V Congreso Latinoamericano de Ingeniería Biomédica (CLAIB2011)

Compresión de señales EEG basada en descomposición en subbandas y codificación de Golomb
Bazán Prieto Carlos Alberto, Blanco Velasco Manuel, Cárdenas Barrera Julián, Cruz Roldán Fernando

Última modificación: 2011-03-26 11:40


The need for electroencephalographic (EEG) data storage, processing and transmission for further analysis has made a great demand for efficient compression techniques. This paper presents a new compression method for EEG signals using subband descomposition and Golomb coding, enabling exploit the prevalence of low amplitude samples.  Uniform scalar quantization is applied after a quality-controlled thresholding process and the elimination of dead zone. Nonzero thresholded values and runs of zeros are coded using Golomb-Rice and Golomb-Exp encoding respectively. The proposed method attains compression ratios that are better than other published results for the same reconstruction quality.