LBTS-Net: A fast and accurate CNN model for brain tumour segmentation.

An accurate tumour segmentation in brain images is a complicated task due to the complext structure and irregular shape of the tumour. In this letter, our contribution is twofold: (1) a lightweight brain tumour segmentation network (LBTS-Net) is proposed for a fast yet accurate brain tumour segmentation; (2) transfer learning is integrated within the LBTS-Net to fine-tune the network and achieve a robust tumour segmentation. To the best of knowledge, this work is amongst the first in the literature which proposes a lightweight and tailored convolution neural network for brain tumour segmentation. The proposed model is based on the VGG architecture in which the number of convolution filters is cut to half in the first layer and the depth-wise convolution is employed to lighten the VGG-16 and VGG-19 networks. Also, the original pixel-labels in the LBTS-Net are replaced by the new tumour labels in order to form the classification layer. Experimental results on the BRATS2015 database and comparisons with the state-of-the-art methods confirmed the robustness of the proposed method achieving a global accuracy and a Dice score of 98.11% and 91%, respectively, while being much more computationally efficient due to containing almost half the number of parameters as in the standard VGG network.

Real-time physiological tremor estimation using recursive singular spectrum analysis.

Physiological hand tremor causes undesirable vibration of hand-held surgical instruments which results in imprecisions and poor surgical outcomes. Existing tremor cancellation algorithms are based on detection of the tremulous component from the whole motion; then adding an anti-phase tremor signal to the whole motion to cancel it out. These techniques are based on adaptive filtering algorithms which need a reference signal that is highly correlated with the actual tremor signal. Hence, such adaptive approaches use a non-linear phase filter to pre-filter the tremor signal either offline or in real-time. However, pre-filtering causes unnecessary delays and non-linear phase distortions as the filter has frequency selective delays. Consequently, the anti-phase tremor signal cannot be generated accurately which results in poor tremor cancellation. In this paper, we present a new technique based on singular spectrum analysis (SSA) and its recursive version, that is, recursive singular spectrum analysis (RSSA). These algorithms decompose the whole motion into dominant voluntary components corresponding to larger eigenvalues and oscillatory tremor components having smaller eigenvalues. By selecting a group of specific decomposed signals based on their eigenvalues and spectral range, both voluntary and tremor signals can be reconstructed accurately. We test the SSA and RSSA algorithms using recorded tremor data from five novice subjects. This new approach shows the tremor signal can be estimated from the whole motion with an accuracy of up to 85% offline. In real-time, tolerating a delay of ≈ 72ms, the tremor signal can be estimated with at least 70% accuracy. This delay is found to be one-tenth of the delay caused by a conventional linear-phase bandpass filter to achieve similar performance in real-time.

Breathing Analysis Using Thermal and Depth Imaging Camera Video Records.

The paper is devoted to the study of facial region temperature changes using a simple thermal imaging camera and to the comparison of their time evolution with the pectoral area motion recorded by the MS Kinect depth sensor. The goal of this research is to propose the use of video records as alternative diagnostics of breathing disorders allowing their analysis in the home environment as well. The methods proposed include (i) specific image processing algorithms for detecting facial parts with periodic temperature changes; (ii) computational intelligence tools for analysing the associated videosequences; and (iii) digital filters and spectral estimation tools for processing the depth matrices. Machine learning applied to thermal imaging camera calibration allowed the recognition of its digital information with an accuracy close to 100% for the classification of individual temperature values. The proposed detection of breathing features was used for monitoring of physical activities by the home exercise bike. The results include a decrease of breathing temperature and its frequency after a load, with mean values -0.16 °C/min and -0.72 bpm respectively, for the given set of experiments. The proposed methods verify that thermal and depth cameras can be used as additional tools for multimodal detection of breathing patterns.

Privacy-preserving clinical decision support system using Gaussian kernel-based classification.

