METHOD FOR BLUR AND DEFOCUS DETECTION BASED ON TEXT MODEL

Kostyukov M.V

238 KB

The paper considers the problem of automatic blur and defocusing detecting on images of documents. The problem is solved based on text model and analysis of second derivative of brightness function along the brightness gradient. The paper gives text model description, examples of text distorted images and some advises on choosing descriptors for blur and defocusing detection and further distortion classification. The way for calculation of invariant number features by local nonuniformities of image from selected descriptors is proposed. The paper describes specificity for the classifier training. Particular attention is paid to the problem for performance of involved algorithms. Principal research is provided on difference of Gaussians image to mimic neural processing of the retina in human eye.

Библиографическая ссылка

Костюков М.В ДЕТЕКТОР СМАЗАННОСТИ И РАСФОКУСИРОВКИ НА ОСНОВЕ МОДЕЛИ ТЕКСТА // Научное обозрение. Физико-математические науки . 2020. № 1. С. 34-35;
URL: https://physics-mathematics.ru/en/article/view?id=39 (дата обращения: 24.06.2026).