Biomedical Data Acquisition and Signal Processing Lecture 3 PDF
Document Details
AUTH
2024
Ilias K. Kitsas, Ph.D., E&CE, AUTh
Tags
Related
- Filtri Digitali PDF - Elaborazione Dati e Segnali Biomedici II - 2023-24
- Biomedical Signals and Signal Processing PDF
- Myoelectric Interfaces and Related Applications: Current State of EMG Signal Processing PDF
- Digital Signal Processing DSP PDF
- Biomedical Data Acquisition and Signal Processing Lecture Notes PDF
- Biomedical Data Acquisition and Signal Processing Lecture Notes PDF
Summary
These lecture notes cover biomedical data acquisition techniques, including devices, sensors, and 1D/2D signal registration. The lecture also reviews biomedical signal processing concepts such as time/frequency domain representation, Fourier series, A/D conversion, Nyquist theorem, and more.
Full Transcript
Biomedical data acquisition and signal processing Lecture 3-Data acquisition. Introduction to data acquisition, devices, sensors, registration of 1D/2D signals Ilias K. Kitsas, Ph.D., E&CE, AUTh MSc in BioMedical Engineering (BME-AUTh) PREVIOUSLY … Biom...
Biomedical data acquisition and signal processing Lecture 3-Data acquisition. Introduction to data acquisition, devices, sensors, registration of 1D/2D signals Ilias K. Kitsas, Ph.D., E&CE, AUTh MSc in BioMedical Engineering (BME-AUTh) PREVIOUSLY … Biomedical Signal Processing Time/Frequency domain signal representation Fourier series A/D conversion Nyquist theorem Discrete time systems & Convolution Periodic sampling DTFT Biomedical Instrumentation Amplifiers & Sensors MSc in BioMedical Engineering (BME-AUTh) ELECTROCARDIOGRAM - ECG P wave: depolarization of SA node QRS complex: ventricular depolarization T wave: ventricular repolarization The atrium repolarization is masked by the QRS complex 3 MSc in BioMedical Engineering (BME-AUTh) ELECTRICAL ACTIVITY IN MYOCARDIUM 4 MSc in BioMedical Engineering (BME-AUTh) ANATOMY OF THE HEART Anatomy of the Heart 5 MSc in BioMedical Engineering (BME-AUTh) ECG – Ischemia 6 MSc in BioMedical Engineering (BME-AUTh) ECG EXAMPLES 7 MSc in BioMedical Engineering (BME-AUTh) ECG recording Bipolar leads 1.Scalar ECG: 2 electrodes + ground [Right arm(RA), Left arm(LA) + GND Right Leg(RL)] 2.Vector ECG: ◼ 3 electrodes + ground (RA, LA, LL, GND at RL) ◼ 4 electrodes + ground (RA, LA, LL, GND at RL, on chest) ◼ Precordial leads (V) (+ esophageal) horizontal planar projection MSc in BioMedical Engineering (BME-AUTh) ◼8 ECG RECORDING - Eindhoven’s Triangle 1903! MSc in BioMedical Engineering (BME-AUTh) Augmented ECG leads – Eindhoven’s Triangle 1942 MSc in BioMedical Engineering (BME-AUTh) Recording ECG-12 leads Precardiac leads MSc in BioMedical Engineering (BME-AUTh) ECG as a vector Electric signal moving towards specific direction Towards positive pole: positive deflection Away from the positive pole: negative deflection Perpendicular to the positive pole: small / no deflection MSc in BioMedical Engineering (BME-AUTh) NORMAL ECG – ECG PAPER MSc in BioMedical Engineering (BME-AUTh) NORMAL ECG – ECG PAPER MSc in BioMedical Engineering (BME-AUTh) ◼14 BIOMEDICAL SIGNALS’ FEATURES MSc in BioMedical Engineering (BME-AUTh) ◼16 ECG RECORDER Electrodes Protection Lead Pre- Isolation Final Recorder circuit selector amplifier circuit amplifier Amplifying Selection Very high Patient protection buffers’ switch CMRR against power Annotating protection supply system MSc in BioMedical Engineering (BME-AUTh) ECG RECORDER MSc in BioMedical Engineering (BME-AUTh) ◼18 ECG amplifiers ECG: low typical values Signal at the body surface is small portion of the source (tissue attenuation) Interfering voltages from other sources near the ECG recording site Information extraction/isolation ECG is a differential signal Noise: almost steady power over the body Differential amplifiers with high Common Mode Rejection Ratio (CMRR) MSc in BioMedical Engineering (BME-AUTh) ECG RECORDING NOISE Types of noise baseline wander muscle artifact (EMG) electrode motion artifact Electrodes’ movement is the main source of noise (resembles the heart beats and cannot be easily rejected by filtering as in other cases of noise) Examples MSc in BioMedical Engineering (BME-AUTh) Pred. baseline wander MSc in BioMedical Engineering (BME-AUTh) Muscle artifact (EMG) MSc in BioMedical Engineering (BME-AUTh) Electrode motion artifact MSc in BioMedical Engineering (BME-AUTh) ECG RECORDING NOISE Electrode displacement SNR = -6dB SNR = 0dB SNR = 6dB SNR = 18dB MSc in BioMedical Engineering (BME-AUTh) ECG ELECTRODES MSc in BioMedical Engineering (BME-AUTh) ECG RECORDERS MSc in BioMedical Engineering (BME-AUTh) ◼26 PORTABLE PERSPECTIVES http://cardiacdesigns.com/ ECG Check MSc in BioMedical Engineering (BME-AUTh) ◼27 DEFINITIONS ◼ Transducer a device which converts energy from one form to another ◼ Sensor a device which converts a physical parameter to an electrical output ◼ Actuator a device which converts electrical energy to a mechanical or physical output 28 MSc in BioMedical Engineering (BME-AUTh) SENSORS ◼ Active sensors generate electrical power directly in response to an applied stimulation or measurand do not require an external voltage source to produce electrical output Solar cell, piezoelectric material, thermocouple, etc ◼ Passive sensors produce a change in some passive electrical quantity, such as capacitance, resistance or inductance in response to an applied stimulus or measurand require an external ac or dc voltage source to convert passive electrical quantity to electrical output Photodiode, thermistor, strain gauge, etc 29 MSc in BioMedical Engineering (BME-AUTh) ACTIVE VS PASSIVE TRANSDUCERS Parameter Active Transducer Passive Transducer The type of transducer which The type of transducer that does not require external requires an additional Basic description power supply to produce the source of power to work is output signal is known as an known as a passive active transducer. transducer. Active transducer draws the energy required to work Passive transducer draws Working energy directly from the physical energy to work from an quantity which is to be external source of power. measured. The design and Active transducers have simple Design construction of passive design and construction. transducer is complex. MSc in BioMedical Engineering (BME-AUTh) ACTIVE VS PASSIVE TRANSDUCERS Parameter Active Transducer Passive Transducer Passive transducer Active transducer naturally produces an produces an output Need of output output signal of high signal of very low amplification amplitude, hence it does amplitude. Therefore, it not require amplification of requires amplification. output signal. Passive transducer Active transducer produces its output as the generally produces its change in passive circuit Output parameters such as output in the form of voltage or current. resistance, inductance, and capacitance. MSc in BioMedical Engineering (BME-AUTh) FIBEROPTIC BIOSENSOR MSc in BioMedical Engineering (BME- MSc in BioMedical Engineering AUTh) (BME-AUTh) BIOMEDICAL INSTRUMENTS’ SENSORS MSc in BioMedical Engineering (BME- MSc in BioMedical Engineering AUTh) (BME-AUTh) BIOMEDICAL INSTRUMENTS’ SENSORS MSc in BioMedical Engineering (BME- MSc in BioMedical Engineering AUTh) (BME-AUTh) SKELETAL MUSCLE ANATOMY Thousand of myofibrils Grouped in Muscle fibers Every myofibril comprises 100 to 10.000 myofilaments MSc in BioMedical Engineering (BME-AUTh) SKELETAL MUSCLE ANATOMY ◼ Every myofilament consists of actin and myosin filaments (motor proteins) MSc in BioMedical Engineering (BME-AUTh) ELECTROMYOGRAM - EMG EMG is the total signal acquired by a single electrode It is the spatial and temporal sum of all active motor units in the area of the electrodes 37 MSc in BioMedical Engineering (BME-AUTh) EMG FEATURES ◼ Amplitude: 2μV-5mV (100μV-1mV surface) ◼ Frequency band: 10-1000 Hz (surface), 20-2000 Ηz (needle) ◼ Noise: 50 Hz, cable artifacts, electrode movement, high frequencies 38 MSc in BioMedical Engineering (BME-AUTh) RECORDING ELECTRODES ◼ Monopolar ◼ Single electrode with reference ◼ Simple amplification ◼ Bipolar ◼ Two electrodes with reference ◼ Differential amplifier 39 MSc in BioMedical Engineering (BME-AUTh) RECORDING ELECTRODES + AMPLIFIER ◼ Monopolar ◼ Single electrode with reference ◼ Simple amplification ◼ Bipolar ◼ Two electrodes with reference ◼ Differential amplifier 40 MSc in BioMedical Engineering (BME-AUTh) EMG RECORDING ELECTRODES Surface Needle Features Features Features Features 41 MSc in BioMedical Engineering (BME-AUTh) SURFACE Advantages Applied easily and fast No medical surveillance needed Minimum discomfort Disadvantages Applied on large muscles Cross-talk interference No typical way of placement Motion sensitive Recording of dynamic muscle activity 42 MSc in BioMedical Engineering (BME-AUTh) SURFACE RECORDING ELECTRODES 43 MSc in BioMedical Engineering (BME-AUTh) NEEDLE Advantages Extremely sensitive Recording of single muscle activity Access to internal muscle structure No cross-talk Disadvantages Extremely sensitive Qualified personnel needed Impossible to reposition Area of recording may not be the actual area of interest Pain and discomfort MSc in BioMedical Engineering (BME-AUTh) 44 EMG ELECTRODE PLACEMENT 45 MSc in BioMedical Engineering (BME-AUTh) LABORATORY SET-UP 46 MSc in BioMedical Engineering (BME-AUTh) LABORATORY SET-UP Negative expression Positive expression 47 MSc in BioMedical Engineering (BME-AUTh) ΕΜG RECORDING PROCESS 48 MSc in BioMedical Engineering (BME-AUTh) ΕΜG SIGNAL PROCESSING … 49 MSc in BioMedical Engineering (BME-AUTh) ΕΜG SIGNAL PROCESSING … Denoising via wavelet packet transform 50 MSc in BioMedical Engineering (BME-AUTh) ΕΜG SIGNAL PROCESSING-MOTION ANALYSIS (gait cycle) Terminal Heel Foot Opposite Heel Opposite Toe Swing Swing Strike Flat Toe-off Rise Heel Strike Off Phase 51 MSc in BioMedical Engineering (BME-AUTh) ΕΜG SIGNAL PROCESSING-MOTION ANALYSIS (gait cycle) EMG receiver A/D Converter Computer Force Platforms Cameras OR Optical Tracking System 52 MSc in BioMedical Engineering (BME-AUTh) ΕΜG SIGNAL PROCESSING-MOTION ANALYSIS (gait cycle) LEFT RIGHT walking walking RF RF HS HS AT AT G/S G/S 0 25 50 TO 75 100 0 25 50 TO 75 100 % stride % stride 53 MSc in BioMedical Engineering (BME-AUTh) EMG RECORDING for the evaluation of Orthopedic and neurological disorders Pharmacological treatments Evolution of motor deficits Use of orthotics Rehabilitation follow-up Athletic task optimization MSc in BioMedical Engineering (BME-AUTh) COMMERCIAL PRODUCTS Blue Sensor Electrodes www.ambuUSA.com MyoSystem 2000 www.noraxon.com FREEEMG https://www.btsbioengineering.com https://www.biopac.com/ 55 MSc in BioMedical Engineering (BME-AUTh) INTERMEDIATE MSc in BioMedical Engineering (BME-AUTh) 56 AXIAL SECTION of the brain Frontal Gyri grey matter Sulci White matter Parental MSc in BioMedical Engineering (BME-AUTh) EEG ORIGIN Tangential dipoles Axial dipoles (gyri) (grooves) Electric Field Magnetic Field MSc in BioMedical Engineering (BME-AUTh) EEG ORIGIN EEG is the sum of field potentials produced by the currents of pyramoid neural cells MSc in BioMedical Engineering (BME-AUTh) 59 EEG rhythms - α, β, θ, δ EEG: superposition of other simpler waveforms One or two dominant rhythms (normal) Other rhythms dominate during sleep (normal case) or in pathological states Main rhythms of different frequency and amplitude MSc in BioMedical Engineering (BME-AUTh) 60 SLEEP STAGES MSc in BioMedical Engineering (BME-AUTh) EEG RECORDING Electrode placement ❖10/20 system ❖Hemisphere symmetry Recording technique ❖Monopolar ❖Biopolar MSc in BioMedical Engineering (BME-AUTh) 62 EEG RECORDING (Medical & electrical engineering, Guger Technologies, www.gtec.at): 4 EEG bipolar channels, passive electrodes, Filters: 0.5–30 Hz, Sensitivity: 100 μV, Data transfer: wireless, Bluetooth ‘Class I’ technology EPOC (www.emotiv.com): Channels: 14 Sampling: 128 Hz Analysis: 16 bits Bandwith: 0.2 - 45 Hz Wireless connection to PC via Bluetooth MSc in BioMedical Engineering (BME-AUTh) EEG RECORDING EPOC (www.emotiv.com): Channels: 14 Sampling: 128 Hz Analysis: 16 bits Bandwith: 0.2 - 45 Hz Wireless connection to PC MSc in BioMedical Engineering (BME-AUTh) via Bluetooth EEG RECORDING g.tec Emotiv EPOC MSc in BioMedical Engineering (BME-AUTh) EEG RECORDING Dense EEG (EGI) Channels: 256 (http://www.atesdevice.it/) MSc in BioMedical Engineering (BME-AUTh) QEEG MSc in BioMedical Engineering (BME-AUTh) QEEG MSc in BioMedical Engineering (BME-AUTh) QEEG Without 2 hours medication after using addictive substances MSc in BioMedical Engineering (BME-AUTh) QEEG EEG rhythms in children and adolescents (closed eyes) MSc in BioMedical Engineering (BME-AUTh) Medical Imaging Use of image (analog or digital) Images are subject to general laws regardless of the recovery medium ❖ finite detail ❖ limited dynamic range of illustration regarding amplitude of information ❖ noise MSc in BioMedical Engineering (BME-AUTh) 71 MAIN CONCEPTS: IMAGE INFORMATION CONTENT Total image information = [number of image’s picture elements (pixels)] x [number of quantization levels for each pixel] (usually equals SNR) MSc in BioMedical Engineering (BME-AUTh) 72 SPATIAL FREQUENCY Photocell scanning single line of image Bright object → Positive output Dark object → Negative output 1 period Uniform succession of bright objects and dark spaces ◼ Object: positive half period of output signal ◼ Gap: negative half period of output signal Identifying an object requires a whole period The cycle rate between brightness transitions is known as the spatial frequency of the image. [units of cycles per mm or line pairs per mm] MSc in BioMedical Engineering (BME-AUTh) 73 IMAGE RESOLUTION 1cm 10 holes / cm 100 holes / cm2 Resolution: 1 lp/mm or 2 pixels/mm 1cm Identification of both hole and the area between them MSc in BioMedical Engineering (BME-AUTh) 74 PIXEL Basic element of image The area on an image surface with half period of the maximum spatial frequency along the horizontal axis and the same dimension along the vertical axis. 1 period The element with the smallest size that we need to identify MSc in BioMedical Engineering (BME-AUTh) 75 CONTRAST/MODULATION How faithfully the minimum and maximum brightness values are transferred from the input to the output of an imaging system. It depends on the spatial frequency MSc in BioMedical Engineering (BME-AUTh) 76 MAIN CONCEPTS: IONIZING/NON INONIZING RADIATION Non-ionizing is low-frequency radiation, which includes radio, television, communications (mobile-fixed), microwaves (microwave ovens), and even low-frequency ultraviolet light. MSc in BioMedical Engineering (BME-AUTh) 77 MAIN TYPES OF IMAGING METHODS Ionizing radiation I ❖Classic X-ray ❖X-ray ❖Angiography ❖Computed Tomography Ionizing radiation II ❖Radioisotopes Ultrasound Magnetic resonance Functional MRI Audiovisual MSc in BioMedical Engineering (BME-AUTh) 78 CLASSIC X-ray X-ray: ionizing radiation ❖has a very short wavelength - it interacts with atoms and molecules - it creates ions ❖ion production leads to uncontrollable chemical reactions Main parts ❖High voltage generator - X-ray tube (on device) ❖beam guidance system ❖auxiliary grid ❖reinforcing plates - film MSc in BioMedical Engineering (BME-AUTh) 79 X-Ray PRODUCTION X-ray photons have a wide distribution of energies Only part of the radiation exits the tube due to strong absorption by the tube materials X-ray exposure unit: roentgen (R) = ionizing radiation of any sign 2.58x10-4 C / Kg in dry air Unit of measurement of absorbed dose: rad = 0.01 J / Kg Patient absorbs 95% - 99% of radiation from the X-ray beam, then R and rad are approximately equal and are used in place of each other MSc in BioMedical Engineering (BME-AUTh) 80 X-Ray DETECTION A percentage of the radiation is