Dataflow Model Design of Gabor Wavelet Transform for ECG Feature Extraction System

摘要

This paper presents a dataflow design of 1-D Gabor Wavelet Transform based on an Algorithm/Architecture Co-exploration (AAC) design methodology for an electrocardiogram (ECG) Gabor feature extraction system Traditional design methodologies only focus on execution time which ignores the relationship between algorithm and architecture design; on the contrary AAC considers overall complexity between algorithm and architecture such as number of operations data storage data transfer and degree of parallelism and these complexity can be extracted by dataflow with multiple abstract levels and various data granularities Due to dataflow exploration as a bridge between algorithm and architecture proposed dataflow achieve multiplication reduction with Gabor symmetry property addition reduction with commonality and critical path reduction with retiming; more efficient than related works and verified by CAL which is a high-level programming language for writing dataflow to model data-driven processing

獎項日期	2016 9月 2
原文	English
監督員	Gwo-Giun Lee (Supervisor)