VLPrompt: Vision-Language Prompting for Panoptic Scene Graph Generation

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zhou, Zijian, Shi, Miaojing, Caesar, Holger
Natura: Preprint
Pubblicazione: 2023
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916294740148224
author Zhou, Zijian
Shi, Miaojing
Caesar, Holger
author_facet Zhou, Zijian
Shi, Miaojing
Caesar, Holger
contents Panoptic Scene Graph Generation (PSG) aims at achieving a comprehensive image understanding by simultaneously segmenting objects and predicting relations among objects. However, the long-tail problem among relations leads to unsatisfactory results in real-world applications. Prior methods predominantly rely on vision information or utilize limited language information, such as object or relation names, thereby overlooking the utility of language information. Leveraging the recent progress in Large Language Models (LLMs), we propose to use language information to assist relation prediction, particularly for rare relations. To this end, we propose the Vision-Language Prompting (VLPrompt) model, which acquires vision information from images and language information from LLMs. Then, through a prompter network based on attention mechanism, it achieves precise relation prediction. Our extensive experiments show that VLPrompt significantly outperforms previous state-of-the-art methods on the PSG dataset, proving the effectiveness of incorporating language information and alleviating the long-tail problem of relations. Code is available at \url{https://github.com/franciszzj/TP-SIS}.
format Preprint
id arxiv_https___arxiv_org_abs_2311_16492
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle VLPrompt: Vision-Language Prompting for Panoptic Scene Graph Generation
Zhou, Zijian
Shi, Miaojing
Caesar, Holger
Computer Vision and Pattern Recognition
Panoptic Scene Graph Generation (PSG) aims at achieving a comprehensive image understanding by simultaneously segmenting objects and predicting relations among objects. However, the long-tail problem among relations leads to unsatisfactory results in real-world applications. Prior methods predominantly rely on vision information or utilize limited language information, such as object or relation names, thereby overlooking the utility of language information. Leveraging the recent progress in Large Language Models (LLMs), we propose to use language information to assist relation prediction, particularly for rare relations. To this end, we propose the Vision-Language Prompting (VLPrompt) model, which acquires vision information from images and language information from LLMs. Then, through a prompter network based on attention mechanism, it achieves precise relation prediction. Our extensive experiments show that VLPrompt significantly outperforms previous state-of-the-art methods on the PSG dataset, proving the effectiveness of incorporating language information and alleviating the long-tail problem of relations. Code is available at \url{https://github.com/franciszzj/TP-SIS}.
title VLPrompt: Vision-Language Prompting for Panoptic Scene Graph Generation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2311.16492