scattered inside the patient (secondary radiation) Most of it is absorbed 1% - 5% of the radiation reaches the detector Low energy radiation is absorbed by the tube Aluminum filter additionally dampens low-energy X-rays that are not able to pass through the body and only increase patient exposure The grid absorbs most of the secondary radiation MSc in BioMedical Engineering (BME-AUTh) 81 X-Ray DETECTION The radiation reaches the detector (film) which, however, is largely transparent to the radiation Reinforcement plates are placed in front / behind the film to improve X-ray detection They consist of plastic sheets coated with high atomic number Z phosphors These sheets increase sensitivity and reduce patient exposure 20-100 times MSc in BioMedical Engineering (BME-AUTh) 82 X-Ray DETECTION The performance of the reinforcement plates increases depending on their thickness Their resolution decreases with increasing thickness Reinforcing plates> 300 μm: very high sensitivity and low resolution The use of high-resolution plates and films causes unnecessary exposure of the patient to radiation Most organs in the human body have low resolution requirements MSc in BioMedical Engineering (BME-AUTh) 83 X-Ray DETECTION MSc in BioMedical Engineering (BME-AUTh) 84 X-ray Real-time function monitoring Requires continuous irradiation, but low dose The use of a contrast medium is usually required ❖swallowing ❖promotion of contrast in vases ❖heart operation Local increase of radiation absorption in cavities where it is traced The images are displayed on a video screen MSc in BioMedical Engineering (BME-AUTh) 85 X-ray Angiography Small Bowel Follow-Through Pyelography MSc in BioMedical Engineering (BME-AUTh) 86 X-ray Tube and camera on arm that allows all inclinations and angles No change in the patient's position is required Image on screen Image amplifier ◼ Digital image processing MSc in BioMedical Engineering (BME-AUTh) 87 THERMOGRAPHY Body temperature varies as a function of blood circulation the metabolism local biological formations local thermal conductivity ambient temperature skin moisture MSc in BioMedical Engineering (BME-AUTh) 88 THERMOGRAPHY It is used to detect many diseases diseases of the joints inflammatory diseases (eg sinusitis, rheumatoid arthritis) sports injuries carpal tunnel syndrome breast cancer MSc in BioMedical Engineering (BME-AUTh) 89 NUCLEAR MEDICINE Nuclear medicine is the branch of medicine that uses radioactive materials for diagnostic purposes. Difference from X-ray: the source of Gamma rays is not external but inside the patient radioactivity can be selectively added in a programmable and controlled manner to materials that are biochemically active in the patient Basic principle: Measurement of the spatial distribution of radioactivity in the patient's body MSc in BioMedical Engineering (BME-AUTh) 90 Gamma CAMERA The Gamma Camera has established itself as the best tool in nuclear medicine (Hal Anger, 1957) Imaging system, capable of spatial resolution in a large field of view, without the need to move the detectors The radiation detector is a sodium iodide (NaI ) crystal 30 to 40 cm in diameter and 1.2 cm thick. MSc in BioMedical Engineering (BME-AUTh) 91 Gamma CAMERA The basic principle of operation of the Gamma camera is that the relative fraction of the total light seen by each photomultiplier uniquely defines the position of the initial point of entry of the Gamma ray. MSc in BioMedical Engineering (BME-AUTh) 92 Gamma CAMERA Photomultipliers in a grid Gamma ray enters the crystal and the light resulting from the sparkle is diffused Each photomultiplier receives a fraction of the total light The output voltages correspond to the x, y coordinates and are reconstructed by the signal of each photomultiplier MSc in BioMedical Engineering (BME-AUTh) 93 Gamma CAMERA Double camera Rotating SPECT camera single photon emission computed tomography MSc in BioMedical Engineering (BME-AUTh) 94 Gamma CAMERA normal bone scintigraphy pathological - metastases MSc in BioMedical Engineering (BME-AUTh) 95 NUCLEAR MEDICINE - PET MSc in BioMedical Engineering (BME-AUTh) 96 ULTRASOUND SCAN Human tissues and organs are good conductors of sound waves The absorption properties are such that a "safe" sound wave is detectable even if it has crossed 20-30 cm of tissue Quite a high frequency sound wave is a "catheter" of the internal structure of the human body MSc in BioMedical Engineering (BME-AUTh) 97 ULTRASOUND SCAN The size of the structure that can be examined is determined by the wavelength of the sound wave wavelength = velocity of propagation in the medium / frequency Structures of the order of 1mm can be imaged by sound waves with frequencies of the order of MHz (ultrasound) A sound wave that crosses the boundary between two regions with different densities is reflected and / or refracted MSc in BioMedical Engineering (BME-AUTh) 98 ULTRASOUND SCAN Typical diagram of ultrasound piezoelectric transducer Mains Connector Power supply Acoustic shielding layer Depreciation material Outer case Piezoelectric element Electrodes Acoustic adaptation and protection material MSc in BioMedical Engineering (BME-AUTh) ULTRASOUND SCAN A piezoelectric transducer is placed in contact with the patient's