A Simple CNN Tutorial with App.

Benan AKCA

1. INTRODUCTION

2. CONVOLUTIONAL NEURAL NETWORKS

   1. Convolution Layer
      • Steps of Convolution Operation
      • Stride
      • An Example of Convolution Operation
      • Why Do We Use Filters?
   2. Padding
      • Why do we use Padding?
      • Equation of Calculating Output Dimension
      • Types of Padding in Keras
   3. Pooling Layer
      • Hyper Parameters of Pooling Operation
      • Why do we use Pooling?
      • Types of Pooling in Keras
   4. Batch Normalization Layer
   5. Drop Out Layer

3. APPLICATION WITH CNN

   1. Import Modules
   2. Understanding the Data
   3. Data Preprocessing
      • Normalizing the Data
      • Train and Test Splitting
      • Reshape Data to Appropriate Sizes
   4. Building and Training a CNN Model
      • Model - Build
      • Model - Compile
        • Categorical Cross Entropy
        • Optimizer
        • On-the-Fly Data Augmentation
      • Model - Fit
   5. Evaluation of the Model
      • Accuracy and Loss Curves *
      • Test Set Accuracy Score
      • Confusion Matrix
      • F1 Score Calculation
      • Evaluate with Another Dataset
      • Submit for Competition

4. CONCLUSION