Reading Text in the Wild with Convolutional Neural Networks

Benchmark Model Rank Results
scene-text-detection-on-icdar-2013Jaderberg et al.F-Measure: 76.8%Precision: 88.5Recall: 67.8