Visual Semantic Re-ranker for Text Spotting
This work improves text-spotting accuracy for applications like document analysis or scene understanding, but it is incremental as it builds on existing methods with a complementary approach.
The paper tackles the problem of text recognition by addressing the lack of semantic context in existing methods, proposing a post-processing re-ranker that uses visual-textual relations to improve accuracy, resulting in boosted performance with low computational cost on the ICDAR'17 dataset.
Many current state-of-the-art methods for text recognition are based on purely local information and ignore the semantic correlation between text and its surrounding visual context. In this paper, we propose a post-processing approach to improve the accuracy of text spotting by using the semantic relation between the text and the scene. We initially rely on an off-the-shelf deep neural network that provides a series of text hypotheses for each input image. These text hypotheses are then re-ranked using the semantic relatedness with the object in the image. As a result of this combination, the performance of the original network is boosted with a very low computational cost. The proposed framework can be used as a drop-in complement for any text-spotting algorithm that outputs a ranking of word hypotheses. We validate our approach on ICDAR'17 shared task dataset.