A clinical decision support system forms a critical capability to link health observations with health knowledge to influence choices by clinicians for improved healthcare. Recent trends toward remote outsourcing can be exploited to provide efficient and accurate clinical decision support in healthcare. In this scenario, clinicians can use the health knowledge located in remote servers via the Internet to diagnose their patients. However, the fact that these servers are third party and therefore potentially not fully trusted raises possible privacy concerns. In this paper, we propose a novel privacy-preserving protocol for a clinical decision support system where the patients' data always remain in an encrypted form during the diagnosis process. Hence, the server involved in the diagnosis process is not able to learn any extra knowledge about the patient's data and results. Our experimental results on popular medical datasets from UCI-database demonstrate that the accuracy of the proposed protocol is up to 97.21% and the privacy of patient data is not compromised.

An online one class support vector machine-based person-specific fall detection system for monitoring an elderly individual in a room environment.

In this paper, we propose a novel computer vision-based fall detection system for monitoring an elderly person in a home care, assistive living application. Initially, a single camera covering the full view of the room environment is used for the video recording of an elderly person's daily activities for a certain time period. The recorded video is then manually segmented into short video clips containing normal postures, which are used to compose the normal dataset. We use the codebook background subtraction technique to extract the human body silhouettes from the video clips in the normal dataset and information from ellipse fitting and shape description, together with position information, is used to provide features to describe the extracted posture silhouettes. The features are collected and an online one class support vector machine (OCSVM) method is applied to find the region in feature space to distinguish normal daily postures and abnormal postures such as falls. The resultant OCSVM model can also be updated by using the online scheme to adapt to new emerging normal postures and certain rules are added to reduce false alarm rate and thereby improve fall detection performance. From the comprehensive experimental evaluations on datasets for 12 people, we confirm that our proposed person-specific fall detection system can achieve excellent fall detection performance with 100% fall detection rate and only 3% false detection rate with the optimally tuned parameters. This work is a semiunsupervised fall detection system from a system perspective because although an unsupervised-type algorithm (OCSVM) is applied, human intervention is needed for segmenting and selecting of video clips containing normal postures. As such, our research represents a step toward a complete unsupervised fall detection system.

A posture recognition based fall detection system for monitoring an elderly person in a smart home environment.

We propose a novel computer vision based fall detection system for monitoring an elderly person in a home care application. Background subtraction is applied to extract the foreground human body and the result is improved by using certain post-processing. Information from ellipse fitting and a projection histogram along the axes of the ellipse are used as the features for distinguishing different postures of the human. These features are then fed into a directed acyclic graph support vector machine (DAGSVM) for posture classification, the result of which is then combined with derived floor information to detect a fall. From a dataset of 15 people, we show that our fall detection system can achieve a high fall detection rate (97.08%) and a very low false detection rate (0.8%) in a simulated home environment.

Motion-compensated noncontact imaging photoplethysmography to monitor cardiorespiratory status during exercise.

With the advance of computer and photonics technology, imaging photoplethysmography [(PPG), iPPG] can provide comfortable and comprehensive assessment over a wide range of anatomical locations. However, motion artifact is a major drawback in current iPPG systems, particularly in the context of clinical assessment. To overcome this issue, a new artifact-reduction method consisting of planar motion compensation and blind source separation is introduced in this study. The performance of the iPPG system was evaluated through the measurement of cardiac pulse in the hand from 12 subjects before and after 5 min of cycling exercise. Also, a 12-min continuous recording protocol consisting of repeated exercises was taken from a single volunteer. The physiological parameters (i.e., heart rate, respiration rate), derived from the images captured by the iPPG system, exhibit functional characteristics comparable to conventional contact PPG sensors. Continuous recordings from the iPPG system reveal that heart and respiration rates can be successfully tracked with the artifact reduction method even in high-intensity physical exercise situations. The outcome from this study thereby leads to a new avenue for noncontact sensing of vital signs and remote physiological assessment, with clear applications in triage and sports training.

Graphic patterns of cortical functional connectivity of depressed patients on the basis of EEG measurements.

