Showing papers by "Ilangko Balasingham published in 2018"

Automatic Colon Polyp Detection Using Region Based Deep CNN and Post Learning Approaches

Abstract: Automatic image detection of colonic polyps is still an unsolved problem due to the large variation of polyps in terms of shape, texture, size, and color, and the existence of various polyp-like mimics during colonoscopy. In this paper, we apply a recent region-based convolutional neural network (CNN) approach for the automatic detection of polyps in the images and videos obtained from colonoscopy examinations. We use a deep-CNN model (Inception Resnet) as a transfer learning scheme in the detection system. To overcome the polyp detection obstacles and the small number of polyp images, we examine image augmentation strategies for training deep networks. We further propose two efficient post-learning methods, such as automatic false positive learning and offline learning, both of which can be incorporated with the region-based detection system for reliable polyp detection. Using the large size of colonoscopy databases, experimental results demonstrate that the suggested detection systems show better performance than other systems in the literature. Furthermore, we show improved detection performance using the proposed post-learning schemes for colonoscopy videos.

Abnormal Colon Polyp Image Synthesis Using Conditional Adversarial Networks for Improved Detection Performance

Abstract: One of the major obstacles in automatic polyp detection during colonoscopy is the lack of labeled polyp training images. In this paper, we propose a framework of conditional adversarial networks to increase the number of training samples by generating synthetic polyp images. Using a normal binary form of polyp mask which represents only the polyp position as an input conditioned image, realistic polyp image generation is a difficult task in a generative adversarial networks approach. We propose an edge filtering-based combined input conditioned image to train our proposed networks. This enables realistic polyp image generations while maintaining the original structures of the colonoscopy image frames. More importantly, our proposed framework generates synthetic polyp images from normal colonoscopy images which have the advantage of being relatively easy to obtain. The network architecture is based on the use of multiple dilated convolutions in each encoding part of our generator network to consider large receptive fields and avoid much contractions of a feature map size. An image resizing with convolution for upsampling in the decoding layers is considered to prevent artifacts on generated images. We show that the generated polyp images are not only qualitatively realistic, but also help to improve polyp detection performance.

RF Channel Modeling for Implant-to-Implant Communication and Implant to Subcutaneous Implant Communication for Future Leadless Cardiac Pacemakers

Abstract: Propagation of radio-frequency signals inside human body is demanding to analyze as it is a highly complex medium consisting of different frequency-dependent lossy materials of varying thickness. Moreover, experimental analyses are also unfeasible because that requires probes to be placed inside a human body to collect the signals. This paper focuses on in-body to in-body implant communication for future multinodal capsule-like leadless cardiac pacemaker technology. The frequency range of 0.3–3 GHz is analyzed using very detailed numerical simulations of digital human models. The results show that the Industrial, Scientific, and Medical radio band of the frequency range of 2.4–2.5 GHz is optimal, having the least attenuation of signals considering the size constraints of the implant antenna. Furthermore, the placement of an additional subcutaneous implant transceiver is studied. The analysis shows that the abdominal wall is the optimal position for the placement of the implant compared to shoulder and lateral side of the body. This result is further validated by an in vivo experiment on an adult pig. The other novelty of the study is the investigation of the channel behavior based on ventricular blood volume of the heart to find out the appropriate timing of the transmission of signals between the implants. The results show that the attenuation of the signal increases with the increase in blood volume inside the heart.

On Bit Error Probability and Power Optimization in Multihop Millimeter Wave Relay Systems

Abstract: 5G networks are expected to provide gigabit data rate to users via the millimeter-wave (mmWave) communication technology. One of the major problems faced by mmWaves is that they cannot penetrate buildings. In this paper, we utilize multihop relaying to overcome the signal blockage problem in an mmWave band. The multihop relay network comprises a source device, several relay devices, and a destination device and uses device-to-device communication. Relay devices redirect the source signal to avoid the obstacles existing in the propagation environment. Each device amplifies and forwards the signal to the next device, such that a multihop link ensures the connectivity between the source device and the destination device. We consider that the relay devices and the destination device are affected by external interference and investigate the bit error probability (BEP) of this multihop mmWave system. Note that the study of the BEP allows quantifying the quality of communication and identifying the impact of different parameters on the system reliability. In this way, the system parameters, such as the powers allocated to different devices, can be tuned to maximize the link reliability. We derive exact expressions for the BEP of $M$ -ary quadrature amplitude modulation and $M$ -ary phase-shift keying in terms of multivariate Meijer’s G-function. Due to the complicated expression of the exact BEP, a tight lower bound expression for the BEP is derived using a novel Mellin-approach. Moreover, an asymptotic expression for the BEP at high SIR regime is derived and used to determine the diversity and the coding gain of the system. In addition, we optimize the power allocation at different devices subject to a sum power constraint such that the BEP is minimized. Our analysis reveals that optimal power allocation allows achieving more than 3-dB gain compared with the equal power allocation. This paper can serve as a framework for designing and optimizing mmWave multihop relaying systems to ensure link reliability.

