PopEval: A Character-Level Approach to End-to-End Evaluation Compatible with Word-Level Benchmark Dataset.
Hong-Seok LeeYoungmin YoonPil-Hoon JangChankyu ChoiPublished in: ICDAR (2019)
Keyphrases
- end to end
- benchmark datasets
- word level
- language independent
- text localization and recognition
- document images
- machine translation
- document analysis
- character recognition
- sentence level
- n gram
- document level
- congestion control
- word segmentation
- computer vision
- language model
- language modeling
- semantic roles
- viterbi algorithm
- co occurrence
- machine learning