JaSPICE: Automatic Evaluation Metric Using Predicate-Argument Structures for Image Captioning Models

November 07, 2023 · Entered Twilight · 🏛 Conference on Computational Natural Language Learning

"No code URL or promise found in abstract"
"Code repo scraped from project page (backfill)"

Evidence collected by the PWNC Scanner

Repo contents: .gitignore, README.md, bleu, cider, eval.py, example, license.txt, meteor, rouge, setup.py, spice, tokenizer

Authors Yuiga Wada, Kanta Kaneda, Komei Sugiura arXiv ID 2311.04192 Category cs.CV: Computer Vision Cross-listed cs.CL Citations 6 Venue Conference on Computational Natural Language Learning Repository https://github.com/salaniz/pycocoevalcap ⭐ 336 Last Checked 2 months ago

Abstract

Image captioning studies heavily rely on automatic evaluation metrics such as BLEU and METEOR. However, such n-gram-based metrics have been shown to correlate poorly with human evaluation, leading to the proposal of alternative metrics such as SPICE for English; however, no equivalent metrics have been established for other languages. Therefore, in this study, we propose an automatic evaluation metric called JaSPICE, which evaluates Japanese captions based on scene graphs. The proposed method generates a scene graph from dependencies and the predicate-argument structure, and extends the graph using synonyms. We conducted experiments employing 10 image captioning models trained on STAIR Captions and PFN-PIC and constructed the Shichimi dataset, which contains 103,170 human evaluations. The results showed that our metric outperformed the baseline metrics for the correlation coefficient with the human evaluation.