AI Paper Atlas

Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning

SVIP: Semantically Contextualized Visual Patches for Zero-Shot Learning

Zhi Chen, Zecheng Zhao, Jingcai Guo, Jingjing Li, Zi Huang

ICCV 2025 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Generative feature synthesis for zero-shot · 14 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 self-supervised patch selection (ablation)SVIP · CUB (GZSL)harmonic mean H: 71.9 -> 75.0Improved
add self-supervised patch selection (ablation)SVIP · SUN (GZSL)harmonic mean H: 47.3 -> 50.7Improved
contextualize unrelated patches instead of removing them (PSC) (ablation)SVIP · CUB (GZSL)harmonic mean H: 72.6 -> 75.0Improved
contextualize unrelated patches instead of removing them (PSC) (ablation)SVIP · AwA2 (GZSL)harmonic mean H: 72.4 -> 74.9Improved
add JSD loss between original and contextualized predictions (ablation)SVIP · CUB (GZSL)harmonic mean H: 74.9 -> 75.0Improved
add JSD loss (ablation)SVIP · SUN (GZSL)harmonic mean H: 50.4 -> 50.7Improved
initialize learnable patch embedding from word embeddings (W2P) (ablation)SVIP · CUB (GZSL)harmonic mean H: 74.5 -> 75.0Improved
initialize learnable patch embedding from word embeddings (W2P) (ablation)SVIP · AwA2 (GZSL)harmonic mean H: 74.4 -> 74.9Improved
add patch-to-attribute mapping (P2A) (ablation)SVIP · CUB (GZSL)harmonic mean H: 72.9 -> 75.0Improved
add patch-to-attribute mapping (P2A) (ablation)SVIP · SUN (GZSL)harmonic mean H: 50.5 -> 50.7Improved
add all SVIP components to ViT baseline (ablation)SVIP · CUB (GZSL)harmonic mean H: 63.8 -> 75.0Improved
add all SVIP components to ViT baseline (ablation)SVIP · SUN (GZSL)harmonic mean H: 36.0 -> 50.7Improved

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