Make Graph-based Referring Expression Comprehension Great Again through Expression-guided Dynamic Gating and Regression
This work solves the challenge of improving graph-based methods for REC, which is important for computer vision applications like image understanding, and it is incremental as it builds on existing graph-based approaches with novel modules.
The paper tackles the problem of referring expression comprehension (REC) by addressing noise from irrelevant objects and inaccurate localization in graph-based methods, introducing a dynamic gate constraint module and expression-guided regression strategy that achieve better performance than state-of-the-art transformer-based methods without pre-training.
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.