Biomedical Digital Signal Processing: C Language Examples and Laboratory Experiments for the IBM PC

Citation

Tompkins, W. J. (Ed.). (1993). Biomedical Digital Signal Processing: C Language Examples and Laboratory Experiments for the IBM PC. Prentice Hall.

Chapter Summary

The book is an early, laboratory-oriented introduction to digital signal processing for biomedical work, published in 1993 and running to 368 pages. The publisher describes it as covering the full scope of digital signal processing in the biomedical field: the basics of the subject, together with C-language programs for designing and implementing simple digital filters. It was issued with a floppy disk containing UW DigiScope, a program that provides an environment for laboratory experiments, for designing digital filters of different types and for viewing the results of processing on screen. The experiments can use either a real analog-to-digital converter or a virtual, software-simulated one. The contents list has twenty-two entries; the headings below group them by what they do.

Computers in Medicine, Electrocardiography and Signal Conversion

The book opens with "Introduction to Computers in Medicine" and "Electrocardiography", which set the clinical and signal context, with the ECG reappearing later in the QRS and monitoring entries, followed by "Signal Conversion", the step from an analogue physiological signal to samples a computer can process.

Digital Filtering

The core of the book is a sequence on filters: "Basics of Digital Filtering", "Finite Impulse Response Filters", "Infinite Impulse Response Filters", "Integer Filters" and "Adaptive Filters". The publisher describes the C-language programs as being for designing and implementing simple digital filters, and UW DigiScope as an environment for designing filters of different types.

Signal Averaging, Data Reduction and Other Techniques

"Signal Averaging", "Data Reduction Techniques" and "Other Time and Frequency Domain Techniques" widen the toolkit beyond filters to ways of improving signal quality, compressing data and analysing signals in time and frequency.

QRS Filters, ECG Monitoring Systems and VLSI

The application end of the main sequence is "QRS Filters" and "ECG Monitoring Systems", where the filtering methods are brought to bear on heartbeat detection and monitoring, and "VLSI in Digital Signal Processing", on implementing processing in hardware.

Software and Laboratory Material

The remaining entries support the laboratory side: "Configuring the PC for UW DigiScope", "Data Acquisition and Control Routines", "Data Acquisition and Control: Some Hints" and the "UW DigiScope User's Manual". Three further entries, "Signal Generation", "Finite Length Register Effects" and "Commercial DSP Systems", cover signal generation, the effects of finite register length and commercial DSP systems.

Key Concepts

Signal Conversion and Digital Filtering

• Analog-to-digital conversion: turning a physiological signal into samples, with real or virtual converters used in the experiments.
• FIR and IIR filters: the two main classes of digital filter, each with its own chapter.
• Integer filters: a class of filter given its own entry between the IIR and adaptive filters.
• Adaptive filters: filters whose behaviour adjusts to the signal.

Analysis Techniques and Applications

• Signal averaging: combining repeated recordings to improve the signal.
• Data reduction: techniques for representing a signal with less data.
• QRS filters: filtering aimed at detecting the QRS complex of the ECG.
• ECG monitoring systems: complete systems that apply these methods.

Implementation

• Finite length register effects: a topic given its own entry, concerning limited word length in digital processing.
• UW DigiScope: the accompanying program used to design filters and run laboratory experiments.

Critical Analysis

Strengths

• Hands-on design. The publisher stresses laboratory experiments and C-language programs, so a reader implements the filters rather than only reading about them.
• Virtual instrument option. Experiments can run on a software-simulated converter as well as on real hardware, which lets a course without acquisition equipment still do the laboratory work.
• Application focus. The contents run from signal conversion through the filter families to QRS filtering and ECG monitoring systems, and finish with implementation topics such as VLSI, finite register length and commercial DSP systems.

Limitations

• Tied to the 1993 PC environment. The title and the library catalogue record note that the software ran on IBM PC-compatible machines with 640K of memory and VGA, EGA or Hercules graphics. Readers today will need a different platform to run the exercises.
• Narrow signal scope. The contents list centres on the ECG; it names no chapters on EEG, EMG or imaging, so it is not a broad survey of biomedical signal types.
• Dated methods. Techniques later developed for nonstationary signals, time-frequency analysis and modern classification are not in a 1993 contents list.
• Basics of DSP only. The publisher presents the book as providing the basics of digital signal processing together with programs, so a reader who needs the mathematics in depth will want a separate DSP text.

Where It Fits in the Course

In the biomedical signal processing course, this is the oldest of the books and the one built around laboratory programs. The Scientist and Engineer's Guide to Digital Signal Processing by Steven W. Smith is a practical introduction to DSP in general, and Circuits, Signals, and Systems for Bioengineers by John Semmlow is a MATLAB-based introduction to the electrical engineering foundations. Biosignal and Medical Image Processing by Semmlow and Griffel covers signal and image processing for biomedical engineers, and Rangayyan's Biomedical Signal Analysis is a case-study-driven text on analysing physiological signals such as the ECG, EEG and EMG.

Tompkins can be read before or beside those books as a compact, ECG-centred course in implementing filters, or later for the historical view of how real-time ECG processing was built. For current coursework, the MATLAB-based texts are the more direct tools; Tompkins is most useful for its treatment of FIR, IIR, integer and adaptive filters and QRS detection in a single sequence.

How to Use This Book

The book is for students who have had calculus and some programming, and who want to understand digital filtering by building it. It also helps readers interested in the history of ECG monitoring.

A practical plan:

• Core: the contents from "Signal Conversion" through the filter chapters, then "QRS Filters" and "ECG Monitoring Systems".
• Choose by interest: "Signal Averaging" and "Data Reduction Techniques" for evoked-response or storage topics; "VLSI in Digital Signal Processing" and "Commercial DSP Systems" for implementation.
• Laboratory work: the software and DigiScope entries are the starting point for the experiments; reimplement them in a current environment.
• Reading alongside: use a modern DSP text for the mathematics and Rangayyan for broader signal types.

For each filter family, write the filter in a current language, run it on a short ECG recording and compare the output with the book's description of what it should do.