Signals, Volume 2, Issue 3 (September 2021) – 14 articles

Spatiotemporal representations learned using 3D convolutional neural networks (CNN) are currently used in state-of-the-art approaches for action-related tasks. However, 3D-CNN are notorious for being memory and compute resource intensive as compared with more simple 2D-CNN architectures. We propose to hallucinate spatiotemporal representations from a 3D-CNN teacher with a 2D-CNN student. By requiring the 2D-CNN to predict the future and intuit upcoming activity, it is encouraged to gain a deeper understanding of actions and how they evolve. The hallucination task is treated as an auxiliary task, which can be used with any other action-related task in a multitask learning setting. Thorough experimental evaluation, it is shown that the hallucination task indeed helps improve performance on action recognition, action quality assessment, and dynamic scene recognition tasks. From a practical standpoint, being able to hallucinate spatiotemporal representations without an actual 3D-CNN can enable deployment in resource-constrained scenarios, such as with limited computing power and/or lower bandwidth. We also observed that our hallucination task has utility not only during the training phase, but also during the pre-training phase. Full article

With power designers always demanding for faster power switches, electromagnetic interference has become an issue of primary concern. As known, the commutation of power transistors is the main cause of the electromagnetic noise, which can be worsened by the presence of unwanted oscillations superimposed onto the switching waveforms. This work proposes a solution to mitigate the oscillations caused by the turn-on of a power transistor by exploiting its source inductance plus an external one. In this context, an optimization method is proposed to find the optimal value of the source inductance as a trade-off between oscillation damping and power dissipation. The experimental results performed on a prototyped power converter assess the proposed technique as the spectrum of the conducted emission is attenuated by 20 dB at the oscillation frequency. With respect to traditional solution based on snubbers, the proposed solution results in a similar oscillation damping, but with a 0.5% higher power efficiency. Full article

Most IoT networks implement one-way messages from the sensor nodes to the “application host server” via a gateway. Messages from any sensor node in the network are sent when its sensor is triggered or at regular intervals as dictated by the application, such as a Smart-City deployment of LoRaWAN traps/sensors for rat detection. However, these traps can, due to the nature of this application, be moved out of signal range from their original location, or obstructed by objects, resulting in under 69% of the messages reaching the gateway. Therefore, applications of this type would benefit from control messages from the “application host server” back to the sensor nodes for enhancing their performance/connectivity. This paper has implemented a cloud-based AI engine, as part of the “application host server”, that dynamically analyses all received messages from the sensor nodes and exchanges data/enhancement back and forth with them, when necessary. Hundreds of sensor nodes in various blocked/obstructed IoT network connectivity scenarios are used to test our DXN solution. We achieved 100% reporting success if access to any blocked sensor node was possible via a neighbouring node. DXN is based on DNN and Time Series models. Full article

Generative Adversarial Networks (GANs) have gained significant attention in recent years, with impressive applications highlighted in computer vision, in particular. Compared to such examples, however, there have been more limited applications of GANs to time series modeling, including forecasting. In this work, we present the Mixture Density Conditional Generative Adversarial Model (MD-CGAN), with a focus on time series forecasting. We show that our model is capable of estimating a probabilistic posterior distribution over forecasts and that, in comparison to a set of benchmark methods, the MD-CGAN model performs well, particularly in situations where noise is a significant component of the observed time series. Further, by using a Gaussian mixture model as the output distribution, MD-CGAN offers posterior predictions that are non-Gaussian. Full article

This paper proposes an automatic image correction method for portrait photographs, which promotes consistency of facial skin color by suppressing skin color changes due to background colors. In portrait photographs, skin color is often distorted due to the lighting environment (e.g., light reflected from a colored background wall and over-exposure by a camera strobe). This color distortion is emphasized when artificially synthesized with another background color, and the appearance becomes unnatural. In our framework, we, first, roughly extract the face region and rectify the skin color distribution in a color space. Then, we perform color and brightness correction around the face in the original image to achieve a proper color balance of the facial image, which is not affected by luminance and background colors. Our color correction process attains natural results by using a guide image, unlike conventional algorithms. In particular, our guided image filtering for the color correction does not require a perfectly-aligned guide image required in the original guide image filtering method proposed by He et al. Experimental results show that our method generates more natural results than conventional methods on not only headshot photographs but also natural scene photographs. We also show automatic yearbook style photo generation as another application. Full article

