Azmy, M. (2022). Recognition of Ultrasound Images of Breast Cancer Based on Discrete Sinc Transform and Deep Learning. International Journal of Cancer and Biomedical Research, 6(1), 41-45. doi: 10.21608/jcbr.2021.65002.1185
Mohamed Azmy. "Recognition of Ultrasound Images of Breast Cancer Based on Discrete Sinc Transform and Deep Learning". International Journal of Cancer and Biomedical Research, 6, 1, 2022, 41-45. doi: 10.21608/jcbr.2021.65002.1185
Azmy, M. (2022). 'Recognition of Ultrasound Images of Breast Cancer Based on Discrete Sinc Transform and Deep Learning', International Journal of Cancer and Biomedical Research, 6(1), pp. 41-45. doi: 10.21608/jcbr.2021.65002.1185
Azmy, M. Recognition of Ultrasound Images of Breast Cancer Based on Discrete Sinc Transform and Deep Learning. International Journal of Cancer and Biomedical Research, 2022; 6(1): 41-45. doi: 10.21608/jcbr.2021.65002.1185
Recognition of Ultrasound Images of Breast Cancer Based on Discrete Sinc Transform and Deep Learning
Biomedical Engineering Department, Medical Research Institute, Alexandria University, Alexandria, Egypt
Abstract
Background: : The most common cause of cancer-related death among women all over the world is breast cancer. Altough ultrasound emaging I useful in diagnosis of the patients, itsaccuracy needs more improvement. Aim: This paper presents a new automated system to precisely classify images of ultrasound of breast cancer to help radiologists for better diagnosis. Materils and Methods: Deep learning neural network algorithm was used in the classification step. Features of images are extracted using discrete wavelet transform. After that, discrete sinc transform (DSNT) was obtained. Then gray-level co-occurrence matrix was used. After that mean was calculated. Several discrete transforms are applied as discrete cosine transform (DCT), discrete tan transform (DTT), discrete sine transform (DST) and discrete sinc transform (DSNT). DSNT was chosen because it has the largest accuracy rate after the classification step. Results: The obtained accuracy percent is 99%. The specificity rate is 98%. The sensitivity rate is 100%. F-measure rate is 99.0%. F-score is 0.99. Conclusion: Our study point pout that future studies are needed to devlope a new feature extraction method to achieve higher accuracy rate.
Top