عکس msnmkh
Recognize English spoken digits using Hidden Markov Model Python
موضوع‌ها
۰
فورک‌ها
۶
ستاره‌ها
۴
تاریخ ایجاد
۲۴ خرداد ۱۳۹۸
آخرین بروزرسانی
حدود ۳ سال قبل


Spoken Digit Recognition

🎙️ Spoken Digit Recognition with HMM

Overview

This project use mfcc feature extractor and Hidden Markove Model classification algorithm to recognition 0 - 9 digit of Kaggle dataset.

Dataset

Classification Algorithm

  • Hidden Markove Model

Feature Extractor Algorithm

  • Mel-frequency Cepstrum

Code Requirements

This code is written in python. To use it you will need:

  • python3
  • hmmlearn
  • librosa.feature
  • numpy
  • librosa
  • random Use pip to install any missing dependencies

General Steps

  • Split train and test data
  • Feature extract each audio using mfcc
  • Transpose each audio signal matrix
  • Vstack each transpose audio signal matrix
  • Create Hidden Markov Modle
  • Test Audio signal and predict them

Accuracy

94 %

Usage

Run python SDR.py