CompoDiff: Versatile Composed Image Retrieval With Latent Diffusion

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 trip...

Ausführliche Beschreibung

Bibliographische Detailangaben

Veröffentlicht in:	arXiv.org 2024-07
Hauptverfasser:	Gu, Geonmo, Chun, Sanghyuk, Kim, Wonjae, HeeJae Jun, Kang, Yoohoon, Sangdoo Yun
Format:	Artikel
Sprache:	eng
Schlagworte:	Datasets Diffusion Retrieval
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!