REO-Relevance, Extraness, Omission: A Fine-grained Evaluation for Image Captioning

TL;DR

This paper introduces REO, a fine-grained evaluation method for image captioning that assesses relevance, extraness, and omission, providing more detailed insights than traditional single-score metrics.

Contribution

The study proposes a novel evaluation approach that offers detailed analysis of caption quality from multiple perspectives, improving correlation with human judgments.

Findings

REO correlates better with human judgments than traditional metrics.

REO provides more intuitive and detailed evaluation results.

Experiments on benchmark datasets validate the effectiveness of REO.

Abstract

Popular metrics used for evaluating image captioning systems, such as BLEU and CIDEr, provide a single score to gauge the system's overall effectiveness. This score is often not informative enough to indicate what specific errors are made by a given system. In this study, we present a fine-grained evaluation method REO for automatically measuring the performance of image captioning systems. REO assesses the quality of captions from three perspectives: 1) Relevance to the ground truth, 2) Extraness of the content that is irrelevant to the ground truth, and 3) Omission of the elements in the images and human references. Experiments on three benchmark datasets demonstrate that our method achieves a higher consistency with human judgments and provides more intuitive evaluation results than alternative metrics.