Depression is a common mental disorder that is usually addressed by outpatient treatments that favour patients’ inclusion in the society. This raises the need for tools to remotely monitor the emotional state of the patients, which can be carried out via telephone or the Internet using speech processing approaches. However, these strategies lead to privacy concerns caused by the transmission of the patients’ speech and its subsequent storage in servers. The use of speech de-identification to protect the privacy of these patients seems straightforward, but the influence of this procedure in the manifestation of the disease in the patients’ speech has not been addressed yet. Hence, this study evaluates the performance of an automatic depression level estimation system when dealing with original and de-identified speech, in order to analyse the influence of the de-identification procedure in the detection of depression. Two de-identification approaches based on voice transformation via frequency warping and amplitude scaling are assessed, which can be applied to any speaker without additional training. Experiments carried out in the framework of the audio/visual emotion challenge 2014 show that the proposed de-identification approaches achieve promising de-identification results at the expense of a slight degradation of depression detection.

Thanks to advancements in smart mobile devices and social media platforms, sharing photos and experiences has significantly bridged the authors’ lives, allowing them to stay connected despite distance and other barriers. Most approaches to protect image visual privacy focus on encrypting or permuting image data, which generate unreadable image or highly distorted visual effect and therefore may not be in users best interest from both usage and perception perspectives. In this study, the authors propose secure JPEG transmorphing, a framework for protecting image visual privacy in a secure, reversible, and highly flexible and personalised manner. Secure JPEG transmorphing allows one to apply arbitrary regional visual manipulation on image regions of interests (ROIs), while secretly preserving the information about the original ROIs in application segments (APPn markers) of the visually obfuscated JPEG image. Objective and subjective experiments have been performed and results indicate that the proposed protection scheme provides near lossless image reconstruction, controllable level of file size expansion, good degree of privacy protection and especially better subjective pleasantness.

This study presents a novel approach that extends face de-identification from person-specific (closed) sets of facial images to open sets of video frames. Inspired by the previous work in facial expression transfer, the authors have introduced an ‘identity shift’ to ensure identity consistency within a de-identified video sequence. The ‘identity shift’ is derived from the first video frame of a person and is then applied in the de-identification of all subsequent frames of the same person. Experimental results show that video frames that are originally associated with the same person will remain related to a common new identity after the application of the proposed approach. In addition, the proposed approach is able to achieve privacy protection as well as preservation of dynamic facial expressions. Finally, MATLAB implementation of the approach has confirmed its potential to operate in real time at the highest standard video frame rate.

Face deidentification is an active topic amongst privacy and security researchers. Early deidentification methods relying on image blurring or pixelisation have been replaced in recent years with techniques based on formal anonymity models that provide privacy guaranties and retain certain characteristics of the data even after deidentification. The latter aspect is important, as it allows the deidentified data to be used in applications for which identity information is irrelevant. In this work, the authors present a novel face deidentification pipeline, which ensures anonymity by synthesising artificial surrogate faces using generative neural networks (GNNs). The generated faces are used to deidentify subjects in images or videos, while preserving non-identity-related aspects of the data and consequently enabling data utilisation. Since generative networks are highly adaptive and can utilise diverse parameters (pertaining to the appearance of the generated output in terms of facial expressions, gender, race etc.), they represent a natural choice for the problem of face deidentification. To demonstrate the feasibility of the authors’ approach, they perform experiments using automated recognition tools and human annotators. Their results show that the recognition performance on deidentified images is close to chance, suggesting that the deidentification process based on GNNs is effective.

High dynamic range (HDR) imaging has been developed for improved visual representation by capturing a wide range of luminance values. Owing to its properties, HDR content might lead to a larger privacy intrusion, requiring new methods for privacy protection. Previously, false colours were proved to be effective for assuring privacy protection for low dynamic range (LDR) images. In this work, the reliability of false colours when used for privacy protection of HDR images represented by tone-mapping operators (TMOs) is studied. Two different TMO techniques are tested, a simple TMO based on the Gamma transform and a more complex local TMO. Moreover, two false colour palettes are also tested, and are applied to images that result from both TMOs and also to an LDR image that represents the centre exposure in the image sequence used to create the HDR image. The degree of privacy protection is analysed through both a subjective test using crowdsourcing and an objective test using face recognition algorithms. It is concluded that the application of the two studied false colour palettes reduces the recognition accuracy with respect to both tests.

The authors introduce a system for person de-identification in video data that de-identifies biometric and non-biometric features, namely faces, hairstyles and clothing colours. The authors’ system detects human faces and silhouettes in the input video and replaces the detected faces with random synthesised faces obtained using a deep convolutional generative adversarial network. Alternative hairstyles are rendered over the synthesised faces, and the human silhouette is recoloured so that skin hues are preserved and clothing hues are altered. Through the use of artificially synthesised faces that look realistic, they ensure that the de-identified image looks natural and at the same time avoid ethical and legal considerations present when using real face images as replacement faces. As they address non-biometric feature de-identification, their system offers a considerably higher level of privacy protection than commonly employed solutions that use simple image processing techniques such as blurring. Qualitative and quantitative evaluation suggests that their system produces de-identified images that look natural, at the same time being resistant to re-identification attacks.

In this study, the authors propose the seismic trace noise reduction by wavelets and double threshold estimation method (STNRW), that is based on the discrete wavelet transform, estimates two thresholds instead of the one threshold estimation of the traditional methods. The authors verify the robustness of the method proving that the probability of classification error for a noisy wavelet coefficient decreases, as the length of the signal increases. The authors perform Monte Carlo simulations considering eight seismic traces obtained from astsa R package with different signal-noise-to-ratio (SNR) values in order to compare the performance of the new method with three denoising methods well-known in the literature. The results show that the STNRW method is efficient.