Considerable evidences have shown a decrease of neuronal activity in the left frontal lobe of depressed patients, but the underlying cortical network is still unclear. The present study intends to investigate the conscious-state brain network patterns in depressed patients compared with control individuals. Cortical functional connectivity is quantified by the partial directed coherence (PDC) analysis of multichannel EEG signals from 12 depressed patients and 12 healthy volunteers. The corresponding PDC matrices are first converted into unweighted graphs by applying a threshold to obtain the topographic property in-degree (K(in)). A significantly larger K(in) in the left hemisphere is identified in depressed patients, while a symmetric pattern is found in the control group. Another two topographic measures, i.e., clustering coefficients (C) and characteristic path length (L), are obtained from the original weighted PDC digraphs. Compared with control individuals, significantly smaller C and L are revealed in the depression group, indicating a random network-like architecture due to affective disorder. This study thereby provides further support for the presence of a hemispheric asymmetry syndrome in the depressed patients. More importantly, we present evidence that depression is characterized by a loss of optimal small-world network characteristics in conscious state.

Sequential blind source extraction for quasi-periodic signals with time-varying period.

A novel second-order-statistics-based sequential blind extraction algorithm for blind extraction of quasi-periodic signals, with time-varying period, is introduced in this paper. Source extraction is performed by sequentially converging to a solution that effectively diagonalizes autocorrelation matrices at lags corresponding to the time-varying period, which thereby explicitly exploits a key statistical nonstationary characteristic of the desired source. The algorithm is shown to have fast convergence and yields significant improvement in signal-to-interference ratio as compared to when the algorithm assumes a fixed period. The algorithm is further evaluated on the problem of separation of a heart sound signal from real-world lung sound recordings. Separation results confirm the utility of the introduced approach, and listening tests are employed to further corroborate the results.

Removal of the eye-blink artifacts from EEGs via STF-TS modeling and robust minimum variance beamforming.

A novel scheme for the removal of eye-blink (EB) artifacts from electroencephalogram (EEG) signals based on a novel space-time-frequency (STF) model of EEGs and robust minimum variance beamformer (RMVB) is proposed. In this method, in order to remove the artifact, the RMVB is provided with a priori information, namely, an estimation of the steering vector corresponding to the point source EB artifact. The artifact-removed EEGs are subsequently reconstructed by deflation. The a priori knowledge, the vector corresponding to the spatial distribution of the EB factor, is identified using the STF model of EEGs, provided by the parallel factor analysis (PARAFAC) method. In order to reduce the computational complexity present in the estimation of the STF model using the three-way PARAFAC, the time domain is subdivided into a number of segments, and a four-way array is then set to estimate the STF-time/segment (TS) model of the data using the four-way PARAFAC. The correct number of the factors of the STF model is effectively estimated by using a novel core consistency diagnostic- (CORCONDIA-) based measure. Subsequently, the STF-TS model is shown to closely approximate the classic STF model, with significantly lower computational cost. The results confirm that the proposed algorithm effectively identifies and removes the EB artifact from raw EEG measurements.

Monaural room acoustic parameters from music and speech.

This paper compares two methods for extracting room acoustic parameters from reverberated speech and music. An approach which uses statistical machine learning, previously developed for speech, is extended to work with music. For speech, reverberation time estimations are within a perceptual difference limen of the true value. For music, virtually all early decay time estimations are within a difference limen of the true value. The estimation accuracy is not good enough in other cases due to differences between the simulated data set used to develop the empirical model and real rooms. The second method carries out a maximum likelihood estimation on decay phases at the end of notes or speech utterances. This paper extends the method to estimate parameters relating to the balance of early and late energies in the impulse response. For reverberation time and speech, the method provides estimations which are within the perceptual difference limen of the true value. For other parameters such as clarity, the estimations are not sufficiently accurate due to the natural reverberance of the excitation signals. Speech is a better test signal than music because of the greater periods of silence in the signal, although music is needed for low frequency measurement.

