VCR: A Task for Pixel-Level Complex Reasoning in Vision Language Models via Restoring Occluded Text

TL;DR

This paper introduces VCR, a new vision-language task requiring models to restore obscured text in images using pixel-level cues, highlighting the challenge of integrating visual and textual information beyond traditional OCR methods.

Contribution

The paper presents a novel pixel-level reasoning task, a synthetic dataset VCR-Wiki with 2.11M images, and demonstrates the limitations of current models in this complex text restoration challenge.

Findings

Current models lag behind human performance in VCR.

Fine-tuning on the dataset does not significantly improve results.

The dataset enables future research in pixel-level reasoning for vision-language models.

Abstract

We introduce Visual Caption Restoration (VCR), a novel vision-language task that challenges models to accurately restore partially obscured texts using pixel-level hints within images. This task stems from the observation that text embedded in images is intrinsically different from common visual elements and natural language due to the need to align the modalities of vision, text, and text embedded in images. While numerous works have integrated text embedded in images into visual question-answering tasks, approaches to these tasks generally rely on optical character recognition or masked language modeling, thus reducing the task to mainly text-based processing. However, text-based processing becomes ineffective in VCR as accurate text restoration depends on the combined information from provided images, context, and subtle cues from the tiny exposed areas of masked texts. We develop a pipeline to generate synthetic images for the VCR task using image-caption pairs, with adjustable caption visibility to control the task difficulty. With this pipeline, we construct a dataset for VCR called VCR-Wiki using images with captions from Wikipedia, comprising 2.11M English and 346K Chinese entities in both easy and hard split variants. Our results reveal that current vision language models significantly lag behind human performance in the VCR task, and merely fine-tuning the models on our dataset does not lead to notable improvements. We release VCR-Wiki and the data construction code to facilitate future research.