- Notifications
You must be signed in to change notification settings - Fork0
[CoNLL23] JaSPICE: Automatic Evaluation Metric Using Predicate-Argument Structures for Image Captioning Models
License
keio-smilab23/JaSPICE
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
[CoNLL23] JaSPICE: Automatic Evaluation Metric Using Predicate-Argument Structures for Image Captioning Models
- Accepted at CoNLL23
- arXiv paper
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.
pip install jaspiceor
git clone git@github.com:keio-smilab23/JaSPICE.gitcd JaSPICEpip install -e .docker build -t jaspice .Run the docker container.
docker run -d -p 2115:2115 jaspicefromjaspice.apiimportJaSPICEbatch_size=16jaspice=JaSPICE(batch_size,server_mode=True)_,score=jaspice.compute_score(references,candidates)
git clone git@github.com:keio-smilab23/JaSPICE.gitcd JaSPICEpip install -e .- juman : v7.01
- juman++ : v1.02
- knp : v4.20
# JUMAN++wget 'https://github.com/keio-smilab23/JaSPICE/releases/download/0.0.1/jumanpp.tar.xz'# JUMANwget 'https://github.com/keio-smilab23/JaSPICE/releases/download/0.0.1/juman.tar.bz2'# KNPwget 'https://github.com/keio-smilab23/JaSPICE/releases/download/0.0.1/knp.tar.bz2'fromjaspice.apiimportJaSPICEbatch_size=16jaspice=JaSPICE(batch_size,server_mode=False)_,score=jaspice.compute_score(references,candidates)
- 「人通りの少なくなった道路で青いズボンを着た男の子がオレンジ色のヘルメットを被りスケートボードに乗っている．」
@inproceedings{wada2023, title = {{JaSPICE: Automatic Evaluation Metric Using Predicate-Argument Structures for Image Captioning Models}}, author = {Wada, Yuiga and Kaneda, Kanta and Sugiura, Komei}, year = 2023, booktitle = {Proceedings of the 27th Conference on Computational Natural Language Learning (CoNLL)}}This work is licensed under the BSD-3-Clause-Clear license. To view a copy of this license, seeLICENSE.
About
[CoNLL23] JaSPICE: Automatic Evaluation Metric Using Predicate-Argument Structures for Image Captioning Models
Resources
License
Uh oh!
There was an error while loading.Please reload this page.
Stars
Watchers
Forks
Packages0
Uh oh!
There was an error while loading.Please reload this page.