CODEWITHSUNDEEP
imagescipydeep-learningpixel
AgentX
AgentX

Reputation: 1462

Image Classification - What is pixel depth?

I am following the Deep Learning course on Udacity. I am slightly confused about pixel depth and how it is being used in below code:

image_size = 28  # Pixel width and height.
pixel_depth = 255.0  # Number of levels per pixel.

image_data = (ndimage.imread(image_file).astype(float) -
                      pixel_depth / 2) / pixel_depth

Can someone explain me why we are doing pixel_depth / 2) / pixel_depth while reading the image into N-d array?

Upvotes: 0

Views: 264

Answers (2)

cedrou
cedrou

Reputation: 2790

Pixel depth is the number of values that a pixel can take. For 8-bits images, this is 256 (but they use 255 here).

The code here is used to normalize and center the pixel values into the interval [-0.5,0.5].

Upvotes: 1

user8502296
user8502296

Reputation:

Depth in CV just refers to the data type. Depth of 255.0 implies 8-bits for each pixel and so on.

pixel_depth / 2) / pixel_depth

This bit of code seems a little weird at first but it's purpose is to normalize the image to a range of -0.5 to 0.5, which is a common tactic to simplify image processing.

Upvotes: 1

Related Questions