Category-level Text-to-Image Retrieval Improved: Bridging the Domain Gap with Diffusion Models and Vision Encoders

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F25%3A00387644" target="_blank" >RIV/68407700:21230/25:00387644 - isvavai.cz</a>

Výsledek na webu

<a href="https://bmva-archive.org.uk/bmvc/2025/assets/papers/Paper_701/paper.pdf" target="_blank" >https://bmva-archive.org.uk/bmvc/2025/assets/papers/Paper_701/paper.pdf</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Category-level Text-to-Image Retrieval Improved: Bridging the Domain Gap with Diffusion Models and Vision Encoders

Popis výsledku v původním jazyce

This work explores text-to-image retrieval for queries that specify or describe a semantic category. While vision-and-language models (VLMs) like CLIP offer a straightforward open-vocabulary solution, they map text and images to distant regions in the representation space, limiting retrieval performance. To bridge this modality gap, we propose a two-step approach. First, we transform the text query into a visual query using a generative diffusion model. Then, we estimate image-to-image similarity with a vision model. Additionally, we introduce an aggregation network that combines multiple generated images into a single vector representation and fuses similarity scores across both query modalities. Our approach leverages advancements in vision encoders, VLMs, and text-to-image generation models. Extensive evaluations show that it consistently outperforms retrieval methods relying solely on text queries. Source code is available at: https://github.com/faixan-khan/cletir

Název v anglickém jazyce

Category-level Text-to-Image Retrieval Improved: Bridging the Domain Gap with Diffusion Models and Vision Encoders

Popis výsledku anglicky

This work explores text-to-image retrieval for queries that specify or describe a semantic category. While vision-and-language models (VLMs) like CLIP offer a straightforward open-vocabulary solution, they map text and images to distant regions in the representation space, limiting retrieval performance. To bridge this modality gap, we propose a two-step approach. First, we transform the text query into a visual query using a generative diffusion model. Then, we estimate image-to-image similarity with a vision model. Additionally, we introduce an aggregation network that combines multiple generated images into a single vector representation and fuses similarity scores across both query modalities. Our approach leverages advancements in vision encoders, VLMs, and text-to-image generation models. Extensive evaluations show that it consistently outperforms retrieval methods relying solely on text queries. Source code is available at: https://github.com/faixan-khan/cletir

Druh

O - Ostatní výsledky

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/GM21-28830M" target="_blank" >GM21-28830M: Učení Univerzální Vizuální Reprezentace s Omezenou Supervizí</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2025

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů