36th Annual Student Research Conference
Generating Neural Network Models to Detect Tumors
Justin A. Caringal
Dr. Ruthie Halma, Faculty Mentor
The accurate identification of tumors is an essential problem in medicine and provides a good foundation in solving a wide range of problems involving artificial intelligence and computer vision. The expanding areas of machine learning and image processing provide a relevant problem domain for creating a functional neural network model with quality-of-life tools, using budget-friendly open-source packages and other publicly available resources. This paper discusses the effectiveness of tumor detection model generation along with the creation of a model with 100 percent validation accuracy at a 0.416 percent validation loss. While the model is not ready for the detection of tumors in a personal or professional setting, the project provides promising results when combining neural networks to detect tumors. This paper comes as the culmination of the research conducted under a TSU Computer Science department grant, but by no means is the model a final product; potential improvements are presented.
Keywords: Artificial Intelligence, Neural Networks, Convolution, Datasets, Deep Q-Learning, Computer Vision, Brain Tumors
Topic(s):Computer Science
Presentation Type: Oral Presentation
Session: 209-2
Location: MG 1098
Time: 10:30