11th Asian-Pacific Conference on Medical and Biological Engineering

Optical coherence tomography (OCT) is widely used in biomedical imaging. However, background noise severely affects the diagnosis and identification of diseased tissues. Here, we propose a background-noise reduction method based on region filling. OCT images of in vivo skin tissues, ex vivo brain tumors and ex vivo brain normal tissues had been analyzed to remove background noise using the method. Results demonstrated that the mean of the noise reduction ratio (NRR) reached more than 95% and the image-to-noise ratio (INR) ranges from −0.5 dB to 2.5 dB. The denoised OCT images demonstrated a better three-dimensional (3-D) visualization effect. The proposed method is useful for removing the background noise of OCT images in medical imaging.

Vessel size imaging (VSI) was required an intravascular superparamagnetic contrast agent for vessel diameter estimation. Apparent diffusion coefficient (ADC) calculated from single refocused pulsed-gradient spin-echo (SRSE-DWI) was influenced by the magnetic susceptibility of vasculature. The influence of the ADC from the vessels can be reduced by using twice refocused spin-echo (TRSE-DWI) sequence. We evaluate the VSI with SRSE-DWI and TRSE-DWI using Monte-Carlo simulations in a vascular model. It was also evaluated in nine patients with cerebral infarction in the penetrating branch area within two weeks after stroke onset. The simulation results show that vessel radius was determined uniquely from the pair of SRSE-DWI and TRSE-DWI signal. Clinical images also suggested the vasodilation in the periphery of cerebral infarction might be detected in the difference image of SRSE-DWI and TRSE-DWI. It suggests that vessel radius might be estimated by acquiring SRSE-DWI and TRSE-DWI. It could be useful for the clinical diagnosis.

Real-time surgical tool tracking is a critical component of computer-assisted surgery, because it is highly instrumental to analyze and understand the surgical activities. Nowadays, many deep learning methods take fully advantage of very deep networks and track by detection. Although these methods work well, but they take up a significant amount of time and computational resources. To address this problem, we propose a new network which use the cascade of refined convolutional neural network and long short-term memory for real-time single tool tracking based on the Real-time Recurrent Regression Networks (Re3). Our method is tested on the publicly available standard dataset from UCL (University College London). The experimental result show that our method achieves better performance than state-of-the-art tracking methods in terms of accuracy and speed.

This study proposes the development of a novel palpation system for use in the laparoscopic surgery setting, which relies on a wearable motion-sensing armband worn on the palpating arm. The system is tested by prodding the laparoscopic tool on a simulated human tissue with an anomaly underneath and measuring the effects of the forces experienced by the palpating arm by recording the accelerometer data embedded in the armband. Results show that the proposed system could detect the anomalies embedded in the gelatin phantom based on the change of the amplitude of the accelerometer data. A decrease in the average peak-to-peak amplitude of accelerometer signal by as much as 32.29% was observed for the single point prodding experiment, while a decrease of up to 22.93% was observed for the raster scan experiment. The results from this feasibility experiment confirmed the potential role of the accelerometer signal in the motion-sensing armband in being able to assist laparoscopic palpation.

This study explores surface-based electromyography (sEMG) to evaluate the effects of tool handling and fatigue on the muscle activation of laparoscopy surgeons. Specifically, the Myo Armband is used because it is a wearable, unobtrusive, and wireless sensor. Collected EMG signatures showed that more complex gestures have more active muscle groups. For the fatigued state, analysis of the signatures using the RMS feature showed that for more complex gestures, RMS increased because of muscle compensation due to fatigue. Accuracy exercise findings showed that there is a decrease in %accuracy when using an endoscope compared to when looking directly and there is a slight decrease in %accuracy once fatigue is induced. 100% accuracy cannot be reached due to awkward tool handling. Through various experiments, this study presented possible methods to evaluate the effects of tool handling and fatigue on the muscle activation of surgeons through use of sEMG. Analysis of these signatures may lead to better understanding on how specific tool design affects muscle activity and muscle fatigue, and to utilize this as a basis for a more ergonomic laparoscopic set-up and tool design.

Recent trends in the development of rehabilitation for stroke patients suggest that the integration of motor and cognitive training is one of the promising techniques. Gyro-roller rehabilitation system has been developed based on the integration of robotics and virtual reality game for bilateral upper extremity training. In this study, the feasibility of the combined training is investigated with ten healthy participants and a chronic stroke patient. Amplitude analysis of electromyography in both arms in relation to training levels is carried out to investigate the effects of virtual reality based bilateral training on motor activity. Also, the integrated system with short term memory test has been found to induce the result metrics as digit span assessment and stimulate muscle activity. Overall, results suggest that combined cognitive and motor training with Gyro-roller rehabilitation system is possible for stroke rehabilitation.

