Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

CompoDiff: Versatile Composed Image Retrieval With Latent Diffusion

About

This paper proposes a novel diffusion-based model, CompoDiff, for solving zero-shot Composed Image Retrieval (ZS-CIR) with latent diffusion. This paper also introduces a new synthetic dataset, named SynthTriplets18M, with 18.8 million reference images, conditions, and corresponding target image triplets to train CIR models. CompoDiff and SynthTriplets18M tackle the shortages of the previous CIR approaches, such as poor generalizability due to the small dataset scale and the limited types of conditions. CompoDiff not only achieves a new state-of-the-art on four ZS-CIR benchmarks, including FashionIQ, CIRR, CIRCO, and GeneCIS, but also enables a more versatile and controllable CIR by accepting various conditions, such as negative text, and image mask conditions. CompoDiff also shows the controllability of the condition strength between text and image queries and the trade-off between inference speed and performance, which are unavailable with existing CIR methods. The code and dataset are available at https://github.com/navervision/CompoDiff

Geonmo Gu, Sanghyuk Chun, Wonjae Kim, HeeJae Jun, Yoohoon Kang, Sangdoo Yun• 2023

Related benchmarks

TaskDatasetResultRank
Composed Image RetrievalCIRR (test)
Recall@132.39
786
Composed Image RetrievalFashionIQ (val)
Average Recall@1041.12
601
Composed Image RetrievalCIRCO (test)
mAP@1017.71
360
Composed Image RetrievalFashion-IQ (test)
Average Recall@100.3981
176
Composed Image RetrievalFashion-IQ
Average Recall@5051.71
129
Composed Image Retrieval (Image-Text to Image)CIRR
Recall@134.7
128
Composed Image RetrievalCIRCO
mAP@515.33
96
Composed Image Retrieval (Image-Text to Image)FashionIQ
Recall@1036
39
Composed Image RetrievalCIRCO 1.0 (test)
mAP@512.6
36
Composed Image RetrievalCIRR Subset (test)
R@155
33
Showing 10 of 24 rows

Other info

Code

Follow for update