A Comprehensive Survey and Adaptive Fuzzy Median Filtering Based Breast Cancer Detection and Segmentation Techniques

Background : Earliest stage detection of breast cancer helps to reduce the death rate and many researchers are put their effort to develop the image processing method to detect and segment the breast cancer tissues from the different MRI and ultrasound images. Methods : This paper mainly focuses on different effective deep learning method for earliest stage breast cancer detection and segmentation techniques. It will be helpful for the society and medical practitioners to easily understand about different type of deep learning techniques available to diagnosis the breast cancer for the women. Here, different deep learning based breast cancerous detection and segmentation technique are discussed because it plays an important role to detect the tumor in the breast. It gives information about the volume of tissue in the breast to the doctor to give treatment and helps to reduce the burden of the practitioners. Results: Measures of accuracy, precision, and false alarm rate are used to evaluate and contrast several contemporary deep learning-based detection and segmentation systems. Conclusions: Proposed DL algorithm achieved better performance by employing different noise filters, combining different methods such as conventional neural network, fuzzy logic, and Long Short-Term Memory and optimization techniques. The implemented simulation results proves that proposed adaptive fuzzy filtering method achieves greater accuracy of 95.9%, sensitivity of 94.3%, 93.8% of specificity, and 97.66% of F1 Score with less Mean Square Error Values of 25 and 40dB of PSNR values.