Stacked auto-encoder based tagging with deep features for content-based medical image retrieval

作者：

Highlights：

• Proposed method provides an effective and efficient solution method for highly unbalanced medical benchmark datasets.

• This is the first study in which data imbalance using the feature vector at the output of the FCL layer.

• Enabling the reduced search area to be used more effectively.

• Converting high-level features into few digits using unsupervised sAE is considerably improves the performance.

摘要

•Proposed method provides an effective and efficient solution method for highly unbalanced medical benchmark datasets.•This is the first study in which data imbalance using the feature vector at the output of the FCL layer.•Enabling the reduced search area to be used more effectively.•Converting high-level features into few digits using unsupervised sAE is considerably improves the performance.

论文关键词：CBMIR,CNN,Retrieval,SMOTE,IRMA,Auto-encoder

论文评审过程：Received 22 April 2020, Revised 23 June 2020, Accepted 24 June 2020, Available online 4 July 2020, Version of Record 10 July 2020.

论文官网地址：https://doi.org/10.1016/j.eswa.2020.113693