showman-sharma / speech_writing-recognition

We are given 2 different problems to solve. 1. Isolated spoken digit recognition 2. Telugu Handwritten character recognition Both these datasets were given as a time series. 2 different methods were used to solve each of the problem: 1. Dynamic Time Warping 2. Hidden Markov Models

SPEECH AND WRITING RECOGNITION

---

Authored by Team 18: V S S Anirudh Sharma, EE18B036 Hema Landa, EE19B036

---

The assignment deals with Time Series and Normal data classification models. The following instructions will help the users execute Regression.py and Classification.py in the right way.

TIME SERIES CLASSIFICATION

With the Speech and Handwriting data class folders, HMM-Code and the following codes in the same folder, here are the commands:

DTW_speech.py

Run the following command:

>>python3 DTW_speech.py <optional: Path of Data super folder>

Please fill in the paths of the .txt dataset files accordingly. Default is current folder path.

DTW_writing.py

Run the following command:

>>python3 DTW_writing.py <optional: Path of Data super folder>

Please fill in the paths of the .txt dataset files accordingly. Default is current folder path.

HMM_speech.py

Run the following command:

>>python3 HMM_speech.py <optional: Path of Data super folder>

Please fill in the paths of the .txt dataset files accordingly. Default is current folder path.

HMM_writing.py

Run the following command:

>>python3 HMM_writing.py <optional: Path of Data super folder>

Please fill in the paths of the .txt dataset files accordingly. Default is current folder path.