딥러닝 기반 영역 정보를 활용한 피부 병변 분류 방법

Abstract

Skin lesions result from abnormal proliferation of cells. Skin lesions have high rate of misdiagnosis due to a wide variety of forms. Recently, a deep learning-based skin lesion classification methods have been proposed but they tend to misclassify lesions with unclear skin lesion boundaries. In this paper, we propose a method for classifying skin lesions using deep neural network with area information. As a result, we show that the performance of our method is improved by 1.7~2.43 in terms of F1-score, compared to the previous methods.