Sampling Types
Uniform sampling
Uniform sampling has the same row and
column spacing in all areas of an image.
Non-uniform sampling
Non-uniform sampling is adaptive
depending on image characteristics and
aims to avoid missing information.
Image areas that contain high detail are
sampled finer, while homogeneous
areas can be sampled more roughly.
Reduction in Sampling Resolution
Effect The Number of Samples
Quantization Quantization corresponds to a discretization of the
intensity values. That is, of the co-domain of the
function.
After sampling and quantization, we get
Quantization
Quantization corresponds to a transformation Q(f)
Typically, 256 levels (8 bits/pixel) suffices to
represent the intensity. For color images, 256
levels are usually used for each color intensity
Quantization
Uniform Quantization
For example, if the output of an image sensor
takes values between 0 and M, and one wants
L quantization levels, the uniform quantizer
would take
Uniform Quantization
Uniform Quantization
Color Depth
The number of bits used to indicate the
color of a single pixel.
Bits per pixel
Direct color
8-bit color → 3 bits on R and G channels,
and 2 bits on B channel
16-bit color → High color (RGBA)
18-bit color → LCD
24-bit color → true color
Different Number of Gray Levels
Different Number of Color Levels
Non-Uniform Quantization
The previous approach considers that all values are equally
important and uniformly distributed.
Image Histogram
Image Histogram Samples
Non-Uniform Quantization
What to do if some values are more
important than others?
In general, we can look for quantization
levels that “more accurately” represent
the data.
To minimize the mean square error (MSE)
we can use the Max-Lloyd algorithm to
find the quantization levels with
minimum MSE.
Non-Uniform Quantization
Max-Lloyd algorithm:
1. Choose initial quantization levels;
2. Assign points to a quantization level and
reconstruct image;
3. Compute the new quantization levels as
the mean of the value of all points
assigned to each quantization level.
4. Go back to 2 until reduction of MSE is
minimal.
Non-Uniform Quantization
