Optimization of Deep Architecture for Classification of Skin Disease (2016)
Undergraduate: Conrad Czejdo
Faculty Advisor: Sambit Battacharya
Department: Chemistry
Deep learning is a fast growing field in machine learning that attempts to classify data through multi-layered transformations that resemble the human brain. It has become an essential tool for a variety of uses including voice translation and handwriting recognition. Rapid development in the area has resulted in a plethora of different architectures, like traditional core layers, convolution layers and recurrent layers. By utilizing Keras, fast testing of each method can be accomplished on a sample of images of various skin diseases in order to find an optimal combination of layers.