Image Quantization

KEY STAGES IN
DIGITAL IMAGE
GENERATION
• Image captured by sensor (camera) are in continuous voltage
waveform
• Continuous in term of x and y coordinates and amplitude
• Digital image are represented in digital form i.e. discrete signals
• conversion of captured continuous signal into discrete signal
1. Sampling
2. Quantization

Image Quantization
• Process of digitizing the amplitude value of the continuous signal
• Continuous grey level intensity is converted in discrete form
• Depicts the grey level resolution of image

General Steps in Image Quantization
• Measuring the grey level intensity of the signal in fixed interval in time
• Value obtained in each instant of time is converted in number and stored
• This number depicts brightness value of a particular point
• Such point is called pixel
QUANTIZATION

Image Matrix
• Represents the intensity value or pixel value
• For n bit image, intensity value ranges form 0 – 2n-1

Drawbacks of quantization
• Generally irreversible
• Results in loss of information
• Introduces distortion which cannot be eliminated

Quantizer
• Used for quantization
• Amount of distortion depends upon the quantizer
• Good quantizer results in better quantization of image

Classification of Quantizer
Quantizer
Uniform Quantizer
Non-uniform quantizer
Zero Memory Quantizer

Zero Memory Quantizer
• Simplest type of quantizer
• Quantizing a sample is independent of other sample
• Maps amplitude variable to a discrete set of quantization levels, {r1,r2…,rl}
• Based on simple comparison / thresholding with certain values, tk
• tk = transition/ decision level
• rl = reconstruction level

Uniform Quantizer
• Simplest form of zero memory quantizer
• Quantization level are uniformly spaced
• Shows absolute change in amplitude of stimulus
• tk and rk are equally spaced
• Mathematically given as:

Non-Uniform Quantizer
• Quantization levels are not necessarily equally spaced
• Logarithmic relation between quantization levels
• Shows proportional change in amplitude of stimulus
• Better for human perception
• Quantization level are assigned from histogram analysis

Non-Uniform quantization, 4 level
Uniform quantization, 4 level

Image Quantization

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Similar to Image Quantization

Similar to Image Quantization (20)

Recently uploaded

Recently uploaded (20)

Image Quantization