Fuzzy and Neuro-Fuzzy Systems in Medicine

by Teodorescu, Horia-Nicolai; Kandel, Abraham; Jain, L. C.

• ISBN13: 9780849398063
• ISBN10: 0849398061
• Format: Hardcover
• Copyright: 1998-10-01
• Publisher: CRC Pr I Llc

Summary

Fuzzy and Neuro-Fuzzy Systems in Medicineprovides a thorough review of state-of-the-art techniques and practices, defines and explains relevant problems, as well as provides solutions to these problems.After an introduction, the book progresses from one topic to another - with a linear development from fundamentals to applications.Chapters discuss:o a historical perspective of fuzzy systems technology and neuro-fuzzy systems technology in medicine and biologyo the relationship of fuzzy logic to the human braino analysis and classification of signals using fuzzy, neuro-fuzzy, and wavelet methodso wavelet analysis combined with neuro-fuzzy systems in contouring gated SPECT images of ventricleso a detailed application based on a knowledge-based system that uses fuzzy techniques, multispectral analysis, and image processing algorithmso applications in the field of dentistryo a dedicated system for myocardial ischemia diagnosiso a typical expert system used in intensive care unitso designing and tuning fuzzy rules for medical diagnosiso knowledge processing, decision-making, and control strategies combined with control methods in medical equipmento current technological problems and trends in the neural and fuzzy hardware implementation fieldThe well-balanced chapters cover all the major fields in medicine and biomedical engineering, including imaging, prosthetics, psychology, medical equipment, diagnosis, and treatment.

Table of Contents

Preface
A guide for using the volume in the classroom
Index of symbols
Contributors
Part 1. Fundamentals and Neuro-Fuzzy Signal Processing	3	(170)
Chapter 1. Fuzzy Logic and Neuro-Fuzzy Systems in Medicine and Bio-Medical Engineering. A Historical Perspective	3	(14)
Horia-Nicolai L. Teodorescu
Abraham Kandel
Lakhmi C. Jain
1. The first period: the infancy	3	(6)
2. Further developments and background	9	(2)
3. Neuro-fuzzy systems and their applications in medicine and biology	11	(2)
4. Genetic algorithms, fuzzy logic, and neuro-fuzzy systems	13	(1)
5. Bibliographies	13	(2)
6. Conclusions and predictions	15	(2)
Chapter 2. The Brain As A Fuzzy Machine: A Modeling Problem	17	(40)
Robert P. Erickson
Mircea I. Chelaru
Catalin V. Buhusi
1. The fuzzy approach in neurobiology: a historical perspective	18	(1)
2. The generality of Young's hypothesis	19	(6)
2.1. Simple stimuli	19	(3)
2.2. Neural organization of cryptic events: from tastes to faces	22	(1)
2.2.1. The general approach	22	(1)
2.2.2. Solutions possible: Taste	22	(1)
2.2.3. Solutions possible: Faces	23	(1)
2.3. Neural codes	24	(1)
3. Fuzzy models for taste	25	(15)
3.1. Grades of membership in fuzzy sets	25	(1)
3.2. A fuzzy model	26	(1)
3.2.1. The model	26	(6)
3.2.2. The synthesis of the fuzzy model	32	(1)
3.2.3. Simulating the dynamics of taste neurons	32	(8)
4. The brain as a fuzzy machine	40	(9)
4.1. A neural network implementing a fuzzy machine?	40	(1)
4.2. An artificial neuron implements a fuzzy membership function	41	(1)
4.3. A layer of neurons implements a fuzzifier	42	(3)
4.4. A "hidden" neuron implements a fuzzy rule	45	(4)
5. Applications of fuzzy logic to neural systems	49	(3)
5.1. Quantitative aspects of the fuzzy neural sets	49	(1)
5.1.1. Neural mass	49	(1)
5.1.2. Sensitivity to fine gradations in input	50	(1)
5.1.3. Intelligence	50	(1)
5.2. Defuzzification and responses	50	(1)
5.3. Memory: input and retrieval	51	(1)
6. Conclusions	52	(1)
Appendix 1. Abbreviations	53	(1)
Appendix 2. Terminology	53	(1)
References	54	(3)
Chapter 3 Brain State Identification and Forecasting of Acute Pathology Using Unsupervised Fuzzy Clustering Of EEG Temporal Patterns	57	(38)
Amir B. Geva
Dan H. Kerem
