Breast Cancer Diagnosis: Two-Dimensional Discrete Orthonormal Stockwell Transform and Histogram of Oriented Gradients
Year: 2026
Type: article
Abstract: Thermographic imaging has emerged as a non-invasive modality for breast abnormality assessment. This paper presents a reliable hybrid feature extraction approach utilizing Two-Dimensional Discrete Orthonormal Stockwell Transform (2D-DOST) and Histogram of Oriented Gradients (HOG) for breast cancer recognition in thermal imaging. The proposed method leverages the complementary strengths of both techniques: 2D-DOST captures spatial-frequency characteristics, and HOG extracts discriminative edge-based descriptors essential for malignancy detection. The model was evaluated on 252 thermogram images from the DMR dataset using 5-fold cross-validation. Among four different classifiers, Random Forest (RF) achieved the highest performance with 97% accuracy, 95% sensitivity, 99% specificity, 97% F1-score, and an AUC of 96.72%. These findings underscore the potential of advanced image processing techniques to improve diagnostic accuracy in breast cancer screening and support the development of more effective early detection methods.
Language: English
Subfield: Radiology, Nuclear Medicine and Imaging
Field: Medicine
Domain: Health Sciences
Open Access status: diamond
OpenAlex ID: W7237200706
