An efficient approach in Iris genetic lesion detection algorithm using SURF and SVM

In this paper, we propose a filtering algorithm that can extract lesions by using lesion feature points based on medical images. We use representative feature extraction methods for lesion recognition based on extraction of shape features and feature points from lesion patterns. I suggest. The preprocessing process of medical image processing removes unnecessary noise from complex medical images and performs the binary-coded process by effectively separating lesions from them. The feature points are extracted from the lesion objects extracted with various shapes and sizes and the image outline is detected after normalization. It has an advantage that it can be extended to global features by comparing similarities between different vectors existing among various medical images. In addition, compared with the existing methods, the proposed method improves the error rate of the lesion matching by 1.02% on average and the processing speed of the medical image of 512 × 512 in the experiment by about 40%.