Medical image classification via multiscale representation learning

作者：

Highlights：

• Multiscale representation learning method is used to capture the intrinsic scales.

• Fisher vector technique is used to encode the extracted features.

• A fixed-length image representation is obtained regardless of the input size.

• Providing more abundant information of high-order statistics.

摘要

•Multiscale representation learning method is used to capture the intrinsic scales.•Fisher vector technique is used to encode the extracted features.•A fixed-length image representation is obtained regardless of the input size.•Providing more abundant information of high-order statistics.

论文关键词：Multiscale feature learning,Sparse autoencoder,Fisher vector,Image classification

论文评审过程：Received 24 February 2017, Revised 19 May 2017, Accepted 20 June 2017, Available online 29 June 2017, Version of Record 12 August 2017.

论文官网地址：https://doi.org/10.1016/j.artmed.2017.06.009