Related Experiment Video
Updated: Jun 14, 2025
04:48
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
Published on: July 5, 2024
379
Dual-Tree Complex Wavelet Pooling and Attention-Based Modified U-Net Architecture for Automated Breast Thermogram
Lalit Garia1,2, Hariharan Muthusamy3
1Department of Electronics Engineering, National Institute of Technology Uttarakhand, Srinagar (Garhwal), 246174, Uttarakhand, India.
Journal of Imaging Informatics in Medicine
|September 3, 2024
Summary
This study introduces a novel AI method for breast thermogram analysis, achieving 99.90% accuracy in detecting breast cancer. The advanced technique enhances early detection and improves treatment outcomes for patients.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Biomedical Engineering
Background:
- Thermography offers a non-invasive method for early breast cancer detection by analyzing thermal variations.
- Preprocessing techniques like region of interest (ROI) segmentation are crucial for accurate thermogram analysis.
Purpose of the Study:
- To develop and evaluate a modified U-Net architecture (DTCWAU-Net) for precise breast thermal image segmentation.
- To enhance the accuracy of breast cancer classification from thermograms using advanced feature extraction and machine learning.
Main Methods:
- A modified U-Net architecture incorporating dual-tree complex wavelet transform (DTCWT) and attention gates was utilized for ROI segmentation.
- Texture, histogram, and deep features were extracted from segmented thermograms.
- Neighborhood Component Analysis (NCA) was employed for feature selection, followed by Random Forest classification.
Main Results:
- The DTCWAU-Net achieved an average Dice coefficient of 93.03% and 94.82% sensitivity for breast thermal image segmentation.
- The proposed method, using VGG16 deep features with NCA and Random Forest, reached a classification accuracy of 99.90% for breast cancer detection.
- The methodology demonstrated superior performance compared to existing state-of-the-art approaches.
Conclusions:
- The proposed DTCWAU-Net and subsequent classification framework show significant potential for accurate breast cancer screening using thermography.
- The study highlights the effectiveness of AI in improving early detection rates and patient outcomes in breast cancer care.
