AI Paper Atlas

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

ClevrTex: A Texture-Rich Benchmark for Unsupervised Multi-Object Segmentation

Laurynas Karazija, Iro Laina, Christian Rupprecht

NeurIPS 2021 · Detection and segmentation › Open-set and few-shot detection · approach: Amodal and open-world instance segmentation · 3 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
Train and evaluate on textured ClevrTex instead of CLEVRSPAIR* · CLEVR vs ClevrTexmIoU (%): 65.95 ± 4.02 -> 0.0 ± 0.0Hurt
Train and evaluate on textured ClevrTex instead of CLEVRGNM · CLEVR vs ClevrTexmIoU (%): 59.92 ± 3.72 -> 42.25 ± 0.18Hurt
Use varied textured backgrounds with simple colored objects instead of textured objects and backgroundSA · VarBG vs ClevrTexmIoU (%): 22.58 ± 2.07 -> 62.57Improved
Use a consistent grass background instead of varied textured backgroundsGNM · GrassBG vs ClevrTexmIoU (%): 42.25 ± 0.18 -> 53.15Improved
Use a consistent grass background instead of varied textured backgroundsSPACE · GrassBG vs ClevrTexmIoU (%): 9.14 ± 3.46 -> 33.85Improved
Use textured objects on a plain background instead of textured objects and backgroundSPAIR* · PlainBG vs ClevrTexmIoU (%): 0.0 ± 0.0 -> 39.32Mixed
Use varied textured backgrounds with simple colored objects instead of textured objects and backgroundSPAIR* · VarBG vs ClevrTexmIoU (%): 0.0 ± 0.0 -> 0.00No clear effect
Use one consistent grass background instead of varied backgroundspixel-space methods · GrassBG vs ClevrTexmIoU (%)No clear effect
Use one common material on every object and the backgroundmodels trained on ClevrTex · CAMOmIoU (%)Hurt
Test on unseen materials and shapesmodels trained on ClevrTex · OODmIoU (%)Mixed

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