Abstract:
This study develops and tests a global mammographic image feature-based computer-aided diagnosis (CADx) scheme to predict malignancy. 1,959 cases were retrospectively imaged. All suspicious lesions were biopsied. 737 are malignant and 1,222 benign.
Each case has four craniocaudal and mediolateral oblique left and right breast mammograms. CADx pre-processes mammograms, generates two frequency-domain image maps using discrete cosine transform and fast Fourier transform, computes bilateral image feature differences from left and right breasts, and uses a support vector machine (SVM) to predict malignancy.
The original mammograms and two transformation maps yielded three image feature subgroups. Four SVMs with three subgroups of image features and fusion of all features were trained and tested using 10-fold cross-validation.
Image features from one of three subgroups yield AUCs of 0.85 to 0.91. The fourth SVM performs significantly better with AUC = 0.96±0.01 (p<0.01). This study shows that a high-performance global image feature analysis-based mammogram CADx scheme is possible. This new CADx approach avoids breast lesion segmentation problems, making development faster and application more reliable.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