In order to analyze stress and flow field around the aortic valve, various numerical method have been developed. The validity of the results has been confirmed from the point of hemo dynamical, fluid dynamical, view. However, it has not been evaluated from the point of view of structural mechanics. The purpose of this study is to construct a stress measurement experiment using a phantom model and to establish a numerical analysis method for it. The phantom model of the aortic valve was created with reference to the valve shape used for aortic valve reconstruction surgery. The displacement field of the phantom model is measured using Digital Image Correlation (DIC), which measures the three-dimensional displacement field on object surface using a stereo camera. In the numerical analysis, a finite element model simulating a phantom experiment is created, and quasi-static analysis is performed by an explicit method using LS-Dyna. Though the experimental results have not been obtained yet, according to the numerical analysis, max principle strain will increase at the fixed commissure and center of the coaptation.

How ultrastructural alterations in cardiomyocytes lead to heart disease remains unclear. Therefore, there is interest in simulate cardiac myocytes with realistic 3-D models of myocyte shape and structure. Using deep learning and image processing, we developed an efficient method to generate 3-D models of cardiac myocytes without the need for manual segmentation. Image processing was used to segment borders between mitochondria in two series of serial-block-face scanning electron microscopy (SBF-SEM) images of healthy mouse ventricular cardiomyocytes. Computed mitochondrial size and morphologies were in good agreement with published measurements. These new automated methods will facilitate more realistic simulation studies of cardiac myocytes and enable new investigations of the links between altered mitochondrial morphology and heart diseases.

The aortic valve is always under great mechanical stress. It is known that stress is greatly involved in the progression of the disease. For treatment, information such as valve stress, strain, and blood flow are helpful. Until now, the relationship between mechanical stress and disease progression has been studied in clinical research, phantom experiments, and numerical simulations.The purpose of this study is to show a new diagnostic indicator of the aortic valve based on numerical simulation. Based on the three-dimensional shape of the aortic valve reconstructed based on CT images, a personalized mesh was constructed. NURBS is used to construct personalized mesh. Using the mesh, we perform numerical simulations and obtain useful information for diagnosis.Deep learning is used to reconstruct three-dimensional shapes from CT images. As an analysis method, ALE fluid-structure interaction analysis is used. The Navier‐Stokes equation is weakly formalized, finite elements discrete, linearized, and conformed with structural analysis formulas. SUPG and PSPG stabilization are used for fluid analysis.Three mesh based on CT images of patients(patient1–3) were created and analyzed. The stress condition of the valve, the flow field of the blood, and the pressure field was visualized and compared. There was a significant difference in the analysis results for each personalized mesh. The length of the perimeter, area, and Young's modulus greatly affected the way the valve opened.

The multiple cylindrical model to estimate urinary volume using image sensors without collecting urine in to the cup has been proposed. However, it was not supposing that the measurement is performed on a western style toilet. So, it is necessary to devise a way to embed an image sensor into a western style toilet where it can keep the geometric parallelism of the multiple cylindrical model. In this study, we constructed an environment where a multiple cylindrical model can be applied by using a raising toilet seat in a western style toilet. And we confirmed the accuracy of estimation of liquid flow rate using simulated urination.

As shown in cases of photosensitive epilepsy, audio-visual stimuli can have various effects on the human body. However, until now, there have been no reports investigating the effect of television commercials on the human body. In this study, for which we obtained ethical approval, we are the first ones to investigate the effect of television commercials on the human cardiovascular and autonomic nervous systems. Cardiovascular and autonomic nervous functions were evaluated using the human face blood flow analyzing system invented in Tohoku University (Jpn.Pt. 4789203, 5161671, 5390851, 5408751). No epilepsy and no patho-physiological reactions were observed after viewing of soft TV commercials including commercials from the advertising awards. Autonomic nerve responses were shown to be altered depending on the interests and preferences of the viewer. Our results suggest that assessing the response of the human body to TV commercials will become a key aspect of the process of creating TV commercials.

