Yi Chun Du

Adaptive wavelet network for multiple cardiac arrhythmias recognition

This paper proposes a method for electrocardiogram (ECG) heartbeat detection and recognition using adaptive wavelet network (AWN). The ECG beat recognition can be divided into a sequence of stages, starting with feature extraction from QRS complexes, and then according to characteristic features to identify the cardiac arrhythmias including the supraventricular ectopic beat, bundle branch ectopic beat, and ventricular ectopic beat. The method of ECG beats is a two-subnetwork architecture, Morlet wavelets are used to enhance the features from each heartbeat, and probabilistic neural network (PNN) performs the recognition tasks. The AWN method is used for application in a dynamic environment, with add-in and delete-off features using automatic target adjustment and parameter tuning. The experimental results used from the MIT-BIH arrhythmia database demonstrate the efficiency of the proposed non-invasive method. Compared with conventional multi-layer neural networks, the test results also show accurate discrimination, fast learning, good adaptability, and faster processing time for detection.

Portable hand motion classifier for multi-channel surface electromyography recognition using grey relational analysis

This paper proposes the portable hand motion classifier (HMC) for multi-channel surface electromyography (SEMG) recognition using grey relational analysis (GRA). SEMG provides information on motion detection of flexion and extension of fingers, wrist, forearm, and arm. A portable HMC is developed to identify hand motion from the SEMG signals with an electrode configuration system (ECS) and GRA-based classifier. The ECS consists of seven active electrodes place around the forearm to acquire the multi-channel SEMG signals of corresponding muscle groups. Six parameters are extracted from each electrode channel and various 42 (7 Channels by 6 Parameters) parameters could be constructed as specific patterns. Sequentially, these patterns are sent to the GRA-based classifier to recognize 11 hand motions. Twelve subjects including eight males and four females participate in this study. Compared with the multi-layer neural networks (MLNNs) based classifier, GRA demonstrates the processing time, computational efficiency, and accurate recognition for recognizing SEMG signals. It takes about 0.05s CPU time to identify each hand motion which is close to a real-time process. Therefore, the GRA-based classifier could be further recommend to implement in prosthesis control, robotic manipulator or hand motion classification applications.

Long-term effects of baroreflex function after stenting in patients with carotid artery stenosis

Baroreflex sensitivity is recognized for its prognostic relevance to cardio-vascular and cerebro-vascular risks. However, little is known about the long-term outcome of baroreflex function in patients with carotid stenosis undergoing carotid stenting. Heart rate variability and cardio-vascular autonomic function, including baroreflex sensitivity, were examined using non-invasive methods in 22 adult patients who underwent carotid stenting. They were compared with the normal control group with 22 sex- and age-matched normal volunteers and the risk control group with 10 adult patients with severe stenosis or even total occlusion of the carotid artery without stenting. The groups of patients with stenting and risk controls had significantly reduced valsalva ratio and baroreflex sensitivity measured by the valsalva method compared to normal controls. However, there was no significant difference between patients with stenting and risk controls. There was significant decrease in heart rate response to deep breathing and to head-up tilt in patients with carotid stenting compared to normal controls. Other parameters of cardio-vascular autonomic function showed no difference among the three groups. Reduced baroreceptor function in patients with carotid stenting may be due to underlying diseases rather than the stenting itself. There was no short-term parasympathetic hyperactivity after the stenting, suggesting that the effect is transient rather than permanent.

Synchronizing chaotification with support vector machine and wolf pack search algorithm for estimation of peripheral vascular occlusion in diabetes mellitus

Abstract This study proposes a method for the estimation of peripheral vascular occlusion (PVO) in diabetic foot using a support vector machine (SVM) classifier with the wolf pack search (WPS) algorithm. The long-term presence of elevated blood sugar levels commonly results in peripheral neuropathy, peripheral vascular disease, nephropathy, and retinopathy in patients with Type 2 diabetes mellitus. Patients with PVO disease have decreased walking capability and life quality in diabetes mellitus and poor peripheral circulation of PVO causes morbidity like infection and amputation of the legs or feet of diabetics. This progressively vascular occlusion is often ignored by the patients and primary care physicians in early stage. Therefore, a reliable method of diagnostic assistance is crucial for early diagnosis and monitoring of PVO and prevention of amputation. Photoplethysmography (PPG) is a non-invasive technique for detecting blood volume changes in peripheral vascular bed. Literature indicates that the pulse transit time increases and waveform shape changes increase in PPG of the vascular occlusion. PPG pulses of feet gradually become asynchronous due to the different speed of deteriorating patency and collateral circulation in the peripheral arteries. We utilized synchronizing chaotification to compare the bilateral similarity and asymmetry of PPG signals, and applied SVM to estimate three degrees of PVO. Among 33 subjects tested, this classification technique could recognize various butterfly motion patterns representing severities successfully including normal condition, lower-degree disease, and higher-degree disease. The proposed method has potential for providing diagnostic assistance for PVO of diabetics and other high-risk populations, with efficiency and higher accuracy.

Residual Stenosis Estimation of Arteriovenous Grafts Using a Dual-Channel Phonoangiography With Fractional-Order Features

The residual stenosis estimation of an arteriovenous shunt is a valuable for evaluating outcomes of percutaneous transluminal angioplasty (PTA) treatment and surgical revision. This paper proposes a dual-channel phonoangiography (PCG) with fractional-order features to estimate the residual of stenosis estimation of arteriovenous shunt. The auscultation technique provides a noninvasive tool to monitor the degrees of arteriovenous grafts (AVGs). Then, support methods, such as the Burg autoregressive (AR) method and self-synchronization error formulation (SSEF), are used to extract fractional-order features between the loop site (L-site) and venous anastomosis site (V-site). Using 2-D patterns (nonlinear mapping), a generalized regression neural network (GRNN) is designed as a nonlinear estimate model to indicate the outcome of surgical revision or AVG stenosis upon routine monthly examinations. For 42 long-term follow-up patients, the results of examination show the proposed GRNN-based screening model efficiently estimates residual stenosis.

