WITHDRAWN: Breast cancer images classification by clustering of ROI and mapping of features by CNN with XGBOOST learning

Abstract

Abstract The efficiency of most traditional classification systems is based on an appropriate representation of data, and many of the attempts are made to achieve feature-based engineering and to construct useful features through the use of prior expert knowledge of data. But on the other hand, profound or deep learning will collect and organize the discriminatory data without a domain expert engineering extractor. Convolutional neural networks (CNNs) are a deep, feed-forward network with a focus on the scientific community and business sector, which is able to achieve empirical achievements in tasks including object recognition, signal processing, speech recognition, transfer learning, and processing. This paper uses a deep learning experience to identify breast cancer and do some preliminary studies. Utilizing DDSm and Mias mammography images in the proposed technique. In the proposed methodology, CNN hybridization through the improvement of cumulative and class-wise performance, and optimization tree-based learning.