ConvolutionalNeuralNetworks Part-1

In this series of deep learning sessions on CNN, we will go through variety of applications and also understand the basic concepts behind CNN and implement them.

Welcome to the House of CNN: In this series of sessions, we will cover variety of things such as,

1. Step by Step understanding and implementing of CNN layers.
2. Hand gesture classification using CNN on tensorflow framework.
3. The Deep learning Happy House Application
4. Face verification and Face recognition system using CNN.
5. Self-driving car application (Ex. Car detection)

Part 1: Step by Step understanding and implementing of CNN layers.

In this session, we will go through step by step analysis of CNN layers, we will implement CONVOLUTION and POOLING layers using numpy library, we will implement forward propagation and backward propagation for single layer and expand it to multiple layers.

Step by Step Implementation of CNN Layers:

• Install all the latest dependencies.
• Clone the repository in your local system.
• Make sure all folders are in same location
• Open any python3.6 IDE and execute StepByStepCNN.py
• Visualize the results of different layers

Dependencies:

1. Python 3.6
2. Numpy
3. Matplotlib
4. h5py

The Repository also has notebook(.ipynb) file, which has detailed explnation of project and also implmentation steps, do check it out once

Visualization of max_pooling layer:

Thank You, cheers. 👍