Cat vs Dog Image Classification

Welcome to the "Cat vs Dog Image Classification, making prediction" project! This project focuses on building a powerful image classification model to distinguish between cats and dogs using a Convolutional Neural Network (CNN).

Project Overview

• Reference: Building Powerful Image Classification Models using Very Little Data
• Dataset: Cat vs Dog Image Classification Dataset
• Model Accuracy: 89%

Project Structure

The project is structured as follows:

• Data Preprocessing: Utilized the Keras ImageDataGenerator for augmenting and preprocessing the training and test sets.
• CNN Architecture: Implemented a Convolutional Neural Network with multiple layers for feature extraction and classification.
• Training and Evaluation: Compiled and trained the CNN using the training set and evaluated on the test set.
• Single Prediction: Demonstrated making a single prediction using a test image.

Code Steps

1. Data Preprocessing: Create training and test sets using ImageDataGenerator.
2. CNN Architecture: Build a CNN with convolutional, pooling, flattening, and fully connected layers.
3. Training: Compile and train the model using the training set.
4. Evaluation: Evaluate the model on the test set to measure accuracy.
5. Prediction: Make a single prediction using a sample image.

Model Accuracy

The model achieved an Training Set:

Accuracy: 89.92%, Loss: 0.2320, Validation Set (Test Set): Accuracy: 80.70%, Loss: 0.5082, showcasing its effectiveness in classifying cats and dogs.

CNN Concept used

Kaggle Code

Explore the complete code on Kaggle: Cat vs Dog Image Classification Kaggle Notebook