Computational Mechanisms for Exploiting Temporal Redundancies Supporting Multichannel EEG Compression

Abstract

Multichannel electroencephalogram (MCEEG) recordings generally result in humongous volume of data that places constraint on space and time. Online transmission of such data demands schemes rendering significant performance with lesser computations. To compact such data, numerous compression algorithms have been introduced in the literature. Heretofore single channel algorithms when extended to multichannel applications do not accomplish remarkable results. If achieved it generally results in higher computational cost. Much of this chapter deals with the development of computationally simple algorithms that aim to reduce the computational aspects without comprising on the compression and decompression performance. Also, the amicability of this implementation supporting efficient storage with low bandwidth is performed. The objective of this chapter is to introduce some of the basic supporting concepts for exploiting data representation redundancies that aid compression. Accordingly, simple and novel compression schemes with its simulation comparison are presented in this chapter. The potency of the direct domain compression model is assessed in terms of compression ratio and reconstruction error between the original and reconstructed dataset. A significant compression with substantially low Percent Root-mean-square Difference (PRD) is accomplished by the novel compression schemes, thereby upholding diagnostic information of EEG for telemedicine applications with higher reconstruction accuracy and reduced computational load.