1. Introduction	57	(1)
2. Background	58	(9)
2.1. The electroencephalogram (EEG) signal	58	(1)
2.2. Brain states and the EEG	59	(3)
2.3. Stimulus-evoked EEG patterns	62	(3)
2.4. Underlying processes	65	(1)
2.5. Fuzzy systems and the EEG	65	(2)
3. Tools	67	(8)
3.1. Data acquisition	68	(1)
3.1.1. Spontaneous ongoing signal	68	(1)
3.1.2. Evoked responses	68	(1)
3.2. Feature extraction	69	(1)
3.2.1. Spectrum estimation	70	(1)
3.2.2. Time-frequency analysis	70	(1)
3.2.2.1. Multiscale decomposition by the fast wavelet transform	70	(1)
3.2.2.2. Multichannel model based decomposition by matching pursuit	71	(1)
3.3. The unsupervised optimal fuzzy clustering (UOFC) algorithm	71	(2)
3.4. The weighted fuzzy k-means (WFKM) algorithm	73	(1)
3.5. The clustering validity criteria	74	(1)
4. Examples of uses	75	(9)
4.1. Sleep-stage scoring	75	(2)
4.2. Forecasting epilepsy	77	(4)
4.3. Classifying evoked and event-related potentials by waveform	81	(3)
5. Concluding remarks and future applications	84	(2)
5.1. Dynamic version of state identification by UOFC	85	(1)
5.2. Data fusion	85	(1)
Appendix 1. The fast wavelet transform	86	(1)
Appendix 2. Multichannel model-based decomposition by matching pursuit	87	(1)
Appendix 3. Feature extraction and reduction by principal component analysis	88	(3)
List of acronyms	91	(1)
References	91	(4)
Chapter 4 Contouring Blood Pool Myocardial Gated SPECT Images With a Sequence of Three Techniques Based on Wavelets, Neural Networks, and Fuzzy Logic	95	(42)
Luis Patino
Andre Constantinesco
Ernest Hirsch
1. Introduction	95	(1)
2. Anatomy of the G-SPECT images	96	(3)
3. Strategy of the proposed method	99	(26)
3.1. Overview of the method	99	(1)
3.2. Wavelets-based image pre-processing	99	(1)
3.3. Neural network based image segmentation	100	(2)
3.4. Fuzzy logic based recognition of the regions of interest (ventricles)	102	(1)
3.4.1. Definition of the required fuzzy sentences	102	(5)
3.4.2. Combining neuronal approaches and fuzzy logic based inference systems	107	(7)
3.5. Training the recognition system using a neuro-fuzzy technique	114	(1)
3.5.1. Automated generation of rules and membership functions (ALGORAM)	114	(5)
3.5.2. Adjustment of membership functions using a descent method (FUNNY)	119	(3)
3.5.3. Combining the automated generation of rules and membership functions and the adjustment of their parameters in a parallel implementation (FUNNY-ALGORAM)	122	(3)
4. In vitro experiments and application to medical cases	125	(5)
4.1. Experiments with phantoms	125	(3)
4.2. Clinical test cases	128	(2)
4.3. Implementation issues	130	(1)
5. Conclusions	130	(2)
Appendix 1. Automatic determination of diastolic and systolic images	132	(1)
Appendix 2. Trust limits of the estimated regression coefficients	133	(1)
List of acronyms	134	(1)
References	134	(3)
Chapter 5 Unsupervised Brain Tumor Segmentation Using Knowledge-Based and Fuzzy Techniques	137	(36)
Matthew C. Clark
Lawrence O. Hall
Dimitry B. Goldgof
Robert Velthuizen
Reed Murtaugh
Martin S. Silbiger
1. Introduction	137	(2)
2. Domain background	139	(4)
2.1. Slices of interest for the study	139	(1)
2.2. Basic MR contrast principles	140	(1)
2.3. Knowledge-based systems	141	(1)
2.4. System overview	142	(1)
3. Classification stages	143	(14)
3.1. Stage zero: pathology detection	143	(1)
3.2. Stage one: building the intra-cranial mask	144	(2)
3.3. Stage two: multi-spectral histogram thresholding	146	(2)
3.4. Stage three: "Density screening" in feature space	148	(2)
3.5. Stage four: region analysis and labeling	150	(1)
3.5.1. Removing meningial regions	150	(1)
3.5.2. Removing non-tumor regions	151	(3)
3.6. Stage five: final T1 threshold	154	(3)
4. Results	157	(5)
4.1. Knowledge-based vs. supervised methods	160	(2)
4.2. Evaluation over repeat scans	162	(1)
5. Discussion	162	(4)
Abbreviations	166	(1)
References	166	(7)
Part 2. Neuro-Fuzzy Knowledge Processing	173	(146)
