FG-CLIP 2: A Bilingual Fine-grained Vision-Language Alignment Model

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xie, Chunyu, Wang, Bin, Kong, Fanjing, Li, Jincheng, Liang, Dawei, Ao, Ji, Leng, Dawei, Yin, Yuhui
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918520505237504
author Xie, Chunyu
Wang, Bin
Kong, Fanjing
Li, Jincheng
Liang, Dawei
Ao, Ji
Leng, Dawei
Yin, Yuhui
author_facet Xie, Chunyu
Wang, Bin
Kong, Fanjing
Li, Jincheng
Liang, Dawei
Ao, Ji
Leng, Dawei
Yin, Yuhui
contents Fine-grained vision-language understanding requires precise alignment between visual content and linguistic descriptions, a capability that remains limited in current models, particularly in non-English settings. While models like CLIP perform well on global alignment, they often struggle to capture fine-grained details in object attributes, spatial relations, and linguistic expressions, with limited support for bilingual comprehension. To address these challenges, we introduce FG-CLIP 2, a bilingual vision-language model designed to advance fine-grained alignment for both English and Chinese. Our approach leverages rich fine-grained supervision, including region-text matching and long-caption modeling, alongside multiple discriminative objectives. We further introduce the Textual Intra-modal Contrastive (TIC) loss to better distinguish semantically similar captions. Trained on a carefully curated mixture of large-scale English and Chinese data, including a newly released 12M Chinese region-text dataset, FG-CLIP 2 achieves powerful bilingual performance. To enable rigorous evaluation, we present a new benchmark for Chinese multimodal understanding, featuring long-caption retrieval and bounding box classification. Extensive experiments on 29 datasets across 8 tasks show that FG-CLIP 2 outperforms existing methods, achieving state-of-the-art results in both languages. We release the model, code, and benchmark to facilitate future research on bilingual fine-grained vision-language alignment.
format Preprint
id arxiv_https___arxiv_org_abs_2510_10921
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle FG-CLIP 2: A Bilingual Fine-grained Vision-Language Alignment Model
Xie, Chunyu
Wang, Bin
Kong, Fanjing
Li, Jincheng
Liang, Dawei
Ao, Ji
Leng, Dawei
Yin, Yuhui
Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
Fine-grained vision-language understanding requires precise alignment between visual content and linguistic descriptions, a capability that remains limited in current models, particularly in non-English settings. While models like CLIP perform well on global alignment, they often struggle to capture fine-grained details in object attributes, spatial relations, and linguistic expressions, with limited support for bilingual comprehension. To address these challenges, we introduce FG-CLIP 2, a bilingual vision-language model designed to advance fine-grained alignment for both English and Chinese. Our approach leverages rich fine-grained supervision, including region-text matching and long-caption modeling, alongside multiple discriminative objectives. We further introduce the Textual Intra-modal Contrastive (TIC) loss to better distinguish semantically similar captions. Trained on a carefully curated mixture of large-scale English and Chinese data, including a newly released 12M Chinese region-text dataset, FG-CLIP 2 achieves powerful bilingual performance. To enable rigorous evaluation, we present a new benchmark for Chinese multimodal understanding, featuring long-caption retrieval and bounding box classification. Extensive experiments on 29 datasets across 8 tasks show that FG-CLIP 2 outperforms existing methods, achieving state-of-the-art results in both languages. We release the model, code, and benchmark to facilitate future research on bilingual fine-grained vision-language alignment.
title FG-CLIP 2: A Bilingual Fine-grained Vision-Language Alignment Model
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2510.10921