Classification of brain MRI using hyper column technique with convolutional neural network and feature selection method
作者：
Highlights：
• A novel deep CNN model equipped with the hypercolumn masking technique is proposed for brain tumor MRI classification.
• The fused deep feature set is considered to exploit the generalization abilities of pretrained models.
• The RFE feature selection method is embedded in the model to reveal the most efficient deep features.
• The overall accuracy was obtained as 96.77% using only 200 deep features.
• A consistent and precise diagnostic model is ensured for the brain tumor MRI classification.
摘要
•A novel deep CNN model equipped with the hypercolumn masking technique is proposed for brain tumor MRI classification.•The fused deep feature set is considered to exploit the generalization abilities of pretrained models.•The RFE feature selection method is embedded in the model to reveal the most efficient deep features.•The overall accuracy was obtained as 96.77% using only 200 deep features.•A consistent and precise diagnostic model is ensured for the brain tumor MRI classification.
论文关键词：Biomedical signal processing,Decision support system,Brain MRI,Hypercolumn technique,Feature selection
论文评审过程：Received 6 July 2019, Revised 13 December 2019, Accepted 2 February 2020, Available online 4 February 2020, Version of Record 12 February 2020.
论文官网地址：https://doi.org/10.1016/j.eswa.2020.113274