skin The transducer is excited by means of sequences of pulses resulting in its vibration for a short time at its resonant frequency Sound radiation has a near-field area with complex properties (not used) and a narrowly defined long-range area (beam) with simple properties (main imaging tool) When the beam crosses the boundary between two regions with different acoustic properties this boundary reflects a small fraction of the incident beam further detected by the transducer. MSc in BioMedical Engineering (BME-AUTh) 100 ULTRASOUND SCAN – Α mode Amplitude mode (A mode) Considering the transducer at fixed position, the distribution of the reflected intensity (as a function of time) is a map of the discontinuity of the tissues along the scan axis. Accurate distance measurement Illustration as peaks in relation to a baseline Used in ophthalmology to measure corneal thickness 1. Ultrasound head 4. Reflected echo 2. Emitted beam 5. Registration 3. Different surfaces 101 MSc in BioMedical Engineering (BME-AUTh) REAL TIME ULTRASOUND SCAN Extension of A-mode The dimension of depth is the vertical axis The amplitude of the reflected pulse modulates the brightness Time evolves on the horizontal axis A-mode for a given moment MSc in BioMedical Engineering (BME-AUTh) 102 B-MODE ULTRASOUND SCAN Brightness mode A 2-D image can be produced by combining A- mode information from many different directions The amplitude of the reflected signal (echo) modulates the brightness of the oscilloscope dot The point of the dot is determined by the time delay of receiving the echo, hence the position of the structure Moving the scan bar creates the image MSc in BioMedical Engineering (BME-AUTh) 103 ULTRASOUND SCAN MSc in BioMedical Engineering (BME-AUTh) 104 ULTRASOUND TRIPLEX The techniques described so far are pulsating Another technique uses a continuous sound wave A wave-sensitive Doppler system detects blood velocity Measurement of the frequency difference between the transmitted and the reflected signal MSc in BioMedical Engineering (BME-AUTh) 105 3D ULTRASOUND SCAN Sound waves are sent to many angles and not just vertically The return echo is processed by an advanced program on a computer resulting in a reconstructed 3D volume It is mainly used in pregnancy and for breast ultrasound scanning MSc in BioMedical Engineering (BME-AUTh) 106 TOMOGRAPHY We are interested in the situation within the organizations but without having direct access (with invasive procedures) By properly stimulating the internal system from the external surface we can get a response that depends on the internal structure By properly combining the various responses, an interior map can be constructed Tomography: 3D Mapping the interior by externally stimulating and properly combining the various responses Mathematical technique for creating the map: Reconstruction of an image from its projections MSc in BioMedical Engineering (BME-AUTh) TOMOGRAPHY Depending on the type of stimulation we have: X-ray tomography Magnetic resonance tomography Ultrasound tomography Other imaging techniques produce a cumulative view of the entire body area E.g. X-ray: Does not show an isolated section, but the sum of all perpendiculars in the direction of the radial planes Tomography offers the possibility of separating the third dimension MSc in BioMedical Engineering (BME-AUTh) IMAGE RECONSTRUCTION BY ITS PROJECTION Based on the cumulative response of the system in various directions Image to identify Cumulative weight in horizontal and vertical Other candidates direction By increasing the number of independent measurements (projections) an ideally accurate and unique reconstruction of the object can be reached MSc in BioMedical Engineering (BME-AUTh) IMAGE RECONSTRUCTION BY ITS PROJECTION Projections are Single projection essentially equations (for 0o) The equations must be Sensors linearly independent of each other In the previous Beam example we have 16 unknowns We need 16 equations. That is, 4 different projections Source MSc in BioMedical Engineering (BME-AUTh) IMAGE RECONSTRUCTION BY ITS PROJECTION Projections are essentially equations The equations must be linearly independent of each other In the previous example we have 16 unknowns We need 16 equations. That is, 4 different projections MSc in BioMedical Engineering (BME-AUTh) IMAGE RECONSTRUCTION BY ITS PROJECTION Projections are essentially equations The equations must be linearly independent of each other In the previous example we have 16 unknowns We need 16 equations. That is, 4 different projections MSc in BioMedical Engineering (BME-AUTh) IMAGE RECONSTRUCTION BY ITS PROJECTION Projections are essentially equations The equations must be linearly independent of each other In the previous example we have 16 unknowns We need 16 equations. That is, 4 different projections MSc in BioMedical Engineering (BME-AUTh) IMAGE RECONSTRUCTION BY ITS PROJECTION The example is an ideal case In practice, the function to be reconstructed has no distinct form Therefore an infinite number of projections are needed to determine it MSc in BioMedical Engineering (BME-AUTh) 1D FT Function – signal of one variable: f(t) (complex) sinusoidal superposition : 𝑒 𝑗𝜔𝑡 1 𝑓 𝑡 = න 𝐹 𝜔 𝑒 𝑗𝜔𝑡 𝑑𝜔 2𝜋 where: 𝐹 𝜔 = න 𝑓 𝑡 𝑒 −𝑗𝜔𝑡 𝑑𝑡 (Temporal) frequency = rate Sampling theorem: 1 ΔΤ𝑚𝑖𝑛 = 2𝑓𝑚𝑎𝑥 MSc in BioMedical Engineering (BME-AUTh) 1D FT MSc in BioMedical Engineering (BME-AUTh) 2D FT Function – signal of two variables: f(x,y) (complex) two dimensional sinusoidal superposition: 𝑒 𝑗(𝑥𝜔𝑥+𝑦𝜔𝑦) 1 𝑓 𝑥, 𝑦 = 𝐹 𝜔𝑥 , 𝜔𝑦 𝑒 𝑗𝑥𝜔𝑥 𝑒 𝑗𝑦𝜔𝑦 𝑑𝜔𝑥 𝑑𝜔𝑦 όπου: 4𝜋2 𝑓 𝑥, 𝑦 𝐹 𝜔𝑥, 𝜔𝑦 = ඵ 𝑓 𝑥, 𝑦 𝑒 −𝑗𝑥𝜔𝑥 𝑒 −𝑗𝑦𝜔𝑦 𝑑𝑥 𝑑𝑦 MSc in BioMedical Engineering (BME-AUTh) 2D FT Spatial frequency MSc in BioMedical Engineering (BME-AUTh) 2D FT Is separable: First 1D FT across columns and then 1D FT across lines 𝐹1 𝜔𝑥 , 𝑦 = න 𝑓 𝑥, 𝑦 𝑒 −𝑗𝑥𝜔𝑥 𝑑𝑥 𝐹 𝜔𝑥, 𝜔𝑦 = න 𝐹1 𝜔𝑥 , 𝑦 𝑒 −𝑗𝑦𝜔𝑦 𝑑𝑦 Sampling theorem: 1 1 Dx𝑚𝑖𝑛 = Dy𝑚𝑖𝑛 = 2𝑓𝑥𝑚𝑎𝑥 2𝑓𝑦 𝑚𝑎𝑥 MSc in BioMedical Engineering (BME-AUTh) PROJECTION SLICE THEOREM Let f(x,y) the distribution of an image and f(x,y) one projection on y axis, Px(y) The FT of Px(y) equals the FT of f(x,y) for kx=0, thus FT[Px(y)] = F(kx,ky)lkx=0 The slice of the 2D FT along dimension kx=0 equals the FT of the projection along the same dimension MSc in BioMedical Engineering (BME-AUTh) PROJECTION SLICE THEOREM Image rotation → 2D FT rotation → Independent of kx (kx=0) The 1D FT of the projection equals the “slice” of the 2D FT of the image along the same dimension as that of the projection MSc in BioMedical Engineering (BME-AUTh) PROJECTION SLICE THEOREM Principal concept for the modern reconstruction techniques A projection in Radon 𝑅1 𝑡 = 𝑅 𝑡, 𝜃 (specific θ) and ο PF for the same angle θ PF1(ω)= PF(ω,θ) form a 1D FT match MSc in BioMedical Engineering (BME-AUTh) PROJECTION SLICE THEOREM for every θ MSc in BioMedical Engineering (BME-AUTh) PROJECTION SLICE THEOREM MSc in BioMedical Engineering (BME-AUTh) PROJECTION SLICE THEOREM Projection Slice Theorem Cartesian to Polar coordinates Radon MSc in BioMedical Engineering (BME-AUTh) RADON TRANSFORM MSc in BioMedical Engineering (BME-AUTh) RADON TRANSFORM MSc in BioMedical Engineering (BME-AUTh) RADON TRANSFORM MSc in BioMedical Engineering (BME-AUTh) RADON TRANSFORM MSc in BioMedical Engineering (BME-AUTh) TOMOGRAPHY TYPES - GENERAL How different types of tomography devices, utilizing different natural phenomena, provide information on projections of unknown distribution The reconstruction step, ie the synthesis of the section image from the projection information, is performed by a computer (Computed Tomography) MSc in BioMedical Engineering (BME-AUTh) X-ray TOMOGRAPHY Very common examination method, known as computed tomography The parameter shown is the attenuation caused by each point of the intersection in an X-ray beam The attenuation depends directly on the density of the material at each point X-rays - ionizing radiation - extensive exposure maybe harmful MSc in BioMedical Engineering (BME-AUTh) X-ray TOMOGRAPHY Two sets of sensors, one for transmitting and one for receiving By mechanical transfer of the patient, the area to be imaged is placed at the level of the sensors To record all views, the arrays of sensors rotate around the axial direction of the CT scanner. Loud and annoying noise due to rotation for image acquisition MSc in BioMedical Engineering (BME-AUTh) EVOLUTION OF X-RAY TOMOGRAPHY Spiral computed tomography X-ray sources and receivers move in a spiral path relative to the subject In essence, the patient moves as the sensors rotate Higher resolution Multisliced computed tomography MSc in BioMedical Engineering (BME-AUTh) EVOLUTION OF X-RAY TOMOGRAPHY Spiral computed tomography X-ray sources and receivers move in a spiral path relative to the subject In essence, the patient moves as the sensors rotate Higher resolution Multisliced computed tomography MSc in BioMedical Engineering (BME-AUTh) ULTRASOUND TOMOGRAPHY It works exactly like a CT scanner, but instead of using X-rays, ultrasound is used. The ratio of the received to the transmitted acoustic volume is measured The attenuation of the acoustic signal is proportional to the path along the section (approximately) MSc in BioMedical Engineering (BME-AUTh) PHOTON EMISSION TOMOGRAPHY The radiation is internal and not external as in other types of tomography It comes from a radioisotope that has been placed inside the body The parameter illustrated is the distribution of the radioisotope inside the organism Information is recorded that has to do with the metabolism and physiology of some organs MSc in BioMedical Engineering (BME-AUTh) PHOTON EMISSION TOMOGRAPHY Two types of CT scanners SPECT (Single Photon Emission Computer Tomography): single photon emission PET (Positron Emission Tomography): positron emission MSc in BioMedical Engineering (BME-AUTh) PET Positrons are emitted (e+, radius b) When a positron collides with an electron then two photons are dissipated and generated The photons move in opposite directions, on the same line Sensors that surround the patient detect photons MSc in BioMedical Engineering (BME-AUTh) PET Thus, it can be found approximately where the radioisoptium is located This records the density of radioisotopes along each line (projection information) MSc in BioMedical Engineering (BME-AUTh) SPECT Similar to PET Photons are not detected at the same time because they are not emitted in pairs C-rays are detected Lower image resolution Cheaper equipment than PET The radioisotopes used have a longer half-life