Embedded image compression based on wavelet pixel classification and sorting

Area of Science:

• Digital image processing
• Signal processing
• Computer vision

Background:

• Embedded image compression relies heavily on effective modeling and ordering of wavelet domain coefficients.
• Previous methods often implicitly or explicitly classify and sort pixels for significance coding.

Purpose of the Study:

• To propose a novel pixel classification and sorting (PCAS) scheme for enhanced embedded image compression.
• To exploit intraband correlation and improve rate-distortion performance through advanced pixel management.

Main Methods:

• Developed a PCAS scheme classifying pixels into quantized contexts using a large context template.
• Implemented pixel sorting based on estimated significance probabilities.
• Utilized fractional bit-plane coding passes for optimization.

Main Results:

• The PCAS scheme demonstrates excellent compression performance.
• The proposed technique is simple yet effective in embedded image coding.
• Achieved superior rate-distortion performance compared to existing methods.

Conclusions:

• The PCAS scheme offers a significant advancement in embedded image compression.
• The algorithm provides flexible spatial or quality scalability with manageable complexity.
• This approach effectively models and orders wavelet pixels for superior compression.