AI Paper Atlas

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

A Zero-Shot Framework for Sketch based Image Retrieval

Sasi Kiran Yelamarthi, Shiva Krishna Reddy, Ashish Mishra, Anurag Mittal

ECCV 2018 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Transductive and semantic-dictionary zero-shot · 204 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
evaluate on unseen classes (zero-shot) instead of traditional splitDSH · SketchyPrecision@200: 0.866 -> 0.153Hurt
evaluate on unseen classes instead of traditional splitSiamese-2 · SketchyPrecision@200: 0.690 -> 0.251Hurt
evaluate on unseen classes instead of traditional splitCoarse-grained triplet · SketchymAP@200: 0.573 -> 0.083Hurt
use generative CVAE instead of SAESAE · Sketchy ZS-SBIR (zero-shot split)mAP@200: 0.136 -> 0.225Improved
use generative CVAE instead of Siamese-2Siamese-2 · Sketchy ZS-SBIR (zero-shot split)Precision@200: 0.251 -> 0.333Improved
use CAAE instead of Siamese-2Siamese-2 · Sketchy ZS-SBIR (zero-shot split)Precision@200: 0.251 -> 0.260Improved
use CVAE instead of CAAECAAE · Sketchy ZS-SBIR (zero-shot split)mAP@200: 0.156 -> 0.225Improved
add reconstruction loss of sketch features (ablation)CVAE · Sketchy ZS-SBIR (zero-shot split)Precision@200: 3% improvement on the precisionImproved
coarse-grained instead of fine-grained negatives for triplet trainingTriplet network · Sketchy ZS-SBIR (zero-shot split)Precision@200: 0.155 -> 0.169Improved
replace Direct Regression with ESZSL regularizationDirect Regression · Sketchy ZS-SBIR (zero-shot split)Precision@200: 0.066 -> 0.187Improved
add reconstructibility regularization (SAE) to linear mappingDirect Regression · Sketchy ZS-SBIR (zero-shot split)Precision@200: 0.066 -> 0.238Improved

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