Chapter 6 An Identification of Handling Uncertainties Within Medical Screening: A Case Study Within Screening for Breast Cancer	173	(22)
Fredrik Georgsson
Patrik Eklund
1. Introduction	173	(1)
2. Screening	174	(4)
2.1. Notations	174	(2)
2.2. The screening program	176	(1)
2.3. The methods	177	(1)
3. The select function	178	(4)
3.1. The decision step	179	(1)
3.2. Disease specific knowledge	180	(2)
4. A breast cancer case study	182	(9)
4.1. Minimizing A(0) as much as possible in one step	182	(2)
4.2. Finding the screening method	184	(2)
4.3. Defining disease specific knowledge	186	(2)
4.4. Performing the refinement	188	(1)
4.5. The final system	189	(2)
5. Conclusions and further work	191	(1)
References	191	(4)
Chapter 7 A Fuzzy System for Dental Developmental Age Evaluation	195	(16)
Masao Ozaki
1. Introduction	195	(1)
2. Technical consideration	196	(6)
2.1. Basic conception of the teeth evaluation system	196	(3)
2.2. Rule evaluation module	199	(3)
3. System optimization by using clinical data	202	(5)
3.1. Material and method	202	(1)
3.2. The dimensionality analysis by principal component analysis	203	(1)
3.3. System optimization by using genetic algorithm	204	(1)
3.4. System evaluation and results	205	(2)
4. Discussion and conclusions	207	(1)
References	208	(3)
Chapter 8 Fuzzy Expert System for Myocardial Ischemia Diagnosis	211	(32)
Sorina Zahan
Christian Michael
Stephanos Nikolakeas
1. Introduction	211	(1)
2. Fuzzy expert systems	212	(3)
3. DIFUS: hierarchical diagnosis fuzzy system	215	(10)
3.1. Characteristics	215	(1)
3.2. Knowledge organization	216	(2)
3.3. Structure	218	(2)
3.4. Operation	220	(5)
4. Multimethod myocardial ischemia diagnosis	225	(1)
5. Multimethod myocardial ischemia diagnosis system	226	(11)
5.1. The implementation of fuzzy score-based tests	226	(1)
5.1.1. Medical patterns	227	(1)
5.1.2. Sequential processes	228	(1)
5.1.3. Compact representation of fuzzy score-based tests	229	(2)
5.2. MMIDS structure and operation	231	(1)
5.2.1. MMIDS secondary group	231	(2)
5.2.2. MMIDS primary groups	233	(2)
5.3. Experimental results	235	(2)
6. Conclusions	237	(2)
References	239	(4)
Chapter 9 Design and Tuning of Fuzzy Rule-Based Systems for Medical Diagnosis	243	(48)
Alexander Rotshtein
1. Introduction	243	(1)
2. Problem statement and general methodology	244	(3)
3. Design and rough tuning of fuzzy rules	247	(8)
3.1. Matrix of knowledge	247	(1)
3.2. Fuzzy model with discrete output	248	(2)
3.3. Fuzzy model with continuous output	250	(1)
3.4. Rough tuning of fuzzy rules	251	(1)
3.4.1. Rough tuning of membership functions	251	(4)
3.4.2. Rough tuning of rules weights	255	(1)
4. Fine tuning of fuzzy rules with continuous output	255	(11)
4.1. Tuning as a problem of optimization	255	(3)
4.2. Quality evaluation of fuzzy inference	258	(1)
4.3. Computer simulation	258	(1)
4.3.1. Experiment 1	258	(4)
4.3.2. Experiment 2	262	(4)
5. Fine tuning of fuzzy rules with discrete output	266	(8)
5.1. Tuning as a problem of optimization	266	(1)
5.2. Quality evaluation of fuzzy inference	267	(1)
5.3. Computer simulation	268	(6)
6. Application to differential diagnosis of ischemia heart disease	274	(11)
6.1. Diagnosis types and parameters of patient's state	274	(1)
6.2. Fuzzy rules	275	(1)
6.3. Fuzzy logic equation	275	(4)
6.4. Rough membership functions	279	(1)
6.5. Algorithm of decision making	280	(1)
6.6. Fine tuning of fuzzy rules	281	(4)
7. Conclusion	285	(1)
References	285	(1)
Appendix 1. Comparison of real and inferred decisions for 65 patients	286	(2)
Appendix 2. Fuzzy expert shell and its application	288	(3)
Chapter 10 Integration of Medical Knowledge in an Expert System for Use in Intensive Care Medicine	291	(28)
Uwe Pilz
Lothar Engelmann
1. Introduction	291	(1)
2. Software design principles	292	(1)
3. Medical knowledge in intensive care medicine	293	(7)
3.1. Structure of the knowledge	293	(1)
