Block permutation coding of images using cosine transform

Abstract-There are several applications in information transfer and storage where the order of source letters is irrelevant at the destination. For these source-destination pairs, multiset communication rather than the more difficult task of sequence communication may be performed. In this work, we study universal multiset communication. For classes of countable-alphabet sources that meet Kieffer's condition for sequence communication, we present a scheme that universally achieves a rate of n + o(n) bits per multiset letter for multiset communication. We also define redundancy measures that are normalized by the logarithm of the multiset size rather than per multiset letter and show that these redundancy measures cannot be driven to zero for the class of finite-alphabet memoryless multisets. This further implies that finite-alphabet memoryless multisets cannot be encoded universally with vanishing fractional redundancy.

show abstract

Block permutation coding of images using cosine transform

Cited by 5 publications

References 25 publications

Source coding with a permutation-based reversible memory-binding transform for data compression in categorical data domains

Source coding with a permutation-based reversible memory-binding transform for data compression in categorical data domains

Partial-Order Bit-Allocation Schemes for Lowrate Quantization

On Universal Coding of Unordered Data

Contact Info

Product

Resources

About