Position-Guided Text Prompt for Vision-Language Pre-Training

Jinpeng Wang · Pan Zhou · Mike Zheng Shou · Shuicheng Yan

West Building Exhibit Halls ABC 252
[ Abstract ] [ Project Page ]
[ Paper PDF [ Slides [ Poster
Thu 22 Jun 4:30 p.m. PDT — 6 p.m. PDT


Vision-Language Pre-Training (VLP) has shown promising capabilities to align image and text pairs, facilitating a broad variety of cross-modal learning tasks. However, we observe that VLP models often lack the visual grounding/localization capability which is critical for many downstream tasks such as visual reasoning. In this work, we propose a novel Position-guided Text Prompt (PTP) paradigm to enhance the visual grounding ability of cross-modal models trained with VLP. Specifically, in the VLP phase, PTP divides the image into NxN blocks, and identifies the objects in each block through the widely used object detector in VLP. It then reformulates the visual grounding task into a fill-in-the-blank problem given a PTP by encouraging the model to predict the objects in the given blocks or regress the blocks of a given object, e.g. filling “P” or “O” in a PTP “The block P has a O”. This mechanism improves the visual grounding capability of VLP models and thus helps them better handle various downstream tasks. By introducing PTP into several state-of-the-art VLP frameworks, we observe consistently significant improvements across representative cross-modal learning model architectures and several benchmarks, e.g. zero-shot Flickr30K Retrieval (+4.8 in average recall@1) for ViLT baseline, and COCO Captioning (+5.3 in CIDEr) for SOTA BLIP baseline. Moreover, PTP achieves comparable results with object-detector based methods, and much faster inference speed since PTP discards its object detector for inference while the later cannot. Our code and pre-trained weight will be released.

Chat is not available.