backgrounds (ablation)Labels4Free · LSUN-Cat samplesmIOU: 0.20 -> 0.81Improved
use in-domain backgrounds instead of MIT Places backgrounds (ablation)Labels4Free · LSUN-Car samplesmIOU: 0.34 -> 0.89Improved
use features from multiple generator layers instead of only the last layer (ablation)Labels4Free · FFHQ samplesmIOU: 0.34 -> 0.90Improved
use multiple-layer features instead of only the last layer (ablation)Labels4Free · LSUN-Horse samplesmIOU: 0.33 -> 0.75Improved
use multiple-layer features instead of only the last layer (ablation)Labels4Free · LSUN-Cat samplesmIOU: 0.31 -> 0.81Improved
use multiple-layer features instead of only the last layer (ablation)Labels4Free · LSUN-Car samplesmIOU: 0.42 -> 0.89Improved
Labels4Free instead of unsupervised PSegPSeg · FFHQ samples (truncation 1.0)mIOU: 0.66 -> 0.82Improved
Labels4Free instead of unsupervised PSegPSeg · LSUN-Horse samplesmIOU: 0.45 -> 0.75Improved

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