다중 해상도를 이용한 딥러닝 기반 창상 분류 방법

Abstract

In medical data, the distance between the wound and the imaging device is not constant, so there is a problem that the deep learning model becomes sensitive to changes in resolution. To solve this problem, in this paper, we propose a network that is robust to resolution changes through multiple resolution input. Experimentally, the proposed method is more robust to resolution changes than the existing method, and shows 1.48% higher performance.