Do Current Video LLMs Have Strong OCR Abilities? A Preliminary Study

TL;DR

This paper introduces a new benchmark to evaluate the OCR capabilities of multimodal large language models on videos, highlighting current limitations and guiding future improvements.

Contribution

It presents a comprehensive benchmark with 1,028 videos and 2,961 QA pairs to assess video OCR performance across multiple challenging subtasks.

Findings

Current video LLMs show limited OCR accuracy

The benchmark reveals gaps in semantic and spatial understanding of OCR objects

The resource facilitates targeted improvements in video OCR abilities

Abstract

With the rise of multimodal large language models, accurately extracting and understanding textual information from video content, referred to as video based optical character recognition (Video OCR), has become a crucial capability. This paper introduces a novel benchmark designed to evaluate the video OCR performance of multi-modal models in videos. Comprising 1,028 videos and 2,961 question-answer pairs, this benchmark proposes several key challenges through 6 distinct subtasks: (1) Recognition of text content itself and its basic visual attributes, (2)Semantic and Spatial Comprehension of OCR objects in videos (3) Dynamic Motion detection and Temporal Localization. We developed this benchmark using a semi-automated approach that integrates the OCR ability of image LLMs with manual refinement, balancing efficiency, cost, and data quality. Our resource aims to help advance research in video LLMs and underscores the need for improving OCR ability for video LLMs. The benchmark will be released on https://github.com/YuHuiGao/FG-Bench.git.