COMPARATIVE ANALYSIS OF DATA BALANCING METHODS FOR THE CLASSIFICATION OF RAMAN SPECTRA OF SKIN GROWTHS BASED ON A CONVOLUTIONAL NEURAL NETWORK
A.O. Kapkaeva1, I.A. Matveeva1;
1Samara National Research University, Samara
Abstract
This work is devoted to the study of data balancing methods for the classification of Raman spectra of skin neoplasms. Two balancing methods, were compared in combination with the ResNet architecture for classifying 4 types of skin growths. The process of preprocessing spectral data and the influence of various balancing methods on learning are described. The results of the model's work on medical data and numerical indicators of classification quality are presented.
Speaker
Kapkaeva Angelina
Samara National Research University
Russia
Discussion
Ask question
