0

I want to automatically detect the cars on an image (rgb) as possible values for each pixel. I would like to create a result image with 0 to 100 possible values,for example, if the probability of a pixel being in the car class is 98%, then the result of this pixel value should be 98.

I want the output image of the input image to be the same size.

my data labels (include only 0-100 pixel value) For instance:

img1.shape = 2238, 3126, 3                  img1.shape   2238, 3126 
img2.shape = 668, 7234, 3                   img2.shape = 668, 7234
img3.shape = 4225, 5598, 3                  img3.shape = 4225, 5598

...

How can I train these different sizes of data and what kind of network should I create for this problem?

user25342
  • 1
  • this is a pretty general question you're asking there. Is it just about the specific dimensions of your images or do you want us to suggest a complete ml routine? – flurble Apr 29 '19 at 10:30
  • I want to learn, how can ı train different size of image using keras beckand tensorflow. – user25342 Apr 29 '19 at 10:46

0 Answers0