Image Sampling and Quantization Experiment
Aryan Chahil
21CS3008
Objective:
The aim of this experiment was to understand the concepts of image sampling and quantization
by implementing them using MATLAB. The process involved resizing an image to observe the
effects of sampling and reducing intensity levels to examine quantization.
Observations and Insights:
1. Visual differences and similarities between the images
● The original image has the highest level of detail and smooth transitions between
intensity levels. It maintains the full resolution and grayscale intensity range.
● The sampled image appears smaller due to the reduction in spatial resolution. As the
sampling factor decreases, fine details start to disappear, and the image becomes
blocky. However, the overall structure of the image remains recognizable.
● The quantized image has reduced grayscale intensity levels, causing noticeable banding
effects (false contours). Unlike the original image, which has smooth intensity variations,
the quantized image may exhibit abrupt transitions between intensity levels.
2. Impact of sampling and quantization on image quality
● Sampling Effect: Lowering the sampling factor reduces spatial resolution, making the
image appear pixelated and less detailed. However, moderate reductions still retain the
key features.
● Quantization Effect: Reducing intensity levels increases the visibility of artificial bands
(false contours). If too few levels are used, the image may lose essential shading details,
making it look unnatural.
● Trade-Off: Increasing either the sampling rate or quantization levels improves quality but
increases storage and processing requirements.
3. Compare the original image, sampled image, and quantized image
● The original image has the highest clarity, detail, and smooth tonal variations.
● The sampled image is a lower-resolution version of the original, with visible pixelation at
extreme reductions.
● The quantized image maintains the original resolution but lacks smooth intensity
transitions, leading to artificial banding.
● When both sampling and quantization are applied, the combined effects cause a
significant loss of detail and visual degradation.
Original Image
Sampled Image
Quantized Image
4. Experiment with different sampling factors and quantization levels
Sampling Experiments:
● Sampling factor = 0.75: The image retains most details with minor size reduction.
● Sampling factor = 0.5: The image is noticeably smaller, with some loss of fine details.
● Sampling factor = 0.25: The image becomes highly pixelated, making small objects
unrecognizable.
Quantization Experiments:
● 256 levels (original): Smooth intensity variations with no visible artifacts.
● 16 levels: Some banding starts appearing in smooth regions.
● 8 levels: False contours become noticeable, reducing realism.
● 4 levels: Severe posterization, making the image look unnatural.
