ajay-goel / neural-networks

Neural-Networks

INFO_6205_Program_Structures_-_Algorithm

Project - Digit and Face Recognition Neural Network

Team Member - Ajay Goel, Akshay N. Mahajanshetti

Introduction

Artificial neural networks (ANNs)

or connectionist systems are computing systems which perform tasks by considering examples, generally without being programmed with any task-specific rules. For example, In image recognition, system might learn to identify images that contain cats by analyzing training images that have been manually labelled as "cat" or "no cat" and using the results to identify cats in other images. It does this without any prior knowledge about cats. The system learns things like cats have fur, tails, whiskers and other cat attributes from the training data.

An ANN is based on a collection of connected units or nodes which can transmit a signal from one artificial neuron to another. An artificial neuron that receives a signal can process it and then signal additional artificial neurons connected to it. Artificial neurons and edges typically have a weight that adjusts as learning proceeds. The weight increases or decreases the strength of the signal at a connection. Different layers may perform different kinds of transformations on their inputs. Signals travel from the first layer (the input layer), to the last layer (the output layer), possibly after traversing the layers multiple times.

Problem Statement

The neural network in our application is used to solve two major problems:

Digit Recognition

In this problem, The idea is to take a large number of handwritten digits (in excel format). In our system, we have taken data set of 42000 digits to train the system then develop a system which can learn from those training examples, we test it by passing 11000 testing digits to the system and at least 85% digits should be predicted correctly. Furthermore, by increasing the number of training examples, the network can learn more about handwriting, and so improve its accuracy.

Face Recognition

In this problem, our aim is to pass images (.png format) of 10 different people in 5 different angles and train the system. After this, to see if the system has learnt to identify different people, we test it by passing close to 35 images to the system and at least 80% images should be recognised correctly.

Impementation

Project source code in the project/src folder.

A detailed report of this project and including basics about Neural Networks in file named: Project_Report.pdf

A short presentation to get the understating of working of the project with code and output snippets in file named: Project Presentation.pptx