Improvement of OCT images classification accuracy using nucleus information

Ting-Yu Chang^1*, Snow H. Tseng¹

¹Graduate Institute of Photonics and Optoelectronics, National Taiwan University, Taipei, Taiwan

* Presenter:Ting-Yu Chang, email:r07941050@ntu.edu.tw

The abnormal enlargement of nucleus is a crucial indicator of skin cancer. In this research, we aim to improve the performance of optical coherence tomography (OCT) classification model using deep learning. Specifically, we extract information about nucleus size from skin OCT images as our training data set. Factors result in the variation of performance are also discussed. Our research may provide a practical and reasonable method for future development of skin cancer diagnosis using machine learning.

Keywords: Deep Learning, Full-Field Optical Coherence Tomography, Convolutional Neural Network, Receptive Fields