Implementation of a Burn Scar Assessment System by Ultrasound Techniques

Tissue injury and its ensuing healing process cause scar formation. In addition to physical disability, the subsequent disfigurements from burns often bring negative psychological impacts on the survivors. Scar hypertrophy and contracture limit the joint motion and body function of the patient. With fast development of the current available technologies regarding the scar therapies, not only the process of wound healing has to be focused, but also the cosmetic and functional outcomes need to be emphasized. Therefore, proper evaluation and assessment of the healing process to nil scar status is highly recommended. However, the currently employed tools for scar evaluation are mostly subjective. For example, Vancouver General Hospital (VGH) scar index uses color, pigmentation, vascularity, pliability, and depth of the scar as dependent variables for scar evaluation. These parameters only estimate the superficial surface of the scar, but they can not evaluate the deeper tissue within dermis. Ultrasound is a safe, inexpensive, and multifunctional technique for probing tissue characteristics. In addition, its resolution is not inferior to other measurement techniques. Although 3D-ultrasound is available in clinical application, its still not widely used in scar evaluation because of its high cost. In this study, we proposed a system for scar assessment using B-mode ultrasonic technique. By utilizing the reconstruction methods to search the scar border, many characteristic parameters, including depth, area and volume, can be estimated. The proposed method is useful in assisting the clinician to evaluate the treatment effect and to plan further therapeutic strategy more objectively. In this report, the quantitative assessment system was used to evaluate the scar of a seriously burned patient. In order to verify the reliability of systematic reconstruction method, we constructed a phantom to imitate the scar tissue. The results show that it can achieve more than 90% in accuracy

Chaos synchronization detector combining radial basis network for estimation of lower limb peripheral vascular occlusive disease

Early detection of lower limb peripheral vascular occlusive disease (PVOD) is important to prevent patients from getting disabled claudication, ischemic rest pain and gangrene. This paper proposes a method for the estimation of lower limb PVOD using chaos synchronization (CS) detector with synchronous photoplethysmography (PPG) signal recorded from the big toes of both right and left feet for 21 subjects. The pulse transit time of PPG increases with diseased severity and the normalized amplitudes decreases in vascular disease. Synchronous PPG pulses acquired at the right and left big toes gradually become asynchronous as the disease progresses. A CS detector is used to track bilateral similarity or asymmetry of PPG signals, and to construct various butterfly motion patterns. Artificial neural network (ANN) was used as a classifier to classify and assess the PVOD severity. The results demonstrated that the proposed method has great efficiency and high accuracy in PVOD estimation.

A statistical approach to infer the minimum setup distance of a well chamber to the wall or to the floor for 192Ir HDR calibration

We propose a new method based on statistical analysis technique to determine the minimum setup distance of a well chamber used in the calibration of 192Ir high dose rate (HDR). The chamber should be placed at least this distance away from any wall or from the floor in order to mitigate the effect of scatter. Three different chambers were included in this study, namely, Sun Nuclear Corporation, Nucletron, and Standard Imaging. The results from this study indicated that the minimum setup distance varies depending on the particular chamber and the room architecture in which the chamber was used. Our result differs from that of a previous study by Podgorsak et al. [Med. Phys. 19, 1311-1314 (1992)], in which 25 cm was suggested, and also differs from that of the International Atomic Energy Agency (IAEA)-TECDOC-1079 report, which suggested 30 cm. The new method proposed in this study may be considered as an alternative approach to determine the minimum setup distance of a well-type chamber used in the calibration of 192Ir HDR.

Room scatter factor modelling and measurement error analysis of 192Ir HDR calibration by a Farmer chamber.

We introduce an empirical formula to directly calculate the room scatter factor in the calibration of (192)Ir HDR using a Farmer chamber. Compared to the data of Selvam et al (2001 Phys. Med. Biol. 46 2299), our formula is accurate to within 0.3%. Our method saves time because the room scatter can be obtained in one calculation rather than being deduced through a series of setups of different source-chamber distances. It could also be cost effective because a calibration jig might be no longer necessary. We only need to position the applicator and chamber at a fixed space in air and measure its distance. We also analysed the effects of two possible errors arising from ignoring the room scatter and the measurement error of the source-chamber distance. We recommend that the source-chamber distance should be at least 10 cm.

A novel bipolar pulse generator for high-frequency ultrasound system

Pulse generator is the critical component in all ultrasound systems for driving a piezoelectric transducer to medical or nondestructive testing (NDT) applications. The transducer in ultrasound system was driven by a pulse train generator, which delivers high-voltage bipolar or unipolar pulse train. Several recent papers have discussed both the importance and the design of unipolar and bipolar pulse generator for ultrasound applications. Bipolar voltage pulse has lower unwanted DC and low-frequency component could decrease the leakage current. Also, its peak-to-peak pulse voltage could achieve twice of the voltage rating of the coaxial cable connecting the generator and the transducer. In current commercial ultrasound systems, bipolar pulse generators are commonly used; however, it still had some disadvantages like longer pulse length, which limited the driving frequency and affect the signal performance. This paper purposed a novel design of bipolar pulse generator based on a novel dual P-N channel MOSFET and FPGA timing control. It could produce multi-cycle pulses with center frequency over 50 MHz and shorter pulse length. This design also preserved the low-cost advantage compared to other commercial design of bipolar pulse generator. It is suitable for high-frequency ultrasound Doppler and B-mode imaging applications.