Vulnerable traffic users, such as bikers and pedestrians, account for a significant number of fatalities on the roadways. Extensive research has been conducted in the literature review to identify factors to those crashes. Studying factors to those crashes is especially important in the Western state in the US, due to one of the highest fatality rates in the nation and its unique geographic conditions. The first step in identifying factors to the severity of cyclist crashes is to find the underlying factors to that type of crash, while accounting for the heterogeneity in the dataset. Various techniques such as mixed parameter or mixed effect models have been employed in the literature to account for the heterogeneity of the dataset. In the mixed effect model, often the random effect parameter has been assigned subjectively, and based on some attributes and engineering intuitions. Those assignments are expected to account for the heterogeneity in the dataset and enhancement of the model fit. However, a question might arise whether those factors could account for an optimum amount of the heterogeneity in the dataset. A more reasonable way might be to let the algorithm such as the finite mixture model (FMM) to identify those clusters based on parameters of the Gaussian model, means and covariance matrices of the dataset, and allocate each observation to the related clusters. Thus, in this study, to capture optimum amount of heterogeneity, first we implemented the finite mixture model in the context of maximum likelihood, due the label switching issue of the method in the context of the Bayesian method. After assignment of the parameters to the observation, the main method of Hamiltonian Monte Carlo (HMC) with random effect was implemented. The results highlighted a significant improvement in the model fit, in terms of Widely Applicable Information Criterion (WAIC). The results of this study highlighted factors such as older biker age, increased number of lanes, nighttime travelling, increased posted speed limit and driving while under emotional conditions are some factors contributing to an increased severity of bikers’ crash severity. Extensive discussion has been made regarding the methodological algorithms and model parameters estimations. Full article

This paper describes an automatic drum transcription (ADT) method that directly estimates a tatum-level drum score from a music signal in contrast to most conventional ADT methods that estimate the frame-level onset probabilities of drums. To estimate a tatum-level score, we propose a deep transcription model that consists of a frame-level encoder for extracting the latent features from a music signal and a tatum-level decoder for estimating a drum score from the latent features pooled at the tatum level. To capture the global repetitive structure of drum scores, which is difficult to learn with a recurrent neural network (RNN), we introduce a self-attention mechanism with tatum-synchronous positional encoding into the decoder. To mitigate the difficulty of training the self-attention-based model from an insufficient amount of paired data and to improve the musical naturalness of the estimated scores, we propose a regularized training method that uses a global structure-aware masked language (score) model with a self-attention mechanism pretrained from an extensive collection of drum scores. The experimental results showed that the proposed regularized model outperformed the conventional RNN-based model in terms of the tatum-level error rate and the frame-level F-measure, even when only a limited amount of paired data was available so that the non-regularized model underperformed the RNN-based model. Full article

We developed a hearing assistance system that enables hearing-impaired people to track the horizontal movement of a single sound source. The movement of the sound source is presented to the subject by vibrating vibrators on both shoulders according to the distance to and direction of the sound source, which are estimated from the acoustic signals detected by microphones attached to both ears. We presented the direction of and distance to the sound source to the subject by changing the ratio of the intensity of the two vibrators according to the direction and by increasing the intensity the closer the person got to the sound source. The subject could recognize the approaching sound source as a change in the vibration intensity by turning their face in the direction where the intensity of both vibrators was equal. The direction of the moving sound source can be tracked with an accuracy of less than 5° when an analog vibration pattern is added to indicate the direction of the sound source. By presenting the direction of the sound source with high accuracy, it is possible to show subjects the approach and departure of a sound source. Full article

This paper presents a novel adaptive probabilistic algorithm to identify damage characteristics by integrating the use of the frequency response function with an optimization approach. The proposed algorithm evaluates the probability of damage existence and determines salient details such as damage location and damage severity in a probabilistic manner. A multistage sequence is used to determine the probability of damage parameters including crack depth and crack location while minimizing uncertainties. A simply supported beam with an open edge crack was used to demonstrate the application of the algorithm for damage detection. The robustness of the algorithm was tested by incorporating varying levels of noise into the frequency response. All simulation results show successful detection of damage with a relatively high probability even in the presence of noise. Results indicate that the probabilistic algorithm could have significant advantages over conventional deterministic methods since it has the ability to avoid yielding false negatives that are quite common among deterministic damage detection techniques. Full article

