Image Textualization: An Automatic Framework for Creating Accurate and Detailed Image Descriptions

TL;DR

This paper introduces Image Textualization, an automatic framework that leverages multi-modal large language models and vision experts to generate detailed, high-quality image descriptions, addressing the limitations of existing datasets.

Contribution

The paper proposes a novel collaborative framework for automatic image description generation and introduces benchmarks for evaluating detailed descriptions.

Findings

IT framework produces richer, more detailed descriptions

LLaVA-7B trained on IT descriptions shows improved detail and reduced hallucination

Benchmark results validate the quality of generated descriptions

Abstract

Image description datasets play a crucial role in the advancement of various applications such as image understanding, text-to-image generation, and text-image retrieval. Currently, image description datasets primarily originate from two sources. One source is the scraping of image-text pairs from the web. Despite their abundance, these descriptions are often of low quality and noisy. Another is through human labeling. Datasets such as COCO are generally very short and lack details. Although detailed image descriptions can be annotated by humans, the high annotation cost limits the feasibility. These limitations underscore the need for more efficient and scalable methods to generate accurate and detailed image descriptions. In this paper, we propose an innovative framework termed Image Textualization (IT), which automatically produces high-quality image descriptions by leveraging existing multi-modal large language models (MLLMs) and multiple vision expert models in a collaborative manner, which maximally convert the visual information into text. To address the current lack of benchmarks for detailed descriptions, we propose several benchmarks for comprehensive evaluation, which verifies the quality of image descriptions created by our framework. Furthermore, we show that LLaVA-7B, benefiting from training on IT-curated descriptions, acquire improved capability to generate richer image descriptions, substantially increasing the length and detail of their output with less hallucination.