Transfer learning with different input size

How to do transfer learning for the input of dimension (64x64x1)using Resnet-50, which has three-channel input (224x224x3)?

• You can use a ResizeLayer and a ReplicateLayer to resize the input to have the same original dimensions. However I don't know what the approved academic method is, it would be worth casting around on the arXiv for some papers that discuss the issue. – Carl Lange Aug 20 '19 at 7:09