Related Experiment Videos
Computer vision-based breast self-examination stroke position and palpation pressure level classification using
Melvin K Cabatuan1, Elmer P Dadios, Raouf N G Naguib
1De La Salle University, Manila, Philippines. melvin.cabatuan@dlsu.edu.ph
Summary
This study introduces a computer vision system for breast self-examination (BSE) training. It accurately identifies BSE stroke positions and palpation levels using artificial neural networks and wavelet transform.
Area of Science:
- Medical imaging
- Computer vision
- Biomedical engineering
Background:
- Breast self-examination (BSE) is a crucial method for early breast cancer detection.
- Effective training is essential for accurate BSE performance.
- Existing training methods lack objective feedback mechanisms.
Purpose of the Study:
- To develop a computer vision-based system for breast self-examination (BSE) training and guidance.
- To classify BSE stroke positions and palpation levels (light, medium, deep).
Main Methods:
- Image frames extracted from BSE videos.
- Processing using color information, shape, and texture analysis.
- Application of wavelet transform and first-order color moments.
- Utilizing artificial neural networks for classification.
Main Results:
- High accuracy in identifying BSE stroke positions (97.8%).
- Significant accuracy in classifying palpation levels (87.5%).
- Demonstrated feasibility of a computer vision approach for BSE guidance.
Conclusions:
- The developed system shows promise for enhancing BSE training.
- Computer vision techniques can provide objective feedback for BSE.
- Further development could improve accuracy and expand system capabilities.