Denoising histopathology images for the detection of breast cancer
Source of Publication
Neural Computing and Applications
One of the leading causes of mortality for women worldwide, both in developing and developed economies, is breast cancer. The gold standard for diagnosing cancer is still histological diagnosis, despite major advances in medical understanding. Admittedly, due to the sophistication of histopathology images and the significant increase in workload, this process takes a long time. Therefore, this field requires the development of automated and precise histopathology image analysis tools. Using deep learning, we proposed a system for denoising, detecting, and classifying breast cancer using deep learning architectures that are designed to solve certain related problems. CNN-based architectures are used to extract features from images, which are then put into a fully connected layer for the classification of malignant and benign cells, as well as their subclasses, in the suggested framework. The effectiveness of the suggested framework is evaluated through experiments leveraging accepted benchmark data sets. We achieve an accuracy of 94% and an F1 score of more than 90%.
Springer Science and Business Media LLC
Breast cancer, CNN, Denoising, Detection
Zeb, Muhammad Haider; Al-Obeidat, Feras; Tubaishat, Abdallah; Qayum, Fawad; Fazeel, Ahsan; and Amin, Muhammad, "Denoising histopathology images for the detection of breast cancer" (2023). All Works. 5922.
Indexed in Scopus