Lung Nodules Detection and Classification by 3D Convolutional Neural Networks


Jingjing Yang, Huayuan Data Technology (Shanghai) Co., Ltd.


2019.09.18 12:20-13:05


Room 306, No.5 Science Building


Deep learning algorithms, in particular, deep convolutional neural networks(CNN), have recently received much attention and have achieved very good results in medical image analysis. In this report, we will share our experience with you in building 3D CNN models to improve the detection accuracy of lung nodules. Compared with other 2D CNN based models, our models can encode richer spatial information and learn more representative features of 3D images. Then we will introduce our best practice on the pipeline of lung CT images preprocess and transformation which is very important in real-world deep learning applications. Moreover, we will show you how our AI technology is helping doctors save lives in real-world cases. Finally, we will discuss our next step in developing more power AI models for the healthcare industry.