Reshaping inputs for convolutional neural network: Some common and uncommon methods
作者：
Highlights：
• 25 techniques for reshaping inputs for convolutional neural networks.
• Some uncommon methods along with common techniques for reshaping.
• Tested on six different datasets of multiple domains.
• Techniques applied on Inception-V3, ResNet 18, and DenseNet-121 architecture.
• Statistics about relative convergence, accuracy, agreement and chi-square test.
摘要
•25 techniques for reshaping inputs for convolutional neural networks.•Some uncommon methods along with common techniques for reshaping.•Tested on six different datasets of multiple domains.•Techniques applied on Inception-V3, ResNet 18, and DenseNet-121 architecture.•Statistics about relative convergence, accuracy, agreement and chi-square test.
论文关键词：Deep learning,Convolutional neural network,Reshaping,Resizing,Input size
论文评审过程：Received 12 July 2018, Revised 8 March 2019, Accepted 9 April 2019, Available online 10 April 2019, Version of Record 17 April 2019.
论文官网地址：https://doi.org/10.1016/j.patcog.2019.04.009
