By Pushkar Jain
It is a binary classifier built using an artificial neural network making it from scratch in Python. It's is Machine Learning project for classifying image data in two different classes
Binary Image Classifier
It is a binary classifier built using an artificial neural network. This is being used to classify images of a dataset into two classes which are “cat” or “not cat”. The neural network is made from scratch with atomic units as neuron which contains the activation functions. The classifier is made in Python language with efficient use of concepts of object-oriented programming.
In the field of information extraction and retrieval, binary classification is the process of classifying a given document/account on the basis of predefined classes.
Cat detection is based on binary classification, in which two classes are predefined i.e. Cat or non-Cat. And datasets are classified on the basis of predefined classes.
The task of classifying images into binary classification has gotten a lot easier in the past decade with high-end libraries like Tensorflow, Keras and PyTorch the task is virtually already in place.
However, this project aims to deep dive into the actual workings of classifiers so as to analyse and discover how layers in a classification network work?
How do activation functions help render conclusive results? How biases affect the basic unit of a neural network i.e. the neuron and above all how this is implemented in a viable way so as to make decisions on classifying images.
Submitted by Pushkar Jain (pushkarjain1009)
Download packets of source code on Coders Packet