Title :
Vector quantization neural network for ECG signal compression
Author :
Bhatt, Nishith S. ; Shah, Satish K.
Author_Institution :
Comput. Eng. Dept., Sarvajanik Coll. of Eng. & Technol., Surat, India
Abstract :
Better compression results can be achieved by coding vectors instead of scalars. The proposed algorithm is used for competitive learning of a vector quantization neural network, which is used to generate a codebook for the vector quantization of ECG signals. The competitive learning algorithm can successfully group the sample group to generate the codebook and reproduce at the time of reconstruction.
Keywords :
adaptive decoding; electrocardiography; medical signal processing; neural nets; signal reconstruction; table lookup; unsupervised learning; vector quantisation; ECG; codebook generation; competitive learning; compression ratio; decoder; percent RMS difference; reconstruction; sample group; signal compression; vector coding; vector quantization neural network; Data compression; Decoding; Educational institutions; Electrocardiography; Heart; Neural networks; Parameter extraction; Postal services; Signal generators; Vector quantization;
Conference_Titel :
TENCON '02. Proceedings. 2002 IEEE Region 10 Conference on Computers, Communications, Control and Power Engineering
Print_ISBN :
0-7803-7490-8
DOI :
10.1109/TENCON.2002.1181352