Heart rate (HR) is one of the vital signs in healthcare monitoring systems. The traditional medical devices are very expensive and the accessibility of these devices is limited. So the objective is to propose an efficient framework to monitor the HR at on-demand and continuous modes during rest conditions, which are a typical use case. The proposed framework is based on combining the time-domain (TD) and frequency-domain (FD) analysis. This hybrid approach would handle the different signal conditions to improve the results. The proposed method is compared with the standard ECG device and the real-time results show that the HR is calculated accurately with an average pass percentage of 97.85, mean absolute percentage error of 1.99%, mean absolute error of 1.51 BPM and reference closeness factor of 0.977. The reliable on-demand HR output from the proposed method comes in 6th second, which is the minimum possible time. The results show that the proposed method is efficient in estimating the HR and the results are comparable with other methods. So the proposed method can be targeted to wearable devices to accurately monitor the HR at rest conditions.

Accurate estimation of power signal frequency is an important requirement for many application areas that include system protection, energy quality monitoring and instrumentation. Though significant efforts have been made since long to develop potent algorithms for accurate estimation of power signal frequency, still their accuracy and convergence speed are a challenge under sudden frequency drift and variations. Therefore, this study focuses on a low complexity adaptive linear element filter using quadratic signal model, whose parameters are adjusted using a fast variable step size fuzzy logic-based learning algorithm to provide better convergence and noise rejection properties for the estimation of frequency from noisy and distorted signals. In addition, the new filter has also been implemented on a field programmable gate array hardware and Xilinx 14.2 with Sysgen software for the tracking of dynamic signal parameters in single and three phase power networks. Various numerical and experimental results are addressed for estimation of frequency of time varying sinusoids.

An orthogonal transformation is applied to the cubature quadrature (CQ) points which are used to approximate the intractable integrals appeared during the estimation of states of a non-linear system. With the help of a theorem, it is shown that the transformed points reduce the higher order moments (HOM) appeared while approximating the mean and covariance. As the HOM act as residue, the proposed method results in an improved filtering accuracy. Further, the proposition is extended in delayed measurement environment. The proposed method has been implemented for two different simulation problems. The simulation results reveal that at the same computational load, the filter with transformed CQ points provides higher accuracy compared with the filter with ordinary CQ points. The performance has been verified in the delayed measurement environment and an improved accuracy has been reported.

This study concerns the issue of jointly enhancing noise robustness and promoting signal sparsity in Sparse Bayesian Learning (SBL), which aims at addressing the performance deficiency of sparse signal recovery due to uninformative data with low signal-to-noise ratios. In particular, the authors propose a hierarchical prior noise model with a signal-dependent parametrisation and incorporate it into developing the robust SBL algorithms for sparse signal recovery. The main contribution of the proposed approach is twofold. The first is the new consideration of noise-robustness enhancement in building SBL algorithms, which devotes to noise awareness in counteracting outliers in measurements. Specifically, the idea of signal-sparsity enforcing is extended to build a Least Absolute Deviation like loss criterion with the proposed hierarchical prior model of measurement noise. The second is the novelty of using the signal-dependent parametrisation in the proposed noise model. Indeed, the signal-dependent mechanism plays an indispensable role in producing the reliable noise parameter estimation jointly with updating signal model parameters under the fast SBL framework. In addition to numerical simulation studies, the real-life application of radio tomographic imaging is presented to validate the proposed approach.

This study proposes a novel subspace method to estimate frequencies of two sinusoids embedded in noise. The estimation process is basically composed of three main steps. Two principal eigenvectors of the autocorrelation matrix for the received vector are first found. Then, the authors use the elements of these eigenvectors to form a set of linear and non-linear equations to solve an ambiguous matrix linked two signal subspaces which contain the information of frequencies. With the product of the matrix formed from these two eigenvectors and the solved ambiguous matrix, they can finally find the frequencies. In addition, they also combine an interleaving technique with the proposed method to improve the estimation performance for two close sinusoidal frequencies. Simulation results for synthetic data and practical vital signals are used to demonstrate the performance of the proposed method. The demonstrated results show that the proposed method is feasible for frequencies estimation of two sinusoids, and it can be applied to the case of very close sinusoidal frequencies.

In many applications like surveillance, it is essential to detect the presence of specific objects by seeing an image or video without being concerned about details of the scene. The reconstruction algorithms proposed in the compressive sensing literature try to iteratively reconstruct the full image. Hence, those are computationally expensive. If some prior knowledge of the object is available, then a closed-form reconstruction algorithm can be formulated. The goal is to reconstruct the object of interest efficiently without being bothered about the quality of reconstruction of the scene. To address this situation, a constraint is formulated and incorporated into linear minimum mean square error (LMMSE) estimator to form a closed-form solution. This compact solution is capable of meaningfully reconstructing the object relative to the scene. In the proposed method, an ill-conditioned matrix inversion problem has been faced and overcome by regularization method. To boost the speed of the algorithm, a modified Euler method is proposed for finding the regularization parameter. To further speedup the reconstruction process, larger images are divided into several pieces and each piece is reconstructed separately using constrained LMMSE. For a given number of measurements, the simulation-based results demonstrate acceptable quality of reconstruction with minimal computational effort.