The use of chromatic pupil responses may be a novel way to diagnose and monitor diseases affecting either the outer or inner retina.Here we developed a model-based approximation of the dynamics of pupillary diameter after an optical stimulation; we assume the response of the pupil is approximated as the output of a 2nd-order linear model. Model parameters are identified by using a least-squares fitting procedure, thus obtaining an optimal estimate of the activation time and of the shape of the pupil response to the stimulation.Results indicate the model adequately represents the curve shape; despite the presence of artifacts, that hinder the fitting procedure, a significant difference in the time constants of the model of controls and of patients is present.

The fit of prostheses and orthoses is important, because, if the fit is inadequate, patients may feel uncomfortable. A poor fit can cause difficulty in gait, and, in the worst cases, it can cause inflammation. An evaluation system for checking the fit that utilizes human models has been developed. Human models of the residual limb and lower thigh with bone, muscle, and fat were composed. These models were used to evaluate the fit of the transfemoral socket and ankle foot orthosis (AFO) by dynamic finite-element analysis (FEA). For the transfemoral socket, insertion simulation was performed, and the validity of the fat part and pressure distribution of the residual limb were evaluated using magnetic resonance imaging and the force measured by sensors built in the socket. For the AFO, gait simulation of the stance phase was performed, and the transition of strain was compared with the measured result by the 3D-printed AFO, on which strain gauges were pasted. In addition, by performing topological optimization, the weight of the AFO can be reduced by maintaining the stiffness of the AFO. The results indicate the qualitative validity of dynamic FEA for analyzing the prosthesis socket and AFO using a human leg model.

Recently, the development of the technological advances of the prosthesis socket has been considered. However, the successful fitting of the prosthesis socket to the residuum stump is still not satisfied and existing limitations. The investigation of the shear stress contacting interface is important to improve the socket design fitting, and to understand the mechanism that happened inside the socket. In this study, the analysis of the resultant shear stress and the shear angle at the contacting interface of the Manual Compression Casting Technique (MCCT) socket type were presented. The results indicated that the higher the value of shear stress at the anterior proximal and posterior proximal in comparison with the other positions. The outcome would be benefit in clinical and improvement of the socket alignment.

Power and endurance athletes have different physiological and metabolic properties of their skeletal muscle. The aim of this study was to compare the muscle activation between these two types of athletes by evaluating the muscle performance accessed from various types of lower limb muscles. A total of 27 subjects; 14 power athletes, and 13 endurance athletes voluntarily participated in this study. Electromyographic (EMG) signals were recorded from five lower limb muscles; vastus lateralis, rectus femoris, tibialis anterior, gastrocnemius and soleus muscle of dominant leg using wireless electromyography system. EMG signals were analysed by extracting several time and frequency domain features in order to evaluate the contribution of each muscles in distinguishing between power and endurance athletes. Comparison between groups was quantified and evaluated by Mann-Whitney U Test. Result showed that significant difference (p < 0.05) was found only in vastus lateralis, rectus femoris and tibialis anterior muscle for several EMG features. Meanwhile, no group differences were observed for gastrocnemius and soleus muscle in all features. Finding shows that physiological differences between power and endurance athletes was much easier to be accessed from vastus lateralis muscle as indicated by significant test result of most selected EMG features.

Ankle Foot Orthosis (AFO) is a rehabilitation device which is used to assist patients with weakness over the ankle, foot and part of the leg especially when surgery is not warranted or during the recovery phase of the foot. In this research, the Pneumatic Artificial Muscle was used to develop and actuate the AFO to mimic the movement of the ankle, namely, plantarflexion, dorsiflexion and eversion. Gait pattern analysis was used to gauge the effectiveness of the AFO in human subjects. Two healthy subjects were recruited to simulate gait patterns. The gait analysis showed that the AFO actuated PAM was able to resist plantarflexion while assisting dorsiflexion. During the stance phase, the AFO were able to allow free ankle movement up to approximately 5° dorsiflexion. It also allowed approximately 9° eversion of the ankle from neutral position. Toe drag in the swing phase was avoided as the foot drop was prevented. In conclusion, the AFO performed well in providing support for a foot drop condition and this presents an opportunity for further testing in real patients with foot drop.

