Дослідження оптимізаційних властивостей порогової техніки обробки зображень
Aim of work is study of possibility of segmentation methodic optimization, that is basic on threshold idea. Authors made mathematical forms and compare about their effectiveness on different areas. Global threshold, semi threshold, multilevel thresholding, variable threshold, threshold selection by maximizing between-class variance, threshold selection using a simple image statistic methods will be studied in the conference. Global threshold is effective on clear area; semi threshold can be used as additional filter; multilevel thresholding is good for situation with different types of objects; variable threshold effective to images with local bimodal diagram; threshold selection by maximizing between-class variance it is method without supervisor; simple statistic method is good for images with low resolution.
3. Krcah M, Szekely G, and Blanc R. Fully automatic and fast segmentation of the femur bone from 3D-CT images with no shape prior. In: Proceedings of IEEE International Symposium on Biomedical Imaging: From Nano to Macro. Chicago; 2011. p. 2087-90. DOI: 10.1109/ISBI.2011.5872823
4. Binary image segmentation using the level set method [Internet]. Habr.com. 2018 [cited Dec 2018]. Available from: https://habr.com/post/332692/ Python + OpenCV + Keras . https://habr.com/ru/post/466565/