Robust Automated Parkinson's Syndrome Detection Based on Phonetic Features

Parkinson?s disease is the slowly progressive neurodegenerative disorder. Parkinson?s disease begins with shaking palsy, is a neurological disease because of the breakdown of cells in the vicinity of the midbrain called substantia nigra. The movement control center of the brain that is answerable fo

Project Title

Robust Automated Parkinson's Syndrome Detection Based on Phonetic Features

Project Area of Specialization

Artificial Intelligence

Project Summary

Parkinson’s disease is the slowly progressive neurodegenerative disorder. Parkinson’s disease begins with shaking palsy, is a neurological disease because of the breakdown of cells in the vicinity of the midbrain called substantia nigra. The movement control center of the brain that is answerable for generating dopamine. An imbalance in dopamine production leads to hypokinetic movement disorder, characterized by out-of-control movements. People over the age of 60 are most likely to suffer from Parkinson's disease. 10-15% of patients with Parkinson's under 40 are classified as having "young onset Parkinson's. It is estimated that over 90% of patients with Parkinson’s disease develop a speech disorder known as hypokinetic dysarthria.

The proposed system has shown  that  changes in speech can be used as a measurable indicator for early Parkinson’s detection.The suggested system used several combinations of feature selection approaches, classification algorithms and designed the model with three feature selection methods such as mutual information gain, Extra tree, genetic algorithm and three classifiers for test .The  speech dataset of UCI (university of california, Irvine) is used.The outcomes indicate that  the use of the feature selection method is advantageousbecause it reduces complexity and increases accuracy. However, the proposed framework is lack of testing on significant speech and voice datasets. Multiple voice samples were collected from the same person, and the same voice sample was used for both the training and testing datasets.

Our Framework utilizes latest datasets of UCI (University of California, Irvine) and Oxford University. Using our proposed model, we can reduce reverberation, background noise and distortion by applying enhancement algorithms. Our proposed system will help medical professinals to diagnose Parkinson’s syndromes at early stage .It will also help medical professionals to avoid lengthy diagnostic methods. According to the proposed web-based system, symptoms such as softening of voice, speaking in the same pitch, vocalizing with a breathy tone will allow us to predict whether a subject is “Healthy” or “NOT”.

Project Objectives

1. Our suggested Robust Automated system will help experts to diagnose Parkinson disease at early stage and also help to slow down its progression in patients.
2. An Automated system will be able to predict tremors in voice, monotone voice, and softening of voice based on symptoms. It will predict the probability that patient is “Healthy Person" or "Diseased person".
3. Advanced Robust Automated Platform will also helpful for medical staff to make precise decisions and save lots of diagnosing time.
4. Proposed vigorous Automated System will also incorporate pronunciation of vowel letters like /ah/, /eh/, /Iamh/ to predict person is "Healthy” or “NOT". 
5.  Parkinson’s disease detection system will provide recommendations such as Regular Exercises, Healthy Diets and information about Expert Consultants, if patient is turned out to be a diseased individual.

Project Implementation Method

1. Proposed Robust Automated System will deliver early detection of Parkinson’s disease, we are using “Librosa” Python library to input the user’s voice signal into the system using a microphone which generates Mel Spectrogram for extracting features of voice signals.
2. The task of feature selection will also be done using minimum redundancy and maximum relevancy (mRMR) & recursive features alimentation (RFE) techniques.
3. Datasets are collected from UCI (University of California Irvine) repository and Oxford University Phonetic Signals  datasets.
4. GUI for user interaction will be based on Flask. HTML, CSS, bootstrap, and JavaScript will be applied for front-end interactive user interface.
5. Proposed model will provide recommendations such as Regular Exercises, Healthy Diets and information about Expert Consultants, if patient found to be a diseased.

Benefits of the Project

1. Our proposed Robust Automated System will provide a platform for the detection of Parkinson's disease by voice signal features.
2. Proposed framework will reduced complexity in phonetic sound signals,  preserve physical function and enhance the quality of life.
3. Reduce the cost of medical treatement and diagnostic resources.
4. The framework will help to slow down the progression of Parkinson disease.
5. Not every patient needs to visit the hospitals this Automated System based on Phonetic signal data will predict the Probability of Parkinson Disease, it will save the medical resources in this crisis.

Technical Details of Final Deliverable

The final deliverable will be a Web-Based advanced system for early detection against Parkinson’s Disease. The self operative system will deliver early and precise detection of Parkinson’s disease by utilizing natural language processing tools and machine learning models.

Parkinson's disease probability prediction based on symptoms (softening of voice, speak in the same pitch (monotone), vocalizing with a breathy tone, pronunciation difficulties while reading and writing, vocal tremors).

The proposed sturdy self operative system will diagnose the early Parkinson's disease which reduce the time consumption and also will enhance the medical professional capabilities.

 The proposed Robust Automated System will provide different recommendations like (Healthy Diets, Regular Exercises, and Details of Expert Consultants) to help the patients to overcome the effect of this disease. Parkinson’s disorder detection system is bendy, easy to use, and can be implemented everywhere in clinical institutes.

Final Deliverable of the Project

Software System

Core Industry

IT

Other Industries

Health

Core Technology

Artificial Intelligence(AI)

Other Technologies

Sustainable Development Goals

Good Health and Well-Being for People

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com