This paper presents a study on the conceptual design of an exoskeleton suit for post-stroke rehabilitation. A lower-limb exoskeleton suit is proposed to facilitate the post-stroke patients in the restoration of their gait motion. The aim of this study is to design and analyze the mechanical properties and strength via the finite element method (FEM). The exoskeleton suit is capable to assist the victims of post-stroke patients from sitting to stand-up positions and vice versa. The design aspect takes the basis of the human anatomy of Asian people with an average mass of 80 kg. The most effective and optimum design of the exo-suit model is evaluated through FEM static stress analysis using CATIA and ANSYS. Results proved that the proposed design is able to fully support the aforementioned average mass with the maximum Von-Mises stress of 9.5887e7 N/m2 (Standing Position) and 160.16e6 N/m2 (Bent Knee Position) of which did not exceed the yield strength of 180e6 N/m2

This study introduced Kinect as motion capturing device and its potential as analysis platform by developing graphical user interface (GUI) for observational gait analysis system. Spatiotemporal parameters and kinematic gait parameters of three normal individuals and two individuals with knee osteoarthritis were recorded using the proposed system and VICON system simultaneously. Correlation study were conducted between the proposed system and VICON system. The proposed system successfully generates and display the observational gait analysis results. The generated results show a strong correlation with VICON system and have no significant difference. As a conclusion, the developed system for gait analysis has high potential to be utilize in clinical settings. Some improvements should be done to increase the practicability of the system.

According to World Health Organization, the number of stroke patients related to upper extremity hemiparesis increases yearly. Due to this, many types of rehabilitation devices have been invented to help these patients, especially to recover their hand functions. These include exoskeleton devices which aims to deliver high intensity therapy for a longer period. Every exoskeleton rehabilitation device is made up of three components; exoskeleton mechanism, actuator and control system. Control system is central to rehabilitation devices. The accurate control algorithm implementation is very crucial for effective rehabilitation. Therefore, the performance and of the control algorithm for any rehabilitation device must be investigated to assess its efficiency. This study aims to investigate the performance of position control based on a new control strategy system of an index finger exoskeleton namely FiCoPRS, by comparing the Heuristic and Ziegler Nicol method based on various controller configurations. The parameter controlled was position of index finger measured through MCP joint angle. PID parameters were tuned using heuristic method and Ziegler Nichols’ closed loop method. In order to assess the performance of the control system, actual data from position control was compared to reference values. This study suggested that the PID controller managed to control the parameter according to the set point better than P, PI and PD controller. The device could be manipulated to move to any position within the range of exoskeleton motion accurately. The results of the study also suggested that the performance of the control system for rehabilitation was very good as the error was less than 10%. The current findings prove that the study is useful and has contributed significant knowledge in understanding better the control system of an index finger exoskeleton.

Many caregivers experience occupational injuries due to patient care activities. An analysis of actual occupational injury reports will be effective to prevent such occupational injuries. However, analyzing these reports is difficult because they comprise text data that have not been quantified or standardized. In this study, we employ a text mining technique and analyze 422 occupational injury reports that include the word “patient.” We obtained Severe Injury Reports published by the Occupational Safety and Health Administration. These reports were investigated using cluster analysis and a co-occurrence network based on Jaccard coefficients. The cluster analysis results indicate that are fall accidents could be caused by cords on the floor, finger injuries were associated with doors, and lower back, or neck pain was related to lifting in patient care activities. The generated co-occurrence network showed that fall accidents and lower back injury are common occupational injuries among caregivers. These results indicate the need to prevent fall accidents and lower back pain among caregivers by improving the work environment and workspace strategies, such as cord placement and caregiver postures.

Cardiovascular Disease has a severe mortality rate in Bangladesh. Due to the unique pattern of cardiovascular diseases in Bangladesh, local health professionals are facing immense difficulties in both the cases of diagnosis and treatment. Machine learning and Deep Learning based approaches to diagnose CVD in early stages will surely provide an edge to tackle this alarming situation. But proper database of cardiovascular patients is still not established due to cost effectiveness (as per local perspective) and other local standpoints. The current situation of CVD has become a barrier to achieve Health Related Sustainable Development Goals set by the government of Bangladesh. Our research group is focused to solve this problem by proposing a new model to build a centralized CVD database after minutely analyzing the local health infrastructure. This cost effective and viable approach will surely lead the way for future machine learning and deep learning based researches in order to detect CVD in early stages for the benefit of local population. The benefit of such model goes beyond the boundaries of Bangladesh, as low income localities with infirm health infrastructure can also adapt such model to build centralized CVD database.