MSc in BioMedical Engineering (BME-AUTh) MAGNETIC RESONANCE IMAGING The reconstructed section is a function of the characteristic properties of each material the density the material the possible motion / diffusion phenomena in the area to be displayed two times of relaxation Ability to display a variety of parameters that each offer a different usable diagnostic function Unlimited resolution as opposed to X-ray/ultrasound scanners (radiation wavelength used is a limiting factor) MSc in BioMedical Engineering (BME-AUTh) MAGNETIC RESONANCE IMAGING (MRI) Sections in any direction we wish, not just axial Possibility of direct reconstruction of a 3D object and not just reconstruction of its sections Ionizing radiation is not used, so it does not burden the body It is based on the phenomenon of nuclear magnetic resonance MSc in BioMedical Engineering (BME-AUTh) NUCLEAR MAGNETIC RESONANCE The fact that the particles have mass and rotate causes an angular momentum vector (J) Because protons are charged, their motion causes a magnetic field or magnetic dipole Therefore, it is characterized by a magnetic moment vector (m) The self-rotation of the neutron (although electrically neutral) creates a magnetic moment vector MSc in BioMedical Engineering (BME-AUTh) NUCLEAR MAGNETIC RESONANCE Overall, the magnetization performs a complex motion (superimposition of the individual motions around Bo and B1) As B1 is applied, the vector m forms an angle θ = yB1tp with the direction of Bo, where tp is the application time of B1 This is called the deflection pulse θ MSc in BioMedical Engineering (BME-AUTh) NUCLEAR MAGNETIC RESONANCE Removing B1, the magnetization vector begins to return to the Z axis, rotating at Larmor frequency. This motion of the magnetization vector produces a weak electromagnetic wave (at Larmor frequency) MSc in BioMedical Engineering (BME-AUTh) NUCLEAR MAGNETIC RESONANCE The rotation of the magnet around Bo can be detected with a coil in the xy plane The movement of the transverse component (at level xy) produces a signal on the receiving coil at the Larmor frequency The signal is called the Free Inductible Decay (F.I.D.) signal and depends on the density of the protons and some relaxation times T1 and T2. MSc in BioMedical Engineering (BME-AUTh) NMR IMAGING Image production including spatial information, ie information on the location of the excited nucleus In other words, the coding of the signal in the space is required The operation of MRI scans is based on the fact that the tuning frequency (rotation) of the nucleus is directly proportional to the intensity of the magnetic field Therefore, the magnetic field must change spatially so that the frequency of rotation of the nuclei also changes spatially. Thus, the cores can be selectively excited depending on their position and the signal to be recorded may have the spatial information required to produce the image. MSc in BioMedical Engineering (BME-AUTh) NMR IMAGING The human body is rich in H nuclei (which satisfy the condition of an odd number of nucleus particles) There is a linear relationship between the rotational frequency (Larmor - ωL) of the elementary magnetizations around the direction of the static field and the intensity of the field 2 populations immersed in different intensities of static magnetic fields, the Larmor frequencies of the protons of the 2 regions will differ MSc in BioMedical Engineering (BME-AUTh) NMR IMAGING When lying under the powerful scanner magnets, the protons in the body line up in the same direction, in the same way that a magnet can pull the needle of a compass. Short bursts of radio waves are then sent to certain areas of the body, knocking the protons out of alignment. When the radio waves are turned off, the protons realign. This sends out radio signals, which are picked up by receivers. MSc in BioMedical Engineering (BME-AUTh) NMR IMAGING These signals provide information about the exact location of the protons in the body. They also help to distinguish between the various types of tissue in the body (protons in different types of tissue realign at different speeds and produce distinct signals). In the same way that millions of pixels on a computer screen can create complex pictures, the signals from the millions of protons in the body are combined to create a detailed image of the inside of the body. MSc in BioMedical Engineering (BME-AUTh) MRI-SCANNERS MSc in BioMedical Engineering (BME-AUTh) Α: 1.5 Τ Β: 3Τ MRI-SCANNERS PD weighted T1 Weighted T2 weighted T1-weighted MRI enhances the signal of the fatty tissue and suppresses the signal of the water. T2-weighted MRI enhances the signal of the water. Proton Density (PD) displays the number of nuclei in the area being imaged. MSc in BioMedical Engineering (BME-AUTh) MRI SCANNER STRUCTURE ◼ Magnet: Produces the constant magnetic field Bo ◼ Gradient coils: coils generating the scalar magnetic fields at 3 axes to form the image ◼ Radio frequency coils (RF coils): Excitation pulse transmitter and transmitting signal system ◼ Computer: controls the operation of the CT scanner, performs FT, reconstructs the images, etc. MSc in BioMedical Engineering (BME-AUTh) WIRELESS CAPSULE ENDOSCOPY 1805 1868 1957 1932 MSc in BioMedical Engineering (BME-AUTh) WIRELESS