Total-Text: A Comprehensive Dataset for Scene Text Detection and Recognition

Text in curve orientation, despite being one of the common text orientationsin real world environment, has close to zero existence in well received scenetext datasets such as ICDAR2013 and MSRA-TD500. The main motivation ofTotal-Text is to fill this gap and facilitate a new research direction for thescene text community. On top of the conventional horizontal and multi-orientedtexts, it features curved-oriented text. Total-Text is highly diversified inorientations, more than half of its images have a combination of more than twoorientations. Recently, a new breed of solutions that casted text detection asa segmentation problem has demonstrated their effectiveness againstmulti-oriented text. In order to evaluate its robustness against curved text,we fine-tuned DeconvNet and benchmark it on Total-Text. Total-Text with itsannotation is available at https://github.com/cs-chan/Total-Text-Dataset