SUPER-RESOLUTION METHOD FOR DOCUMENT IMAGES FROM MOBILE DEVICES

Zagaynov I.G

230 KB

It is common situation that resolution of the original document image is not sufficient for optical character recognition (OCR). This may be the result of limited resolution of matrix in the mobile device camera, the losses in optical system, the image encoding algorithms, and other factors. However, there is an option to take a number of images. This paper considers the problem of obtaining higher quality and higher resolution for document images. Several images or video sequence should be available for the purpose. The aim of improvements is better optical character recognition of the original document. The paper gives the assumptions of the class of document images and proposes motion models to compensate inter-frame shifts. The algorithm for signal accumulation from sequence of images is described. The experimental results confirm that the proposed method improves optical character recognition.

Библиографическая ссылка

Загайнов И.Г МЕТОД ПОВЫШЕНИЯ РАЗРЕШЕНИЯ ДЛЯ ИЗОБРАЖЕНИЙ ДОКУМЕНТОВ С МОБИЛЬНЫХ УСТРОЙСТВ // Научное обозрение. Физико-математические науки . 2020. № 1. С. 30-30;
URL: https://physics-mathematics.ru/en/article/view?id=31 (дата обращения: 24.06.2026).