Scene Text recognition with Full Normalization
This work addresses data scarcity for researchers in scene text recognition, but it is incremental as it applies existing methods to new data.
The authors tackled the problem of limited training data for scene text recognition by creating a new dataset of real smartphone shots and demonstrated the effectiveness of profile normalization, achieving unspecified performance improvements.
Scene text recognition has made significant progress in recent years and has become an important part of the work-flow. The widespread use of mobile devices opens up wide possibilities for using OCR technologies in everyday life. However, lack of training data for new research in this area remains relevant. In this article, we present a new dataset consisting of real shots on smartphones and demonstrate the effectiveness of profile normalization in this task. In addition, the influence of various augmentations during the training of models for analyzing document images on smartphones is studied in detail. Our dataset is publicly available.