Abstract
With the continuous development of OCR technology, the text detection in a wide range of scenarios has gradually become a research hotspot and has very good application value. Compared with OCR applications such as license plate recognition, text recognition in natural scenes with more complex and diverse environmental backgrounds is more difficult. In particular, the presence of a background that is very close to the font in the image background often leads to detection errors. This paper proposes a relational network text detection architecture based on the attention mechanism, which is used to select and weight different granular nodes in the relational network, and also used for the nodes in the fusion of link to further reduce the computational cost. The experimental results show that the proposed framework in this paper can further predict and learn the connection relationships in the relational graph network, and has good accuracy in some indicators.
Export citation and abstract BibTeX RIS
Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.