This study presents a method of computing the coefficients of auto-regressive (AR) model at first and then showing the spectrogram of the 1st order coefficient of the AR model. The spectrogram is a popular time-frequency analysis for studying the lung sound. Researchers usually employed the method to find the components of the specific frequencies to enhance the recognition of the lung sounds. The AR model is a powerful mean for time series analysis. The proposed method was tested by segments of normal breaths sound and those of wheezing. The results showed that the normal breaths seldom display the continuous stripes, but the breaths of wheeze (asthma, bronchiolitis, and COPD) usually present the continuous stripes in a manner of the vertical trend in the spectrograms of the 1st order coefficient of AR model. In conclusion, we find a faster computing process (80% off time-consuming) to recognize the wheeze as a manner of better performance than spectrogram only.

Three-dimensional objects can be recognized regardless of the viewing angles. It was reported that the cells in the temporal association area responded tolerantly in a certain viewing angle range to the objects experienced in discrimination at the same viewing angles. In the present study, machine learning was applied to the population response vectors in the temporal association area. A classifier was trained to create a hyperplane that separated an object from similar object at the same viewing angles and then tested by the population response vectors to the object images at a different viewing angle. The discrimination performance evaluated from true positives, false positives, and d’ values was higher in the objects experienced in discrimination at the same viewing angles than those without discrimination experience in the whole 100–600 ms time window. For the objects experienced in discrimination at the same viewing angles, the d’ values were low in the early phase of the responses of 100–220 ms time window, but became higher in the late phase of 220–600 ms time window. The d’ values for the objects experienced by learning association of different views of the objects were high in the early and late phases of the responses. The results demonstrate computational models for object representation created by using cellular responses in the temporal association area, and differences of the representation of the viewing angle tolerance in different time windows.

Diagnosis using the contrast CT images is a simple and minimally invasive method for estimating the functional severity of coronary stenosis. Since the contrast CT images get the distribution of the contrast agent, it is important to clarify the advection and diffusion of contrast agent in the diagnosis of stenosis severity using a CT value gradient such as Transluminal Attenuation Gradient (TAG). In this study, contrast agent flowing through stenosed tubes was imaged by X-ray simple projection. Contrast agent flowed in water or blood under steady flow conditions. The tubes were diameter of the tube was 3 and 2.1 mm, artificially stenosed with area reduction of 62% and 69% simulating a stenosed coronary artery for the case of water and blood respectively. In the case of water, the stenosis spread the contrast agent uniformly over the entire cross section. On the other hand, in the case of blood, the region with high concentration of the contrast agent existed along the tube wall. In the analysis simulating the case of blood, the jet flow was off the center of the axis and along the tube wall. However the analysis of contrast agent dynamics did not agree with the experiment. This may be affected by the fact that the analysis is sorely in consideration of advection.

An extracorporeal ventricular assist device (VAD) intended for patients with severe heart failure has been developed. The VAD consists of a centrifugal blood pump with a new hydrodynamic bearing system that enables non-contacting rotation of the impeller. Preclinical evaluation of the VAD system was performed by bench tests as well as by a series of chronic animal experiments. The bench tests demonstrated that the impeller successfully rotates without mechanical contact under a wide range of operating conditions. In the 30-day animal experiments, the VAD demonstrated good antithrombogenicity under LVAD conditions as well as RVAD conditions. The system was finally decided to have good biocompatibility for use within 30 days.

Pneumatic artificial muscles (PAMs) have many advantages such as the natural compliant, lightweight, high ratio of weight to power, thus making them an ideal choice for humanoid robots using in the rehabilitation system. However, control strategies of PAMs are facing difficulties due to their high nonlinear characteristics, uncertainties and response lag. The three-element model [1], which consists of a contractile element, spring element and damping element in parallel, was investigated to describe the dynamics of PAMs. To improve the trajectory tracking performance of PAM, a sliding mode control based on nonlinear disturbance observer (SMCBNDO) was applied. The experimental results show that the identified three-elements model of PAM and the SMCBNDO can achieve good accuracy of trajectory tracking controller. This work proved that this model and non-linear control method can be used for the application of the PAM-based rehabilitation system.

