REO-Relevance, Extraness, Omission: A Fine-grained Evaluation for Image Captioning
This addresses the need for more informative evaluation in image captioning, though it is incremental as it builds on existing metrics.
The paper tackles the problem of coarse evaluation in image captioning by proposing REO, a fine-grained metric that assesses captions based on relevance, extraness, and omission, achieving higher consistency with human judgments on three benchmark datasets.
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.