3.2. Meaning of colloquial rules	294	(1)
3.3. Rule processing and result calculation	295	(3)
3.4. Combining different rules	298	(2)
4. Transformation of knowledge into FLORIDA commands	300	(6)
4.1. Introduction	300	(1)
4.2. Comments	300	(1)
4.3. Modules	300	(1)
4.4. Linguistic variables	301	(1)
4.5. Fuzzy variables culator	302	(1)
4.6. Rules -- the knowledge itself	303	(2)
4.7. Changing the normal value	305	(1)
5. Invocation of FLORIDA	306	(1)
6. Explaining more of FLORIDA's functionality -- the knowledge base inflammation	306	(6)
6.1. Structuring the knowledge	306	(1)
6.2. Rules for fever	307	(1)
6.3. Rules for leukocytosis/leukopenia	308	(1)
6.4. Rules for tachycardia/tachypnoe	309	(1)
6.5. Rules for synthesis of acute phase proteins	309	(1)
6.6. Rules for consumption of coagulation components	310	(1)
6.7. Improvement of explanation	311	(1)
7. Differentiation of dysfunctions	312	(1)
8. Visualization of the result	313	(1)
9. Discussion and conclusions	314	(1)
References	314	(5)
Part 3. Neuro-Fuzzy Control and Hardware	319	(72)
Chapter 11 Fuzzy Control and Decision Making in Drug Delivery	319	(22)
Johnnie W. Huang
Claudio M. Held
Rob J. Roy
1. Introduction	319	(2)
1.1. Progress in decision making	320	(1)
1.2. Progress in control	321	(1)
2. System development	321	(14)
2.1. Decision-making: fuzzy decision-making module (FDMM)	324	(1)
2.1.1. Purpose	324	(2)
2.1.2. Operation	326	(2)
2.2. Drug-titration control: fuzzy hemodynamic control module (FHCM)	328	(1)
2.2.1. Purpose	328	(1)
2.2.2. Operation	328	(2)
2.3. Supervisory commands: therapeutic assessment module (TAM)	330	(2)
2.4. System evaluation	332	(1)
2.4.1. Example one	332	(1)
2.4.2. Example two	332	(3)
3. Future prospects	335	(1)
3.1. Design possibilities	335	(1)
3.2. "Curse of dimensions"	335	(1)
3.3. Machine intelligence	336	(1)
Appendix. Additional resources	336	(1)
Appendix. Terminology	337	(2)
References	339	(2)
Chapter 12. Neuro-Fuzzy Hardware in Medical Applications	341	(50)
12 A. System Requirements for Fuzzy and Neuro-Fuzzy Hardware in Medical Equipment	341	(20)
Horia-Nicolai Teodorescu
Abraham Kandel
Daniel Mlynek
1. Introduction	341	(1)
2. Specific requirements of medical applications	342	(1)
2.1. General system and technological requirements	343	(1)
2.2. Reliability requirements	344	(1)
2.3. Precision and sensitivity to parameters	345	(1)
3. Analysis of several applications	346	(1)
3.1. Life-support applications	346	(1)
3.1.1. Artificial heart control	346	(1)
3.1.2. Assisted ventilation	346	(1)
3.2. Anesthesia related equipment	347	(1)
3.3. Fuzzy and neuro-fuzzy-based equipment for prosthetics	347	(1)
3.4. General purpose devices	348	(1)
3.5. Other applications	348	(1)
4. General system design issues	349	(1)
4.1. Nonlinearity implementation -- simulation power	349	(1)
4.2. Dynamical errors	350	(1)
5. Hardware implementation issues	351	(1)
5.1. Implementation choice: analog vs. digital fuzzy processors	351	(1)
5.2. Hardware minimization	351	(1)
5.3. Parallelism vs. number of rule blocks	352	(3)
5.4. A minimal system design	355	(1)
6. Choosing the right design	356	(1)
7. Conclusions	356	(1)
References	356	(5)
12. B. Neural Networks and Fuzzy-Based Integrated Circuit and System Solutions Applied to the Biomedical Field	361	(30)
Alexandre Schmid
Daniel Mlynek
1. Introduction	361	(2)
2. Required properties for embedded medical systems	363	(1)
2.1. Embedding medical systems	363	(1)
2.2. Autonomy	364	(2)
2.3. Reliability -- safety	366	(5)
2.4. Precision of computation	371	(1)
2.5. Application specific requirements	371	(1)
3. Architectures applied to neuro-fuzzy IC design	371	(1)
3.1. Artificial neural network integrated realization	372	(8)
3.2. Fuzzy-based integrated realization	380	(4)
3.3. Hybrid integrated realization	384	(1)
3.4. An example of neuro-fuzzy realization	384	(1)
4. Concluding remarks	385	(1)
References	385	(6)
Index of Terms	391