AI Paper Atlas

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

Learning Class Prototypes via Structure Alignment for Zero-Shot Recognition

Huajie Jiang, Ruiping Wang, Shiguang Shan, Xilin Chen

ECCV 2018 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Generative feature synthesis for zero-shot · 143 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 structure alignment of visual and semantic prototypes (ablation)CDL · ZSL on four datasets (aPY, AwA, CUB, SUNA)accuracyImproved
add the prototype learning term instead of fixing prototypes at class means (ablation)CDL · ZSL on four datasetsaccuracyImproved
add the unseen-class domain adaptation term (ablation)CDL · ZSL on four datasetsaccuracyImproved
combine visual and aligned space similarities instead of using the semantic space alone (ablation)CDL · aPY, CUB, SUNAaccuracyImproved
combine visual space with other spaces (ablation)CDL · AwAaccuracyHurt
recognize in the visual space instead of the semantic space (ablation)CDL · ZSL on four datasetsaccuracyImproved
use CDL instead of ALE for zero-shot recognitionALE · aPYaccuracy (%): 39.7 -> 43.0Improved
use CDL instead of SJE for zero-shot recognitionSJE · AwAaccuracy (%): 65.6 -> 69.9Improved
use CDL instead of ALE for zero-shot recognitionALE · SUNAaccuracy (%): 58.1 -> 63.6Improved
use CDL instead of SYNC on fine-grained CUBSYNC · CUBaccuracy (%): 55.6 -> 54.5Hurt
use CDL instead of ALE in generalized ZSLALE · AwA (generalized ZSL)harmonic mean H (%): 27.5 -> 40.6Improved
use CDL instead of ALE in generalized ZSLALE · aPY (generalized ZSL)harmonic mean H (%): 8.7 -> 28.1Improved

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