During tumor resection, the boundaries of some types of tumors are difficult to distinguish with the naked eye. Doctors also face problems with tissue deformation and avoidance of important blood vessels and nerves. Although optical biopsy provides a rapid intraoperative diagnosis method, how to effectively combine with laser ablation and other treatment methods to achieve accurate ablation or resection of tumor tissue still faces great challenges. Therefore, we designed a theranostics system with non-common optical path structure and an intelligent control method. By fixing the front-end of optical coherence tomography (OCT) sample arm with the high power laser probe, and intelligently driving the micromirror to control high power laser spot based on the diagnosis result of OCT image, the system can excise the tumorous tissues with high precision. We verified the feasibility of the system and method on the ex vivo porcine brainstem model. The experimental results showed that the system can sensitively recognize and quickly ablate lesions in real time. The diagnostic and planning process takes 26 s. The ablation precision can reach millimeter order. The theoretical removal rate of light spot planning method is about 90%. The ablation cavity’s position error is about 0.15 mm.

The transcutaneous energy transmission system (TETS) is the most promising method for transmitting driving energy to a totally-implantable artificial heart (TIAH) without additional invasive procedures. In addition to a rectifier, TETS also requires a regulator to supply current at constant voltage to the TIAH and the implanted secondary battery in the inner circuit. In this study, we describe a double wire spiral coil with center tap that can perform full-wave rectification with a simpler circuit configuration. Hence, it suggests that the number of components in the internal circuit can be reduced while achieving almost the same power transmission efficiency.

A method named brain rhythm sequencing has been proposed to interpret Electroencephalography (EEG) signal for emotion recognition. For this purpose, first, the reassigned smoothed pseudo Wigner-Ville distribution (RSPWVD) has been adopted. Then, all generated sequences have been evaluated for classifying the emotional states based on arousal and valence. Moreover, the classification was by similarity measurements between pairwise of the sequences. After evaluating the accuracies by using the sequences from different channels and time durations, on one hand, the significant channel for classification has been identified so that the recognition can be accomplished by considering single-channel. On the other hand, the particular duration which was valuable for classification has been found so that the size of required data can be reduced. Thus, emotion recognition can be performed by fewer data resource. Results revealed that the accuracy of ~70% has been realized and hence, the proposed method could open a new way to implement data analysis for EEG-based applications such as emotion recognition.

Brain extraction is an essential segmentation step in many neuroimage analysis studies. Convolutional neural network approaches have been proposed in the literature and widely used for medical image segmentation, outperforming traditional segmentation approaches. However, these learning-based methods highly rely on ground truth data which needs significant time and labor inputs from experienced experts. In addition, the independent identical distribution assumption is difficult to comply with in practice, leading to the overfitting problem across different datasets. This paper introduces a 3D U-Net based self-fine-tuning network for brain volumetric segmentation that learns from quickly available but imperfect labeled images generated by a model-based approach. We outline two typical application scenarios of our method: (1) In a ground truth absence setup, an improved brain extraction tool (IBET) algorithm is used to generate brain label images fully automatically. The self-fine-tuning network learns from these imperfect labels and yields more accurate predictions. (2) In a transfer learning setup, the IBET algorithm is used to fast extract dataset-specific labels for fine tuning the pretrained network (overfitting prevention). We have evaluated the performance of the proposed method on four datasets and achieved satisfactory results.

A deep residual learning convolutional neural network (CNN) is applied to improve image quality and speed up the reconstruction of compressed sensing magnetic resonance imaging. The reconstruction performances of proposed method are compared with iterative reconstruction methods.The proposed method adopts a deep residual learning CNN to learn the residual component between the input and output images (i.e., aliasing artifacts) for image reconstruction. The CNN-based reconstruction is compared with iterative reconstruction methods to clarify its characteristics. The peak-signal-to-noise ratio (PSNR) and the structural similarity index (SSIM) are examined for various numbers of sampling rates.The experimental results demonstrate that reconstruction time is drastically reduced to 0.022 s per image compared with that for conventional iterative reconstruction. Compared with iterative reconstruction, superior quality was obtained for low sampling rates (30% or 40%). These results suggest that CNN can learn the rules of the occurrence pattern of aliasing artifact. With the proposed method, the structure and small contrast in the original image are well preserved in the reconstructed image and aliasing artifacts are mostly removed.A deep residual learning CNN can recognize and predict aliasing artifacts. It is demonstrated that reconstruction time is significantly reduced.

