DenseWorld-1M: Towards Detailed Dense Grounded Caption in the Real World

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Xiangtai, Zhang, Tao, Li, Yanwei, Yuan, Haobo, Chen, Shihao, Zhou, Yikang, Meng, Jiahao, Sun, Yueyi, Xu, Shilin, Qi, Lu, Cheng, Tianheng, Lin, Yi, Huang, Zilong, Huang, Wenhao, Feng, Jiashi, Shi, Guang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908428350259200
author Li, Xiangtai
Zhang, Tao
Li, Yanwei
Yuan, Haobo
Chen, Shihao
Zhou, Yikang
Meng, Jiahao
Sun, Yueyi
Xu, Shilin
Qi, Lu
Cheng, Tianheng
Lin, Yi
Huang, Zilong
Huang, Wenhao
Feng, Jiashi
Shi, Guang
author_facet Li, Xiangtai
Zhang, Tao
Li, Yanwei
Yuan, Haobo
Chen, Shihao
Zhou, Yikang
Meng, Jiahao
Sun, Yueyi
Xu, Shilin
Qi, Lu
Cheng, Tianheng
Lin, Yi
Huang, Zilong
Huang, Wenhao
Feng, Jiashi
Shi, Guang
contents Multimodal Large Language Models (MLLMs) demonstrate a complex understanding of scenes, benefiting from large-scale and high-quality datasets. Most existing caption datasets lack the ground locations and relations for visual entities. Several grounded caption datasets face the problems of missing detailed descriptions, relations, and massive object descriptions on high-resolution images. To fill this gap for the community, we present DenseWorld-1M, the first massive, detailed, dense grounded caption dataset in the real world. We design a three-stage labeling pipeline, containing open-world perception, detailed object caption generation, and dense caption merging. The first stage obtains entity-level masks and labels. The second stage generates the object-level, detailed captions with the guidance of masks and labels from the first stage. The final stage merges object captions and masks into spatial and relational dense captions. To accelerate the labeling process and improve caption quality, we present two VLM models: the Detailed Region Caption model and the Spatial Caption Merging model. Extensive experiments on various settings, including vision-language understanding, visual grounding, and region caption generation, demonstrate the effectiveness of our DenseWorld-1M dataset and labeling models.
format Preprint
id arxiv_https___arxiv_org_abs_2506_24102
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle DenseWorld-1M: Towards Detailed Dense Grounded Caption in the Real World
Li, Xiangtai
Zhang, Tao
Li, Yanwei
Yuan, Haobo
Chen, Shihao
Zhou, Yikang
Meng, Jiahao
Sun, Yueyi
Xu, Shilin
Qi, Lu
Cheng, Tianheng
Lin, Yi
Huang, Zilong
Huang, Wenhao
Feng, Jiashi
Shi, Guang
Computer Vision and Pattern Recognition
Multimodal Large Language Models (MLLMs) demonstrate a complex understanding of scenes, benefiting from large-scale and high-quality datasets. Most existing caption datasets lack the ground locations and relations for visual entities. Several grounded caption datasets face the problems of missing detailed descriptions, relations, and massive object descriptions on high-resolution images. To fill this gap for the community, we present DenseWorld-1M, the first massive, detailed, dense grounded caption dataset in the real world. We design a three-stage labeling pipeline, containing open-world perception, detailed object caption generation, and dense caption merging. The first stage obtains entity-level masks and labels. The second stage generates the object-level, detailed captions with the guidance of masks and labels from the first stage. The final stage merges object captions and masks into spatial and relational dense captions. To accelerate the labeling process and improve caption quality, we present two VLM models: the Detailed Region Caption model and the Spatial Caption Merging model. Extensive experiments on various settings, including vision-language understanding, visual grounding, and region caption generation, demonstrate the effectiveness of our DenseWorld-1M dataset and labeling models.
title DenseWorld-1M: Towards Detailed Dense Grounded Caption in the Real World
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2506.24102