In-Body and Off-Body Channel Modeling for Future Leadless Cardiac Pacemakers Based on Phantom and Animal Experiments

Abstract: In this letter, the in-body and off-body channel models at the frequency of 2.4 GHz are studied for development of multinode leadless capsule pacemaker technology based on experiments in homogeneous liquid phantom model of human heart and living animal experiments. For conducting the experiments, we design a battery-operated self-contained transmitter capsule consisting of a small antenna and transmitter printed-circuit board, subcutaneous implant and the off-body antennas. The in-body path-loss model obtained from the phantom experiment is a linear function of distance, whereas the off-body path-loss model between the implant and the off-body antenna is a logarithmic function of distance comparable to the free-space path-loss model. The phantom experiment study shows that coupling between implants decreases linearly at the rate of 3.6 dB/cm for cardiac implants and by 4.1 dB/cm for cardiac to subcutaneous implant at 2.4 GHz. The animal experiment results are in good accordance with the phantom results.

Throughput and Delay Analysis of HARQ With Code Combining Over Double Rayleigh Fading Channels

Abstract: This paper proposes the use of hybrid automatic repeat request (HARQ) with code combining (HARQ-CC) to offer reliable communications over double Rayleigh channels. The double Rayleigh fading channel is of particular interest to vehicle-to-vehicle communication systems as well as amplify-and-forward relaying and keyhole channels. This paper studies the performance of HARQ-CC over double Rayleigh channels from an information theoretic perspective. Analytical approximations are derived for the $\epsilon$ -outage capacity, the average number of transmissions, and the throughput of HARQ-CC. Moreover, we evaluate the delay experienced by Poisson-arriving packets for HARQ-CC. We provide analytical expressions for the average waiting time, the packets sojourn time, the average consumed power, and the energy efficiency. In our investigation, we take into account the impact of imperfect feedback on different performance metrics. Additionally, we explore the tradeoff between energy efficiency and the throughput. The proposed scheme is shown to maintain the outage probability below a specified threshold $\epsilon$ , which ensures the link reliability. Meanwhile, HARQ-CC adapts implicitly the transmission rate to the channel conditions such that the throughput is maximized. Our results demonstrate that HARQ-CC allows improving the achievable communication rate compared to fixed time diversity schemes. To maximize the throughput of HARQ-CC, the rate per HARQ round should be less than the rate required to meet the outage constraint. Our investigation of the performance of HARQ-CC over Rayleigh and double Rayleigh channels shows that double Rayleigh channels have a higher severity of fading and result in a larger degradation of the throughput. Our analysis reveals that HARQ with incremental redundancy achieves a larger throughput compared to HARQ-CC, whereas HARQ-CC is simpler to implement, has a lower decoding complexity, and requires less memory resources.

Development and In Vivo Performance Evaluation of 10–60-MHz Band Impulse-Radio-Based Transceiver for Deep Implantation Having 10 Mbps

Abstract: In view of the requirements for high-speed, highly reliable wireless implant communication, we have developed an implant transceiver working at a 10–60-MHz band. The developed transceiver is based on an impulse radio technology with multipulse position modulation to increase the communication speed and reliability, and utilizes the automatic equalization technique to suppress waveform distortion and intersymbol interference due to frequency-dependent tissue properties. The transmit antenna has a dimension of $2.6\,\,\text {cm} \times 1.6\,\,\text {cm} \times 1.6$ cm and a relative bandwidth of 16% by forming the radiation elements on a flexible magnetic sheet for miniaturization. Through an in vivo experiment on a living swine, we have shown the feasibility of implant communication in a depth up to 26 cm with a minimum data rate of 10 Mbps. These results demonstrate the superiority of this new technology over all others reported so far in the literature.

