How would one go about training the network with one channel images?

Question

How would one go about training the network with one channel images?

patakk opened this issue 7 years ago · 2 comments

Let's say I have normal RGB images and would like to only take the Y component of the YUV color space, how would one go about doing that and setting up tfrecords?

Answer 1 · 2019-04-28T20:41:02.000Z

I don't know TF but I suspect it is similar to pytorch

you simply provide images with dimensions (xsize, ysize, depth) where depth is not 3 (RGB) but 1 (Y)

Answer 2 · 2019-05-02T20:39:34.000Z

I honestly can't remember why I even asked this question, but I guess I wanted to know how to convert RGB images to grayscale and save it to tfrecords, but I could've just googled it..