Lossless compression of VQ index with search-order coding.

Area of Science:

• Image processing
• Data compression
• Information theory

Background:

• Memoryless vector quantization (VQ) quantizes image blocks independently.
• This approach leads to redundancy in the transmitted index data.
• Exploiting interblock redundancy is crucial for efficient compression.

Purpose of the Study:

• To develop a novel lossless algorithm for image compression.
• To leverage interblock correlation in the index domain for improved VQ.
• To reduce bit rates without introducing coding distortion.

Main Methods:

• A new lossless algorithm is proposed for vector quantization.
• The algorithm exploits interblock correlation in the index domain.
• It compares current indices with previous indices along a search path, transmitting the search order.

Main Results:

• Significant reduction in bit rates achieved compared to memoryless VQ.
• No additional coding distortion was introduced by the new algorithm.
• The method is computationally efficient and simple to implement.

Conclusions:

• The proposed algorithm effectively reduces bit rates in image compression.
• Exploiting index domain correlation offers a viable path for lossless VQ enhancement.
• The algorithm provides a practical and efficient solution for image compression challenges.