Related Experiment Video
Updated: Jul 16, 2025
05:56
Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
Published on: April 14, 2023
2.5K
ODQN-Net: Optimized Deep Q Neural Networks for Disease Prediction Through Tongue Image Analysis Using Remora
S V N Sreenivasu1, P Santosh Kumar Patra2, Vasujadevi Midasala3
1Department of Computer Science and Engineering, Narasaraopeta Engineering College (A), Narasaraopet, India.
Big Data
|September 13, 2023
Summary
This study introduces an optimized deep Q-neural network (ODQN-Net) for accurate disease prediction from tongue images, improving upon traditional methods. The ODQN-Net achieves high accuracy in classifying multiple diseases using enhanced image processing and feature extraction techniques.
Area of Science:
- Artificial Intelligence
- Medical Imaging
- Traditional Indian Medicine
Background:
- Traditional Ayurvedic medicine relies on manual tongue analysis for disease diagnosis, which is time-consuming and lacks precision.
- Existing machine learning models for tongue-based disease prediction have not achieved sufficient accuracy, especially for multiclass classification.
- Accurate and automated disease identification from tongue images remains a significant challenge.
Purpose of the Study:
- To develop an optimized deep Q-neural network (ODQN-Net) for enhanced disease identification and classification from tongue images.
- To improve the accuracy and efficiency of disease prediction compared to existing methods.
- To address the limitations of manual diagnosis and current AI approaches in Ayurvedic medicine.
Main Methods:
- Image enhancement using the multiscale retinex approach for quality improvement and noise reduction.
- Feature extraction utilizing the local ternary pattern for color-based analysis and the Remora optimization algorithm for efficient selection.
- Classification of diseases using an optimized deep Q-neural network (ODQN-Net) model.
Main Results:
- The proposed ODQN-Net achieved a high accuracy of 99.17% on a tongue imaging dataset.
- Excellent performance metrics were recorded, including an F1-score of 99.75% and a Mathew's correlation coefficient of 99.84%.
- The ODQN-Net demonstrated superior performance compared to current state-of-the-art approaches.
Conclusions:
- The ODQN-Net model offers a highly accurate and efficient solution for automated disease prediction and classification from tongue images.
- This AI-driven approach has the potential to revolutionize Ayurvedic diagnostics by overcoming the limitations of manual inspection.
- The study highlights the effectiveness of combining advanced deep learning with optimized feature extraction for medical image analysis.