A robust minimum variance beamforming approach for the removal of the eye-blink artifacts from EEGs.

In this paper a novel scheme for the removal of eye-blink (EB) artifacts from electroencephalogram (EEG) signals based on the robust minimum variance beamformer (RMVB) is proposed. In this method, in order to remove the artifact, the RMVB is provided with a priori information, i.e., an estimation of the steering vector corresponding to the point source EB artifact. The artifact-removed EEGs are subsequently reconstructed by deflation. The a priori knowledge, namely the vector corresponding to the spatial distribution of the EB factor, is identified using a novel space-time-frequency-time/segment (STF-TS) model of EEGs, provided by a four-way parallel factor analysis (PARAFAC) approach. The results demonstrate that the proposed algorithm effectively identifies and removes the EB artifact from raw EEG measurements.

Underdetermined blind source separation of temporomandibular joint sounds.

The underdetermined blind source separation problem using a filtering approach is addressed. An extension of the FastICA algorithm is devised which exploits the disparity in the kurtoses of the underlying sources to estimate the mixing matrix and thereafter achieves source recovery by employing the ll-norm algorithm. Besides, we demonstrate how promising FastICA can be to extract the sources. Furthermore, we illustrate how this scenario is particularly appropriate for the separation of temporomandibular joint (TMJ) sounds.

Fetal electrocardiogram extraction by sequential source separation in the wavelet domain.

This paper addresses the problem of fetal electrocardiogram extraction using blind source separation (BSS) in the wavelet domain. A new approach is proposed, which is particularly advantageous when the mixing environment is noisy and time-varying, and that is shown, analytically and in simulation, to improve the convergence rate of the natural gradient algorithm. The distribution of the wavelet coefficients of the source signals is then modeled by a generalized Gaussian probability density, whereby in the time-scale domain the problem of selecting appropriate nonlinearities when separating mixtures of both sub- and super-Gaussian signals is mitigated, as shown by experimental results.

On the efficient estimation of blood velocities.

Pulsed wave (PW) Doppler ultrasound systems are commonly used to examine blood flow dynamics and the technique plays a very important role in numerous diagnostic applications. Commonly, narrow-band PW systems estimate the blood velocity using an autocorrelation-based estimator. Herein, we examine a recently proposed hybrid frequency estimator, and via extensive numerical simulations using simulated blood scatterers show the achievable performance gain of this method as compared to the traditional approach.

A modified ventral fixation for surgical management of atlantoaxial subluxation in 19 dogs.

OBJECTIVE: To describe a modified ventral stabilization technique for surgical management of atlantoaxial subluxation in dogs and to evaluate the outcome. STUDY DESIGN: Retrospective clinical study. SAMPLE POPULATION: Nineteen client-owned dogs. METHODS: Medical records of 19 dogs with a radiographic diagnosis of atlantoaxial subluxation surgically managed by a modified ventral fixation technique (cortical screws, Kirschner wires, polymethylmethacrylate) were reviewed. Data on pre- and post-operative neurologic status, surgical technique, and complications were retrieved. Follow-up evaluation was performed at approximately 1 month. Telephone interview of the owner was used for long-term assessment (median follow-up for 17 surviving dogs was 10.5 months). RESULTS: Adequate reduction and stabilization was achieved in all dogs based on radiographic assessment immediately after surgery. Improved neurologic outcome occurred in 16 dogs at 1 month and in 15 dogs at follow-up; 2 dogs died of post-operative complications within 24 hours of surgery. One dog was euthanatized at the owners' request because of recurrent neck pain associated with implant failure after 1 month. Two dogs required surgery to remove broken and migrated implants, but further stabilization was not necessary. CONCLUSIONS: Adequate stabilization and improved neurologic outcome was achieved in most dogs. However, on account of the small size of the study and the variable neurologic signs of the dogs on admission, the surgical technique described could not be compared to those previously reported. CLINICAL RELEVANCE: The surgical technique described is an effective means of surgical treatment for atlantoaxial subluxation.