Sep, 2019

ICDAR 2019 大规模街景文本部分标注竞赛 --RRC-LSVT

TL;DRICDAR 2019-LSVT competition aims to improve the performance of text reading from street view images through exploring state-of-the-art methods on a Large-scale Street View Text with Partial Labeling (LSVT) challenge, which provides 50,000 and 400,000 fully and weakly annotated images, respectively, leading to 41 teams participating in the two proposed tasks of text detection and end-to-end text spotting.