Make Graph-based Referring Expression Comprehension Great Again through Expression-guided Dynamic Gating and Regression

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Ke, Jingcheng, Wang, Dele, Chen, Jun-Cheng, Jhuo, I-Hong, Lin, Chia-Wen, Lin, Yen-Yu
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917769437511680
author Ke, Jingcheng
Wang, Dele
Chen, Jun-Cheng
Jhuo, I-Hong
Lin, Chia-Wen
Lin, Yen-Yu
author_facet Ke, Jingcheng
Wang, Dele
Chen, Jun-Cheng
Jhuo, I-Hong
Lin, Chia-Wen
Lin, Yen-Yu
contents One common belief is that with complex models and pre-training on large-scale datasets, transformer-based methods for referring expression comprehension (REC) perform much better than existing graph-based methods. We observe that since most graph-based methods adopt an off-the-shelf detector to locate candidate objects (i.e., regions detected by the object detector), they face two challenges that result in subpar performance: (1) the presence of significant noise caused by numerous irrelevant objects during reasoning, and (2) inaccurate localization outcomes attributed to the provided detector. To address these issues, we introduce a plug-and-adapt module guided by sub-expressions, called dynamic gate constraint (DGC), which can adaptively disable irrelevant proposals and their connections in graphs during reasoning. We further introduce an expression-guided regression strategy (EGR) to refine location prediction. Extensive experimental results on the RefCOCO, RefCOCO+, RefCOCOg, Flickr30K, RefClef, and Ref-reasoning datasets demonstrate the effectiveness of the DGC module and the EGR strategy in consistently boosting the performances of various graph-based REC methods. Without any pretaining, the proposed graph-based method achieves better performance than the state-of-the-art (SOTA) transformer-based methods.
format Preprint
id arxiv_https___arxiv_org_abs_2409_03385
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Make Graph-based Referring Expression Comprehension Great Again through Expression-guided Dynamic Gating and Regression
Ke, Jingcheng
Wang, Dele
Chen, Jun-Cheng
Jhuo, I-Hong
Lin, Chia-Wen
Lin, Yen-Yu
Computer Vision and Pattern Recognition
Multimedia
One common belief is that with complex models and pre-training on large-scale datasets, transformer-based methods for referring expression comprehension (REC) perform much better than existing graph-based methods. We observe that since most graph-based methods adopt an off-the-shelf detector to locate candidate objects (i.e., regions detected by the object detector), they face two challenges that result in subpar performance: (1) the presence of significant noise caused by numerous irrelevant objects during reasoning, and (2) inaccurate localization outcomes attributed to the provided detector. To address these issues, we introduce a plug-and-adapt module guided by sub-expressions, called dynamic gate constraint (DGC), which can adaptively disable irrelevant proposals and their connections in graphs during reasoning. We further introduce an expression-guided regression strategy (EGR) to refine location prediction. Extensive experimental results on the RefCOCO, RefCOCO+, RefCOCOg, Flickr30K, RefClef, and Ref-reasoning datasets demonstrate the effectiveness of the DGC module and the EGR strategy in consistently boosting the performances of various graph-based REC methods. Without any pretaining, the proposed graph-based method achieves better performance than the state-of-the-art (SOTA) transformer-based methods.
title Make Graph-based Referring Expression Comprehension Great Again through Expression-guided Dynamic Gating and Regression
topic Computer Vision and Pattern Recognition
Multimedia
url https://arxiv.org/abs/2409.03385