Convolutional neural network improvement for breast cancer classification

作者：

Highlights：

• Propose a deep classification algorithm for mammogram images.

• The deep classification performance is improved by the feature wise pre-processing.

• Application of proposed technique to detect and classify breast cancer.

• An intricate designed classification trained using extracted features.

• Achieved classification accuracy of 90.50% and specificity of 90.71%.

摘要

•Propose a deep classification algorithm for mammogram images.•The deep classification performance is improved by the feature wise pre-processing.•Application of proposed technique to detect and classify breast cancer.•An intricate designed classification trained using extracted features.•Achieved classification accuracy of 90.50% and specificity of 90.71%.

论文关键词：Supervised learning,Artificial neural network,Image processing,Medical imaging,Breast cancer classification

论文评审过程：Received 14 March 2018, Revised 11 October 2018, Accepted 5 November 2018, Available online 7 November 2018, Version of Record 22 November 2018.

论文官网地址：https://doi.org/10.1016/j.eswa.2018.11.008