Benchmarking and Improving Detail Image Caption

TL;DR

This paper introduces a new benchmark and a reliable evaluation metric for detailed image captioning, and demonstrates how to enhance LVLMs' captioning abilities through a self-supervised data synthesis pipeline.

Contribution

It proposes a high-quality dataset and a novel CAPTURE metric for detailed image captioning evaluation, and develops a self-looping data construction pipeline to improve LVLM performance.

Findings

CAPTURE achieves higher consistency with expert judgments than existing metrics.

The data synthesis pipeline significantly enhances LVLMs' captioning quality.

Model performance improves with iterative self-looping data refinement.

Abstract

Image captioning has long been regarded as a fundamental task in visual understanding. Recently, however, few large vision-language model (LVLM) research discusses model's image captioning performance because of the outdated short-caption benchmarks and unreliable evaluation metrics. In this work, we propose to benchmark detail image caption task by curating high-quality evaluation datasets annotated by human experts, GPT-4V and Gemini-1.5-Pro. We also design a more reliable caption evaluation metric called CAPTURE (CAPtion evaluation by exTracting and coUpling coRE information). CAPTURE extracts visual elements, e.g., objects, attributes and relations from captions, and then matches these elements through three stages, achieving the highest consistency with expert judgements over other rule-based or model-based caption metrics. The proposed benchmark and metric provide reliable evaluation for LVLM's detailed image captioning ability. Guided by this evaluation, we further explore to unleash LVLM's detail caption capabilities by synthesizing high-quality data through a five-stage data construction pipeline. Our pipeline only uses a given LVLM itself and other open-source tools, without any human or GPT-4V annotation in the loop. Experiments show that the proposed data construction strategy significantly improves model-generated detail caption data quality for LVLMs with leading performance, and the data quality can be further improved in a self-looping paradigm. All code and dataset will be publicly available at https://github.com/foundation-multimodal-models/CAPTURE.