Progressions in wound controlling to encourage rapid and successful healing of wound by providing a warm humid environment at the site of application of treatment aims to facilitate patient immediate rescue. Ceramide has an extravagant prospective in providing structural support and preserving the water-retaining properties of the epidermis. Hydrogel and Composite sponges based on chitosan/gelatin, with and without incorporating ceramide, was synthesized by chemical crosslinking and evaluated as wound dressing material. Physical characteristics of the prepared Hydrogels such as water vapor transmission rate (WVTR) and composite sponges (water absorption and weight loss) were inspected. The lowest WVTR was showcased by the sample containing chitosan and hydrogel in 1:1 ratio, with ceramide included in them. Addition of ceramide led to lower WVTR. FTIR of hydrogel membranes and composite sponges authenticated the existence of free hydroxyl groups. N-H and O-H stretching and the intramolecular hydrogen bonds were indicated by presence of bands for the composite material in the region 3272.6 − 1028.7 cm−1 and 3280.1 − 1028.7 cm−1 in hydrogel membrane. Stretching of C = O bonding due to the formation of hydrogen bonds are detected in amide I mode, which implies the presence of hydrogen bonding. The SEM images revealed that the surface of hydrogel membranes were compact in nature, which prevents microbes from entering. The porous nature of the internal structure of the composite sponges allows for improved water retention. Increasing the amount of chitosan by threefold that of hydrogen improved the porosity of the structure.

Hyperthermia utilizing magnetic particles has been widely studied as a possible method for cancer treatment in recent years. This method induces apoptosis (cell death) using heat generation from magnetic particles which are injected into a tumor region and subjected to an external high-frequency magnetic field. In previous studies, we succeeded in developing microsize thermosensitive ferromagnetic particles with a low Curie temperature of around 40–45 °C as a self-controlled heating element that possibly prevents the local overheating surrounding healthy tissues. Further, by utilizing the change of its permeability around its Curie temperature resulting in the change in the magnetic flux density around tumor region, we also developed a wireless temperature measurement. In addition, to localize magnetic particles which cannot be seen from the body surface, we also developed an automatic localization system by using a robot arm equipped with three detection coils symmetrically installed in heating coil. To make our proposed elemental technologies feasible in clinical settings, we have also been developing a magnetic hyperthermia system that can treat deep-seated tumors up to 5 cm. In this paper, we report our elemental technologies which have been developed so far such as heating elements, temperature and position monitoring technologies.

Recent advances in targeting cancer therapy with stimuli responsive materials provide the opportunity to reduce systemic toxicities. Targeting cancer therapy employing ultrasound with phase-change nano-droplets (PCNDs) potentially has several distinct advantages over current cancer therapeutic modalities. PCND is a liquid perfluorocarbon nanoparticle that its liquid phase is transformable to gas phase upon ultrasound exposure. The vaporization phenomenon of PCND can selectively treat cancers mechanically by conjugating active targeting moiety to it. Here, we present that adenosine triphosphate (ATP) and high-mobility group box 1 (HMGB1), which are known as damage-associated molecular patterns (DAMPs), were released from a cultured cancer cell line by selective intracellular vaporization of PCNDs. The results indicate that not only destruction of cancer cells by intracellular vaporization but also anti-cancer immune response can be expected. The activated immune system can potentially promote systemic anti-cancer effect in vivo. The in vitro DAMPs release motivate further research toward the in vivo investigation.

Engineering three-dimensional tissues in vitro for regenerative therapy is highly desired. Vascular beds with connectable arteries and veins play crucial roles in fabricating three-dimensional tissues in vitro and improving the survival rate of transplanted three-dimensional tissues in vivo. Here, we developed a method to reconstruct a new vascular bed that could be implanted into humans for clinical applications. Porcine small intestine with an arteriovenous loop was selected as a vascular bed skeleton and decellularized using transvascular perfusion with both detergent and enzyme. Subsequently, the decellularized intestinal graft was re-endothelialized with human cells to allow blood perfusion into the vascular bed. The small intestinal graft was successfully decellularized without severe damage to the burst pressure of the graft, and the decellularized intestinal graft could be transplanted without a severe inflammatory response in rats for two weeks. In addition, the decellularized intestinal graft was partially vascularized with human endothelial cells in seven days. Therefore, the reconstructed intestinal vascular bed may serve as a human transplantable vascular bed with potential in clinical investigations.

This study aimed to investigate the effect of tactile stimulation by touching the agonist muscle during the arm curl exercise. Nine healthy males (age, 20–35 years) performed five repetitions of arm curl under two different focus instructions; for the external focus condition, participants were instructed to lift the bar, whereas for the internal focus condition, the participants’ biceps brachii was lightly touched by the investigator and they were instructed to concentrate on muscle contraction at the touched point. Muscular activity in the biceps brachii was evaluated during the exercise. The muscle activity in the internal focus condition was significantly greater than that of the external focus condition (P = 0.01, d = 0.19). This result suggests that internal focus caused by an individual’s consciousness, as well as a focus attention through touching the agonist area increased the agonist muscle activity.

