Training deep convolution regression network with multi dimensional output

Question

Stephen Martinez 2017-5-6

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/339080-training-deep-convolution-regression-network-with-multi-dimensional-output

回答： Gautham Sholingar 2017-5-16

I'm taking in an input image of 512x512 and running it through an alexnet type architecture. The output needs to be another image. The image can be arranged as either [512pixels, 512pixels,1channel,N number of examples] or as [262144,N]. Niether of them are working. The trainNetwork function is being used. Any help you could provide would be greatly appreciated.

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Gautham Sholingar 2017-5-16

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/339080-training-deep-convolution-regression-network-with-multi-dimensional-output#answer_267266

在 MATLAB Online 中打开

1) Could you share some of your code to explain how you are setting up the neural network?

2) Is there a specific error you notice when you try to run the 'trainNetwork' function?

3) Have you had a chance to look at some of the shipped examples which explain how to use an 'm by m by 1' channel image dataset for training a network? Run the following command in the MATLAB command window to look at the example which shows this:

>> openExample('nnet/UseDataInImageDatastoreForTrainingCNNExample')

The following documentation link explains this example: https://www.mathworks.com/help/nnet/ref/trainnetwork.html#bvg3o5h-1