Handcrafted vs. non-handcrafted features for computer vision classification

作者：

Highlights：

• Deep layers of Convolutional Neural Networks are used for feature extraction.

• Handcrafted and learned features are used together to extract information.

• Different architectures for combining handcrafted and learned features are proposed.

• Combination of different features is used to solve image classification problems.

摘要

•Deep layers of Convolutional Neural Networks are used for feature extraction.•Handcrafted and learned features are used together to extract information.•Different architectures for combining handcrafted and learned features are proposed.•Combination of different features is used to solve image classification problems.

论文关键词：Deep learning,Transfer learning,Non-handcrafted features,Texture descriptors,Texture classification,Ensemble of descriptors

论文评审过程：Received 26 May 2016, Revised 16 April 2017, Accepted 28 May 2017, Available online 3 June 2017, Version of Record 12 June 2017.

论文官网地址：https://doi.org/10.1016/j.patcog.2017.05.025