Evaluation codes for MS COCO caption generation.
- java 1.8.0
- python 2.7
./
- cocoEvalCapDemo.py (demo script)
./annotation
- captions_val2014.json (COCO 2014 caption validation set)
- More detials can be found under download tab on COCO dataset
./results
- captions_val2014_fakecap_results.json (example fake results for running demo)
- More detials can be found under evaluate->format tab on COCO dataset
./pycocoevalcap: The folder where all evaluation codes are stored.
- evals.py: includes COCOEavlCap class that includes multiple scores.
- tokenizer: PTBTokenizer
- bleu: Bleu evalutation codes
- meteor: Meteor evaluation codes
- rouge: Rouge-L evaluation codes
- cider: CIDEr evaluation codes
- BLEU: BLEU: a Method for Automatic Evaluation of Machine Translation
- Meteor: Project page with related publications. We used the latest version (1.5) of the Code. Changes have been made to the source code to properly aggreate the statistics for the entire corpus.
- Rouge-L: ROUGE: A Package for Automatic Evaluation of Summaries
- CIDEr: [CIDEr: Consensus-based Image Description Evaluation] (http://arxiv.org/pdf/1411.5726.pdf)