Isolated-Speech-Recognition-using-LCNN

The aim of the project is to classify and evaluate spoken digits (0-9) by the help of Signal Processing for feature extraction and Machine Learning for Classification. By using Lightweight Convolutional Neural Network (LCNN), Rectifier Linear unit (ReLU) activation function has not been used in favor of using Max-Feature-Map (MFM) as it does faster classification and better results. Managed to have the accuracy between 70% and 95%. The programming language that has been used is Python.