The must-have clinical criterion or main symptom associated with knee osteoarthritis (OA) is knee pain. Gait alteration usually occur mainly to compensate the pain experienced at the knee joint. It consequently causes the deterioration of functional ability and affects activity of daily. The aim of the study is to extract kinematic gait features that characterize the knee OA gait and its association with pain assessment. Eight knee OA patients and five asymptomatic subjects were recruited and kinematic and spatiotemporal features were obtained using 3D Vicon motion analysis system during gait trial. There were 13 kinematic features extracted using Principal Component Analysis (PCA) and only two features which are hip flexion/extension phase shift and knee flexion/extension range of motion have significant correlation with pain assessment using ICOAP and KOOS questionnaire.

The PET-laparoscope system has been proposed to diagnose lymph node metastases intraoperatively in gastric cancer surgeries. The system uses two detector units, a fixed detector array and a movable detector probe to acquire PET images. The detector probe can be inserted into a patient’s stomach through laparoscopic ports. A surgeon can move the detector inside the patient’s body but the movement is limited. In this study the system was evaluated with an abdomen phantom so that the probe movement was realistic. Moreover, performance of an MLEM algorithm was investigated for the PET-laparoscope system. Experiments with the abdomen phantom confirmed that the system performs well when the probe movement is similar to that achievable during a laparoscopic surgery. MLEM algorithm provided better images. Image artifacts were reduced and the spatial resolution in lateral directions were better than 3 mm and the spatial resolution in depth direction was around 10 mm. Intraoperative PET image was overlaid on the laparoscopic video feed.

While strain gauges are commonly used in a force sensor, we used multiple digital Hall-Effect magnetic sensors for developing a three-dimensional force sensor. We are aiming to develop a round shape three-dimensional force sensor. In order to get three-dimensional unknown force components, three independent outputs from the sensor are at least necessary. Four or more outputs however, would not be useless but redundant data sets, which could be useful to improve sensor accuracy. We calculated the unknown three-dimensional force values using nine outputs or less and estimated how many redundant outputs we should use to get good accuracy. As a result, when three outputs were used, the sensor accuracy was ranging from 8.2 to 15.7% depending on which outputs were used. When six outputs were used, the sensor accuracy was ranging from 5.5% to 5.6%. When nine outputs were used, the sensor accuracy was 4.1%. Although the best accuracy was acquired when we used all nine outputs, improvements were minimal when we used six or more outputs.

The study of falling, especially for the elderly, is a very critical problem in geriatric care and medicine. This study investigated the potential role of a surface-based electromyography (sEMG) and inertial measurement unit (IMU) sensor package for analyzing and detecting humans falling. Previous studies have only relied on motion sensors only, such as accelerometers and gyroscopes, detecting falling, which completely ignores the muscular force reaction of extremities during the precise moment before a person actually falls. This experiments conducted in this study show promising results of combining the sEMG and accelerometer signals to describe the muscular activity generated as the reflex reaction of the subjects in their arms as they grab onto something to prevent the fall. Further simulations of falling were conducted at a foam pit, specifically deliberate falling and surprised falling. Similar findings were obtained in this scenario where the subjects did not have anything to grab onto upon falling, which cancan be attributed to the fact that the subjects feel the same sensation of reflex and panic whether deliberately falling or surprised falling.

This paper describes wireless biotelemetry system implementation in order to improve the flexibility of existing tele-cardiac monitoring system. The proposed method is an integration of single-lead ECG front end circuit and TCP/IP based wireless communication System on Chip (SoC) running 3 parallel tasks including modified Amplitude Zone Time-Epoch Coding (AZTEC) data reduction algorithm. The electrode wireless node is designed to minimize the size and power consumption meanwhile maintaining acceptable performance of the instrument. On the other hand, an ECG server is implemented as a TCP/IP data receiving unit featuring data recovery algorithm with store and forward capability. This feature has enabled integration with existing tele-cardiac monitoring system, in addition to raw ECG data logging for manual analysis. The experiment has demonstrated the implementation feasibility of ECG acquisition system with commercial ECG front-end IC and an on-chip ADC. A moderate performance and acceptable error have been observed. The proposed system provides a potential improvement in telemedicine service.