AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

IRGen: Generative Modeling for Image Retrieval

Yidan Zhang, Ting Zhang, Dong Chen, Yujing Wang, Qi Chen, Xing Xie, Hao Sun, Weiwei Deng, Qi Zhang, Fan Yang, Mao Yang, Qingmin Liao and 2 more

ECCV 2024 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 25 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 semantic RQ identifiers instead of random codes (ablation)IRGen · In-shop ClothesPrecision@10: 75.4 -> 89.7Improved
use proposed semantic tokenizer instead of hierarchical k-means identifiers (ablation)IRGen · In-shop ClothesPrecision@10: 81.7 -> 89.7Improved
use PQ codes as identifiers instead of residual quantization (ablation)IRGen · In-shop ClothesPrecision@10: 89.7 -> 66.1Hurt
use RQ-VAE tokenizer with long sequences (ablation)IRGen · In-shop ClothesPrecisionHurt
increase identifier length from 2 to 4 (ablation)IRGen · In-shop ClothesPrecision@1: 72.1 -> 92.4Improved
increase identifier length from 4 to 8 (ablation)IRGen · In-shop ClothesPrecision@1: 92.4 -> 92.9No clear effect
add autoregressive decoder on top of RQ prefix-tree beam search (ablation)IRGen · ImageNetMAP@100: 56.7 -> 76.0Improved
add partial reconstruction losses to classification loss in tokenizer training (ablation)IRGen tokenizer · In-shop ClothesPrecision@10: 85.1 -> 87.0Improved
IRGen generative retrieval versus FT-CLIP with ANN or linear scanIRGen · In-shop ClothesPrecision@10: 66.8 -> 87.0Improved
IRGen versus ScaNN/SPANN baselines on recall at larger KIRGen · In-shop ClothesRecall@10: 97.8 -> 96.8Hurt
retrieve with IRGen on newly added unseen data without retrainingIRGen · ImageNet unseen 5% dataPrecision@1: 70.6 -> 77.0Improved

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