Image Processing: Sampling

Image Sampling

• Image sampling is a process of creating a digital image from data acquired by a camera or other imaging instrument
• It requires a two-dimensional pattern to represent measurements (light intensity or color) numerically
• Examples: Lena image sampled with two different sampling structures, resulting in different resolutions (256x256 and 64x64 pixels)

Sampling Frequency

• Sampling frequency affects the resolution of the image
• Lower sampling frequency results in a lower resolution image (e.g., 64x64 pixels)
• Higher sampling frequency results in a higher resolution image (e.g., 256x256 pixels)

Pixels

• A digital image is a two-dimensional array of pixels
• Each pixel corresponds to a small square area of the image
• Pixels are indexed by x and y coordinates, with x and y taking integer values

Quantization

• Quantization is the process of digitizing the value of the image function f(x, y)
• It involves replacing a continuously varying f(x, y) with a discrete set of quantization levels
• The accuracy of quantization depends on the number of quantization levels used

Effects of Reducing Number of Samples

• Reducing the number of samples can result in a loss of detail and a checkerboard pattern effect
• Examples: subsampling an image from 1024x1024 to 512x512, 256x256, and 32x32 pixels
• The effects of reducing the number of samples become more pronounced as the image size decreases

Quantization of an Image

• Quantization of an image involves representing the brightness of each pixel by an integer value
• The number of quantization levels determines the accuracy of the representation
• Examples: quantization with 8 bits per pixel (256 reconstruction levels), 4 bits per pixel (16 reconstruction levels), and 2 bits per pixel (4 reconstruction levels)