A Review on Breast Cancer Diagnosis in Mammography Images Using Deep Learning Techniques

Abstract

The Breast Cancer is second of the dangerous type diseases that is very effective for women in the world. Breast Cancer is spread very fast in last stage. It’s not give a more time for patients for treatment. We need accurate and efficient diagnosis system over a short period of time that can processed a large amount of data in a short time. It’s only possible using deep learning techniques. We are studying on a new diagnosis system for detecting Breast cancer in early stage. But in this paper we are describing the all techniques and images processing method for segmentation and filter images for breast cancer diagnosis. In this paper we have describe all algorithm and techniques for image pre-processing and post progressing. We have reviewed more than 20 papers from different publisher and find out that many researcher working on breast cancer diagnosis techniques. It’s a very latest and effective techniques for helping and saving the human life. In this paper we have also describe some experimental part of images segmentation and filter images with different filter method. It’s a very interesting area for researcher. In this paper we have used (Mammography Image Analysis Society) MIAS dataset having 95 images in .pgm format with 1024 X 1024 pixel. All paper divided in two main part first is data collection and image pre-processing. In this paper we have describe some most used deep learning pre-trained model that is very effective for images pre-processing. As per reviewed study the AlexNet, GoogleNet and VGG16 are provided best accuracy for detecting breast cancer in mammography images. We also used many filter method for image quality improvement. Filters are used to filter unwanted things or object in a spatial domain or surface. The main objective of the filter are to improve the quality of the images by enhancing and improve interoperability of the images. In this paper we also showing the comparison between most used pre-trained model that are used for breast cancer diagnosis.