Dynamic adversarial domain adaptation based on multikernel maximum mean discrepancy for breast ultrasound image classification

作者：

Highlights：

• MK_DAAN is proposed for breast ultrasound image classification.

• An adaptive layer is added to MK_DAAN to further align the feature distribution.

• MK_DAAN adopts MK-MMD in the adaptive distance measurement.

摘要

•MK_DAAN is proposed for breast ultrasound image classification.•An adaptive layer is added to MK_DAAN to further align the feature distribution.•MK_DAAN adopts MK-MMD in the adaptive distance measurement.

论文关键词：Deep transfer learning,Breast ultrasound image,Adversarial domain adaptive,Unsupervised domain adaptation

论文评审过程：Received 15 February 2022, Revised 10 May 2022, Accepted 22 June 2022, Available online 28 June 2022, Version of Record 1 July 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.117978