# Machine Learning Shapes

This repository contains code and resources for a machine learning project focused on classifying shapes using convolutional neural networks (CNNs).

Overview

The project aims to demonstrate the application of CNNs in classifying simple geometric shapes such as circles and triangles. It utilizes the Keras deep learning library with TensorFlow backend for building and training the neural network models.

Key Features

• Dataset generation: Script to generate synthetic images of shapes with variations in size, position, and rotation.
• Model architecture: Implementation of CNN models using Keras for shape classification.
• Training scripts: Python scripts for training the models on the generated dataset.
• Evaluation: Evaluation of trained models on test data and visualization of model performance.
• Grad-CAM visualization: Utilization of Grad-CAM (Gradient-weighted Class Activation Mapping) for visualizing CNN model predictions.

Repository Structure

• data/: Contains scripts for generating synthetic shape images and dataset.
• models/: Includes scripts defining CNN model architectures.
• train/: Scripts for training the models on the generated dataset.
• evaluation/: Evaluation scripts for assessing model performance.
• visualization/: Scripts for visualizing model predictions and Grad-CAM heatmaps.
• utils/: Utility functions used across different parts of the project.
• requirements.txt: Python dependencies required for running the project.

Data_Sets

https://drive.google.com/drive/folders/1U_lz3Pn5p6mRxWoqeCcJD4OGTdL5aopc

Getting Started

1. Clone this repository: git clone https://github.com/DominikRoczan/MachineLearning-Shapes.git

Contributor

• Dominik Roczan