Structural Languages and Biomedical Signal Analysis Using Interactive Graphics,
Abstract
The analysis of biological waveforms by computer or by special-purpose hardware can reduce the burden on trained medical manpower, make mass screening of a population more tenable, and allows the automatic analysis of large quantities of research data. Waveforms generated in electrocardiography, phonocardiography, vectorcardiography, cardiac output recording, oxygen consumption recording, electroencephalography, electromyography, and in similar applications are candidates for analysis. A characteristic of many biomedical waveforms in such applications is that many of the features by which classes of waveforms, e.g., abnormal vs. normal EKG's, are distinguished, are generally expressed by fuzzy descriptions of structural details; in EKG analysis, one hears reference to, for example 'a depressed S-T segment,' 'an inverted T wave,' 'increased duration of the QRS interval,' 'late onset of the intrinsicoid deflection.' It is the intent of this paper to describe a generalized approach to obtaining a quantitative measure of the degree of existence of such a characteristic in a given waveform. Such measures have a clear interpretation and can be used in a heuristic or linguistic program; they are continuous by 'continuously' measured features which can be used alone or in combination with more abstract features in statistical pattern recognition algorithms. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1972
- Accession Number
- AD0739258
Entities
People
- D. C. Collins
- W. S. Meisel
Organizations
- Technology Service Corporation