Recovery of Independent Sparse Sources From Linear Mixtures Using Sparse Bayesian Learning

Abstract: Classical algorithms for the multiple measurement vector (MMV) problem assume either independent columns for the solution matrix or certain models of correlation among the columns. The correlation structure in the previous MMV formulation does not capture the signals well for some applications like photoplethysmography (PPG) signal extraction where the signals are independent and linearly mixed in a certain manner. In practice, the mixtures of these signals are observed through different channels. In order to capture this structure, we decompose the solution matrix into multiplication of a sparse matrix composed of independent components, and a linear mixing matrix. We derive a new condition that guarantees a unique solution for this linear mixing MMV problem. The condition can be much less restrictive than the conditions for the typical MMV problem in previous works. We also propose two novel sparse Bayesian learning (SBL) algorithms, independent component analysis sparse Bayesian learning, (ICASBL), and fast independent component sparse Bayesian learning, which capture the linear mixture structure. Analysis of the global and local minima of the ICASBL cost function is also provided, and similar to the typical SBL cost function it is shown that the local minima are sparse and that the global minima have maximum sparsity. Experimental results show that the proposed algorithms outperform traditional approaches and can recover the signal with fewer number of measurements in the linear mixing MMV setting.

High Gain and Wideband Stacked Patch Antenna for S-Band Applications

Abstract: In this paper design of a stacked circular patch antenna is presented for high gain and wideband applications. The main radiator of this design is a circular patch antenna, which is fed by using a coupling mechanism. Wide impedance bandwidth of 40% and linear polarization of gain 9.5 dBi at the center frequency of 2.4 GHz is measured. The antenna gain is further increased by using regular period of N circular patch directors on top of the main radiator. The first director is placed in a distance of about one half of the wavelength and the next directors are placed with regular distances of about a quarter of the wavelength. The antenna gain is tuned and increased with the number of the directors in the range of 9.5–16.5 dBi with N up to seventeen. The antenna impedance change due to the added directors is adjusted by using two parasitic circular patches between the main radiator and the first director. A prototype antenna is designed, manufactured and measured. The antenna operation can be further extended using dual feed geometry in which we can obtain two orthogonal radiation patterns or circular polarizations.

Coherent query scheme for wireless backscatter communication systems with single tag

Abstract: An un-coded multi-transmitter query scheme is introduced for wireless backscatter communication systems in which M transmitters and N receivers are used for single-tag connectivity (M × 1 × N) The main idea is to harden the wireless communication channel with a tag device for high data rate readings The proposed method is designed for multipath fading channels in which the backscatter channel is a multiplicative Rayleigh A coherent transmit query scheme is used to increase the tag-reflected signals and simultaneously alter the fading statistics in the forward path by implementing a receiver feedback Full-diversity performance and array gain is achieved using the receiver diversity without requiring any tag antenna diversity Therefore, the tag device remains simple Mathematical expressions for the probability density function (PDF) of the backscatter channel are presented using closed-form equations Bit error rate (BER) simulations for the binary phase shift keying (BPSK) modulation are computed numerically Diversity gain of 10 dB is obtained by using a 2 × 1 × 1 scheme The results show that the transmit diversity for single-tag usage performs the same as the tag antenna diversity, at the expense of a moderate transmitter complexity The tag device remains intact as a requirement for the simplicity and size constraints Also, the system realization becomes more feasible due to the available space on the transmitter side to accomplish the uncorrelated forward channel conditions The feasibility study is demonstrated using software-defined radio (SDR) implementations

Wireless Channel Modeling for Leadless Cardiac Pacemaker: Effects ofVentricular Blood Volume

Abstract: In this paper, the effects of ventricular blood volume change on the cardiac wireless channel modeling at 2.4 GHz are presented for future multi-node leadless capsule pacemakers. Numerical electromagnetic simulations are implemented in the anatomical model of human body and experiments are done in a homogeneous liquid phantom. The results show that the coupling between the capsule antennas decreases with increasing blood volume, as the blood is a highly attenuating medium. The received power varies by 3.1 dB for the implant-to-implant link inside the heart and varies by 1.3 dB for the heart implant to sub-cutaneous implant, during the diastole and systole cycle. The findings are useful to develop an intra-body communication between leadless pacemakers providing timing information for synchronized pacing in a multi-chamber leadless pacemaker system.

Electromechanical Model to Predict Cardiac Resynchronization Therapy

Abstract: Cardiac resynchronization therapy (CRT) can substantially improve dyssynchronous heart failure and reduce mortality. However, one-third of the CRT patients derive no measurable benefit from CRT, due to suboptimal placement of the left ventricular (LV) lead. We introduce a pipeline for improved CRT-therapy by creating an electromechanical model using patient-specific geometric parameters allowing individualization of therapy. The model successfully mimics expected changes when variables for tension, stiffness, and conduction are entered. Changing LV pacing site had a notable effect on maximum pressure gradient (dP/dt max ) in the presence of cardiac scarring, causing non-uniform excitation propagation through the LV. Tailoring CRT to the individual requires simulations with patient-specific biventricular meshes including cardiac geometry and conductivity properties.

