Method: OpenCV + Tesseract - Sample num 2 - Task 5 - Still image mode End-to-End - ICDAR 2017 Robust Reading Challenge on Omnidirectional Video - Robust Reading Competition

Sample 2 of 2893

Evaluation:

Sample Ranking	Recall	Precision	Hmean
google vision api	15.38%	100.00%	26.67%
OpenCV + Tesseract	7.69%	50.00%	13.33%