CAPSULE ENDOSCOPY Capsule endoscopy was invented based on the principles of electro- optics by engineer Gavriel Iddan to detect gastrointestinal bleeding and related problems that can be life-threatening. MSc in BioMedical Engineering (BME-AUTh) WIRELESS CAPSULE ENDOSCOPY Iddan, Meron, Glukhovsky, Swain (2000): Revolution in the field of endoscopy Painless - non-invasive – effective Imaging of the entire small intestine, something that is almost impossible and very painful for a doctor and a patient Consists of: Wireless capsule Data recording system Image reader software MSc in BioMedical Engineering (BME-AUTh) WIRELESS CAPSULE ENDOSCOPY A) Capsule ingestion and recording for 8 hours B) The capsule is the size of a pill and contains a camera C) Recording of a bleeding ulcer inside the small intestine D) The image of the bleeding ulcer is available to the doctor MSc in BioMedical Engineering (BME-AUTh) WIRELESS CAPSULE Single use It moves with natural peristalsis 11mm diameter - 27mm length - 3.7gr weight 140ο view angle- 256x256 resolution - 2 images/sec CMOS sensor - 8 hour battery Improved version (PillCam SB 3) 2-6 images/sec dynamic recording speed 156ο view angle - 320x320 resolution MSc in BioMedical Engineering (BME-AUTh) Optical dome Lens base Lens WIRELESS CAPSULE LED light sources Photo sensor Battery MSc in BioMedical Engineering (BME-AUTh) IC Transmitter Antenna DATA RECORDING SYSTEM 8 sensors are mounted on the abdominal wall Walkman type digital recorder Battery lasting 8 - 9 hours Mounting zone The sensors detect the images emitted by the capsule The images are stored in the digital recorder MSc in BioMedical Engineering (BME-AUTh) IMAGE PROCESSING SOFTWARE Uploading images from the recorder to the computer Examining images in video format The playback rate of the images is determined by the doctor Ability to view 2 or 4 images simultaneously Save findings Approximate determination of the position of the capsule MSc in BioMedical Engineering (BME-AUTh) WIRELESS CAPSULE Healthy small intestine Capsule in the Crohns disease small intestine MSc in BioMedical Engineering (BME-AUTh) PATENCY CAPSULE uncovered area uncovered area for corrosion by for corrosion by digestive fluids digestive fluids RFID sensor self-dissolving body Check the passability of the digestive system Same dimensions - without camera, batteries, electronics - self- dissolving With RFID sensor (comes with RFID scanner) MSc in BioMedical Engineering (BME-AUTh) 2D SIGNAL REGISTRATION In the medical field, image registration is used for diagnostic purposes when images of the same anatomical structure must be superimposed on each other. Registration methods are used for combining computer tomography (CT) and magnetic resonance imaging (MRI) data to obtain more complete information about the patient, for monitoring tumor growth, for treatment verification, for comparison of the patient’s data with anatomical atlases. MSc in BioMedical Engineering (BME-AUTh) 2D SIGNAL REGISTRATION Feature detection Feature matching Transform model estimation Image resampling and transformation Steps involved in image registration - Zitova, B.and Flusser, J. Image registration methods: a survey, Image Vision Comput. 21 (11) (2003) 977–1000 - Mambo, S. , Djouani, K. , Hamam, Y. , Wyk, B. , Siarry, P. (2018). 'A Review on Medical Image Registration Techniques'. World Academy of Science, Engineering and Technology, Open Science Index 133, International Journal of Computer and Information Engineering, 12(1), 48 - 55. MSc in BioMedical Engineering (BME-AUTh) APPROACHES OF IMAGE REGISTRATION Transformations using Fourier Analysis Cross correlation approach using Fourier Analysis Sum of squares search technique Eigenvalue Decomposition Moment matching techniques Warping Techniques Procedural approach Anatomic Atlas Internal landmarks MSc in BioMedical Engineering (BME-AUTh) 2D SIGNAL REGISTRATION Application of Image registration in MR mammography Maximum intensity projection of contrast controlled MR mammography (a) with and (b) without registration Wyawahare, M.V., Patil, P.M., Abhyankar, H.K., et al., 2009. Image registration techniques: an overview. International Journal of Signal Processing, Image Processing and Pattern Recognition 2, 11–28. MSc in BioMedical Engineering (BME-AUTh) REGISTRATION PROBLEM DEFINITION q = (912,632) p = (825,856) q = T(p;a) Pixel location in first image Homologous pixel location in second image Pixel location mapping function MSc in BioMedical Engineering (BME-AUTh) IMAGE REGISTRATION Define a transform T that will map one image onto another image of the same object such that some image quality criterion is maximized. A mapping between two images both spatially and with respect to intensity I2 = g (T(I1)) MSc in BioMedical Engineering (BME-AUTh) IMAGE REGISTRATION Spatial transform that maps points from one image to corresponding points in another image. Rigid: Rotations & translations Affine: also skew and scaling Deformable: Free-form mapping MSc in BioMedical Engineering (BME-AUTh) IMAGE REGISTRATION Rigid: Rotation One parameter (angle θ) 𝑐𝑜𝑠θ −𝑠𝑖𝑛𝜃 0 𝑅 = 𝑠𝑖𝑛𝜃 𝑐𝑜𝑠𝜃 0 0 0 1 MSc in BioMedical Engineering (BME-AUTh) IMAGE REGISTRATION Affine transformation MSc in BioMedical Engineering (BME-AUTh)