Rethinking 3D Dense Caption and Visual Grounding in A Unified Framework through Prompt-based Localization

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Luo, Yongdong, Lin, Haojia, Zheng, Xiawu, Jiang, Yigeng, Chao, Fei, Hu, Jie, Jiang, Guannan, Zhang, Songan, Ji, Rongrong
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916529319182336
author Luo, Yongdong
Lin, Haojia
Zheng, Xiawu
Jiang, Yigeng
Chao, Fei
Hu, Jie
Jiang, Guannan
Zhang, Songan
Ji, Rongrong
author_facet Luo, Yongdong
Lin, Haojia
Zheng, Xiawu
Jiang, Yigeng
Chao, Fei
Hu, Jie
Jiang, Guannan
Zhang, Songan
Ji, Rongrong
contents 3D Visual Grounding (3DVG) and 3D Dense Captioning (3DDC) are two crucial tasks in various 3D applications, which require both shared and complementary information in localization and visual-language relationships. Therefore, existing approaches adopt the two-stage "detect-then-describe/discriminate" pipeline, which relies heavily on the performance of the detector, resulting in suboptimal performance. Inspired by DETR, we propose a unified framework, 3DGCTR, to jointly solve these two distinct but closely related tasks in an end-to-end fashion. The key idea is to reconsider the prompt-based localization ability of the 3DVG model. In this way, the 3DVG model with a well-designed prompt as input can assist the 3DDC task by extracting localization information from the prompt. In terms of implementation, we integrate a Lightweight Caption Head into the existing 3DVG network with a Caption Text Prompt as a connection, effectively harnessing the existing 3DVG model's inherent localization capacity, thereby boosting 3DDC capability. This integration facilitates simultaneous multi-task training on both tasks, mutually enhancing their performance. Extensive experimental results demonstrate the effectiveness of this approach. Specifically, on the ScanRefer dataset, 3DGCTR surpasses the state-of-the-art 3DDC method by 4.3% in CIDEr@0.5IoU in MLE training and improves upon the SOTA 3DVG method by 3.16% in Acc@0.25IoU. The codes are at https://github.com/Leon1207/3DGCTR.
format Preprint
id arxiv_https___arxiv_org_abs_2404_11064
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Rethinking 3D Dense Caption and Visual Grounding in A Unified Framework through Prompt-based Localization
Luo, Yongdong
Lin, Haojia
Zheng, Xiawu
Jiang, Yigeng
Chao, Fei
Hu, Jie
Jiang, Guannan
Zhang, Songan
Ji, Rongrong
Computer Vision and Pattern Recognition
Artificial Intelligence
3D Visual Grounding (3DVG) and 3D Dense Captioning (3DDC) are two crucial tasks in various 3D applications, which require both shared and complementary information in localization and visual-language relationships. Therefore, existing approaches adopt the two-stage "detect-then-describe/discriminate" pipeline, which relies heavily on the performance of the detector, resulting in suboptimal performance. Inspired by DETR, we propose a unified framework, 3DGCTR, to jointly solve these two distinct but closely related tasks in an end-to-end fashion. The key idea is to reconsider the prompt-based localization ability of the 3DVG model. In this way, the 3DVG model with a well-designed prompt as input can assist the 3DDC task by extracting localization information from the prompt. In terms of implementation, we integrate a Lightweight Caption Head into the existing 3DVG network with a Caption Text Prompt as a connection, effectively harnessing the existing 3DVG model's inherent localization capacity, thereby boosting 3DDC capability. This integration facilitates simultaneous multi-task training on both tasks, mutually enhancing their performance. Extensive experimental results demonstrate the effectiveness of this approach. Specifically, on the ScanRefer dataset, 3DGCTR surpasses the state-of-the-art 3DDC method by 4.3% in CIDEr@0.5IoU in MLE training and improves upon the SOTA 3DVG method by 3.16% in Acc@0.25IoU. The codes are at https://github.com/Leon1207/3DGCTR.
title Rethinking 3D Dense Caption and Visual Grounding in A Unified Framework through Prompt-based Localization
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2404.11064