Does this model supports opencv dnn inference module?

[ghimiredhikura]

Thank you @Zheng222 for this wonderful work.

I am trying to use opencv dnn inference module for your SR method. I will be grateful if you can explain these lines of code in your IDN-Caffe/test/test_IDN.m file.

   im_input=permute(im_l,[2,1,3]);
    net.blobs('data').reshape([size(im_input),1,1]);
    net.reshape();
    net.blobs('data').set_data(im_input); 
    net.forward_prefilled();
    im_result=net.blobs('upsample').get_data();

    im_h=im_result'+mycrop(im_b,up_scale);
    time_idn(i)=toc;

Best, Deepak

[Zheng222]

@ghimiredhikura Hello, the above code segment indicates that preparing the input image first and then sent it to the network. It is a basic operation in MatCaffe, I recommend you learn the official guidance of matcaffe. If you are familiar with pytorch, you can reference a third party implementation of IDN.

[ghimiredhikura]

Thank you @Zheng222 . Actually I successfully reproduce results in using opencv dnn.