Voice transformation, for example, from a male speaker to a female speaker, is achieved here using a two-level dynamic warping algorithm in conjunction with an artificial neural network. An outer warping process which temporally aligns blocks of speech (dynamic time warp, DTW) invokes an inner warping process, which spectrally aligns based on magnitude spectra (dynamic frequency warp, DFW). The mapping function produced by inner dynamic frequency warp is used to move spectral information from a source speaker to a target speaker. Artifacts arising from this amplitude spectral mapping are reduced by reconstructing phase information. Information obtained by this process is used to train an artificial neural network to produce spectral warping information based on spectral input data. The performance of the speech mapping compared using Mel-Cepstral Distortion (MCD) with previous voice transformation research, and it is shown to perform better than other methods, based on their reported MCD scores. Full article

Inaccurate estimates of the linear prediction coefficient (LPC) and noise variance introduce bias in Kalman filter (KF) gain and degrade speech enhancement performance. The existing methods propose a tuning of the biased Kalman gain, particularly in stationary noise conditions. This paper introduces a tuning of the KF gain for speech enhancement in real-life noise conditions. First, we estimate noise from each noisy speech frame using a speech presence probability (SPP) method to compute the noise variance. Then, we construct a whitening filter (with its coefficients computed from the estimated noise) to pre-whiten each noisy speech frame prior to computing the speech LPC parameters. We then construct the KF with the estimated parameters, where the robustness metric offsets the bias in KF gain during speech absence of noisy speech to that of the sensitivity metric during speech presence to achieve better noise reduction. The noise variance and the speech model parameters are adopted as a speech activity detector. The reduced-biased Kalman gain enables the KF to minimize the noise effect significantly, yielding the enhanced speech. Objective and subjective scores on the NOIZEUS corpus demonstrate that the enhanced speech produced by the proposed method exhibits higher quality and intelligibility than some benchmark methods. Full article

This article presents the advantages and limitations of a recently developed Ultrasonic Guided Wave Leakage (UGWL) method in comparison to the well-known Half-Cell Potential (HCP) method in their ability to detect corrosion in reinforced concrete (RC) bridge decks. This research also establishes a correlation between UGWL data and chloride content in concrete RC slabs. Concrete slabs submerged in a 10% NaCl solution were monitored using both methods over a period of six months. The chloride content from the three cores (0.84, 0.55, and 0.18%) extracted from the slab after the 6-month long process all exceeded the chloride threshold values suggested in ACI 318, which is 0.05 to 0.1% by weight of concrete. Further, the UGWL method detected changes due to corrosion approximately 21 days earlier than the HCP method. Full article

During the pregnancy, fetal electrocardiogram (FECG) is deployed to analyze fetal heart rate (FHR) of the fetus to indicate the growth and health of the fetus to determine any abnormalities and prevent diseases. The fetal electrocardiogram monitoring can be carried out either invasively by placing the electrodes on the scalp of the fetus, involving the skin penetration and the risk of infection, or non-invasively by recording the fetal heart rate signal from the mother’s abdomen through a placement of electrodes deploying portable, wearable devices. Non-invasive fetal electrocardiogram (NIFECG) is an evolving technology in fetal surveillance because of the comfort to the pregnant women and being achieved remotely, specifically in the unprecedented circumstances such as pandemic or COVID-19. Textiles have been at the heart of human technological progress for thousands of years, with textile developments closely tied to key inventions that have shaped societies. The relatively recent invention of smart textiles is set to push boundaries again and has already opened the potential for garments relevant to medicine, and health monitoring. This paper aims to discuss the different technologies and methods used in non-invasive fetal electrocardiogram (NIFECG) monitoring as well as the potential and future research directions of NIFECG in the smart textiles area. Full article

Brain source localization has been consistently implemented over the recent years to elucidate complex brain operations, pairing the high temporal resolution of the EEG with the high spatial estimation of the estimated sources. This review paper aims to present the basic principles of Electrical source imaging (ESI) in the context of the recent progress for solving the forward and the inverse problems, and highlight the advantages and limitations of the different approaches. As such, a synthesis of the current state-of-the-art methodological aspects is provided, offering a complete overview of the present advances with regard to the ESI solutions. Moreover, the new dimensions for the analysis of the brain processes are indicated in terms of clinical and cognitive ESI applications, while the prevailing challenges and limitations are thoroughly discussed, providing insights for future approaches that could help to alleviate methodological and technical shortcomings. Full article