Hailemicael/Amharic-Character-Recognition-with-ML

This project focuses on recognizing handwritten Amharic characters using machine learning techniques. It uses PCA and LDA for dimensionality reduction and classifiers like SVM, Logistic Regression, and KNN for character recognition. The goal is to build an accurate model that can classify Amharic characters. # Amharic Character Recognition # Table of Contents 1. Introduction 2. Motivation 3. Dataset 4. Methodology - Data Loading and Preprocessing - Train and Test Data Splitting - Dimensionality Reduction - PCA - LDA - Model Training Evaluation and Analysis 5. Results 6. Usage 7. Contributors ## Introduction This project aims to recognize Amharic characters using machine learning techniques. Its primary objective is to develop a model capable of accurately classifying Amharic characters from images. The project employs Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) for dimensionality reduction. Additionally, it utilizes various classifiers such as Support Vector Machines (SVM), Logistic Regression, and K-nearest neighbors (KNN) for classification. The performance of each classifier is evaluated before and after applying feature extraction techniques. ## Motivation The motivation behind this project is to bridge the gap in existing character recognition systems that primarily focus on Latin characters. Amharic is one of the most widely spoken languages in Ethiopia, and having accurate character recognition systems can facilitate tasks such as text processing, language learning, and document digitization for Amharic speakers. ## Dataset The dataset consists of images of handwritten Amharic characters. Each image is labeled with the corresponding character it represents. The dataset has been preprocessed to ensure consistency in image size and format, making it suitable for training machine learning models. It comprises 4200 characters, including some augmented images, with a total of 14 distinct characters for classification. ## Methodology ### Data Loading and Preprocessing The images are loaded and preprocessed to convert them into a format suitable for training machine learning models. This includes resizing, normalization, and flattening of the image data, as well as shuffling. ```python # Function to load images and labels def load_ima …