 |
General: 663 |
P8: Classiffication of Hand-Written Digits Using Scattering Convolutional Network
Author: Dongmian Zou , Advisor: Radu Balan (Math)
Problem Statement Presentation
Project Proposal
Abstract
In image classiffication it is advantageous to build classiffiers based on feature extractors. It is demanded that the feature extractors unmask the identity of the image and remain stable with regard to small-scale deformations. A scattering convolutional network is such a feature extractor. In this project, we propose to use modern machine learning techniques to train the scattering convolutional network and test them on a standard database (MNIST) for the task of recognizing handwritten digits.
MidYear Progress Report and
Presentation
Final Presentation ,
Final Report