On Path Length Estimation for Wireless Capsule Endoscopy

Abstract: Wireless capsule endoscopy (WCE) is a non-invasive technology used for inspection of the gastrointestinal tract. Localization of the capsule is a vital part of the system enabling physicians to identify the position of anomalies. Due to intestinal motility, the positions of the intestines will change significantly within the abdominal cavity over time. However, the distance from one position to another within the intestines changes much less. In this paper a method for calculating the pathlength travelled by a WCE is proposed. The method is based on Kalman-and particle filters and is simulated using a model that approximately replicates the movement through the small intestine. The travelled distance was estimated to an accuracy within a few millimeters.

Recovery of Linearly Mixed Sparse Sources from Multiple Measurement Vectors Using L1-Minimization

Abstract: Multiple measurement vector (MMV) enables joint sparse recovery which can be applied in wide range of applications. Traditional MMV algorithms assume that the solution has independent columns or correlation among the columns. This assumption is not accurate for applications like signal estimation in photoplethysmography (PPG). In this paper, we consider a structure for the solution matrix decomposed into a sparse matrix with independent columns and a square mixing matrix. Based on this structure, we find the uniqueness condition for l 1 minimization. Moreover, an algorithm is proposed that provides a new cost function based on the new structure. It is shown that the new structure increases the recovery performance especially in low number of measurements.

Blind Source Separation Using Temporal Correlation, Non-Gaussianity and Conditional Heteroscedasticity

Abstract: Independent component analysis separates latent sources from a linear mixture by assuming sources are statistically independent. In real world applications, hidden sources are usually non-Gaussian and have dependence among samples. In such case, both attributes should be considered jointly to obtain a successful separation. To capture sample dependence, a latent source is sometimes modeled by autoregressive or moving average models with an independent and identically distributed error or residual. However, these models are limited by assuming only linear dependence among a source’s samples. This paper proposes a new blind source separation algorithm based on an autoregressive-autoregressive conditional heteroscedasticity (AR-ARCH) model, which captures linear correlations, non-Gaussianity, and squared residuals’ dependence. The AR part of the AR-ARCH model captures the correlation among samples. The ARCH part of the model captures the non-Gaussianity and nonlinear dependence among samples. The ARCH model also assumes the time-varying conditional variances for sources. We derive the Cramer Rao lower bound (CRLB) for the mixing matrix based on the AR-ARCH model. We perform simulations on both synthetic and real data. The results show that the proposed method outperforms the baseline algorithms especially for a small number of samples and approaches the CRLB.

Capacity estimation in MIMO synaptic channels

Abstract: Designing novel artificial intra-body networks and/or synthetic neurons, which interact with operating cells and compensate for malfunctioning cells, requires understanding and quantifying the information transfer in neural networks. However, the latter is not studied enough in the existing literature. Here we quantify the information rate transmitted between two neurons by analyzing Poisson Multiple-Input Multiple-Output (MIMO) synaptic channels. The results provided are intuitive and prove that multiple synapses working in cooperation improve the reliability of the neuron-to-neuron communication channel. The results serve as a progressive step in the evaluation of the performance of biological neural networks and the development of artificial cells and networks.

On the intracellular signaling of cardiomyocytes for energy efficient leadless pacemakers

Abstract: We studied the three signaling-relevant compartments of single cardiac cells (cardiomyocytes) to understand the excitation mechanisms and accordingly determine a suitable stimulation strategy. We demonstrated that stimulation/pacing with a pulse train signal may reduce power consumption. This may be of importance when applied to leadless pacemakers.

Reliable and High-Speed Implant Ultra-Wideband Communications with Transmit–Receive Diversity

Abstract: For realizing high-speed implant communications, ultra-wideband (UWB) transmission is one of the promising candidates. In this paper, an UWB transmitter diversity antenna was presented, and then the fundamental performances were numerically and experimentally evaluated. In addition, this paper discussed the path loss performance in an implant communication link a living animal. Consequently, the communication performance improvement by the implant side polarization diversity with the antenna developed in this study was evaluated and discussed. Our evaluation results demonstrated that the developed antenna can improve the communication performance by 7 dB to establish an outage rate of 0.01.