In this paper we address the problem of hallucinating high-resolution facial images from low-resolution inputs at high magnification factors. We approach this task with convolutional neural networks (CNNs) and propose a novel (deep) face hallucination model that incorporates identity priors into the learning procedure. The model consists of two main parts: i) a cascaded super-resolution network that upscales the low resolution facial images, and ii) an ensemble of face recognition models that act as identity priors for the super-resolution network during training. Different from most competing super-resolution techniques that rely on a single model for upscaling (even with large magnification factors), our network uses a cascade of multiple SR models that progressively upscale the low-resolution images using steps of 2×. This characteristic allows us to apply supervision signals (target appearances) at different resolutions and incorporate identity constraints at multiple-scales. The proposed C-SRIP model (Cascaded Super Resolution with Identity Priors) is able to upscale (tiny) low-resolution images captured in unconstrained conditions and produce visually convincing results for diverse low-resolution inputs. We rigorously evaluate the proposed model on the Labeled Faces in the Wild (LFW), Helen and CelebA datasets and report superior performance compared to the existing state-of-the-art. This research was selected among the top-10 research achievements of the University of Ljubljana in the calendar year 2019 and also received the 2019 Max Snijder award, which is given out annually by the European Association of Biometrics (EAB) for the research work conducted within a PhD thesis in Europe.
COBISS.SI-ID: 12800852
The area of ocular biometrics is among the most popular branches of biometric recognition technology. This area has long been dominated by iris recognition research, while other ocular modalities such as the periocular region or the vasculature of the sclera have received significantly less attention in the literature. Consequently, ocular modalities beyond the iris are not well studied and their characteristics are today still not as well understood. While recent needs for more secure authentication schemes have considerably increased the interest in competing ocular modalities, progress in these areas is still held back by the lack of publicly available datasets that would allow for more targeted research into specific ocular characteristics next to the iris. In this paper, we aim to bridge this gap for the case of sclera biometrics and introduce a novel dataset designed for research into ocular biometrics and most importantly for research into the vasculature of the sclera. Our dataset, called Sclera Blood Vessels, Periocular and Iris (SBVPI), is, to the best of our knowledge, the first publicly available dataset in the world designed specifically with research in sclera biometrics in mind. The dataset contains high-quality RGB ocular images, captured in the visible spectrum, belonging to 55 subjects. Unlike competing datasets, it comes with manual markups of various eye regions, such as the iris, pupil, canthus or eyelashes and a detailed pixel-wise annotation of the complete sclera vasculature for a subset of the images. Additionally, the datasets ship with gender and age labels. The unique characteristics of the dataset allow us to study aspects of sclera biometrics technology that have not been studied before in the literature (e.g. vasculature segmentation techniques) as well as issues that are of key importance for practical recognition systems. Thus, next to the SBVPI dataset we also present in this paper a comprehensive investigation into sclera biometrics and the main covariates that affect the performance of sclera segmentation and recognition techniques, such as gender, age, gaze direction or image resolution. Our experiments not only demonstrate the usefulness of the newly introduced dataset, but also contribute to a better understanding of sclera biometrics in general.
COBISS.SI-ID: 1538534595
The aim of this paper is to present a new analytical model developed for open-circuit magnetic field calculation in U-shape interior permanent magnet (IPM) machine. The model is developed based on the 2-D subdomain-model approach by solving Poisson’s and Laplace’s equations. The comprehensive theoretical derivations of the final analytical equations for the magnetic flux density distribution in the air-gap and in the permanent magnets are explained in detail. We also provide the calculation algorithm in form of a flow diagram suitable for engi- neers to compute the radial and the tangential component of the magnetic flux density in all subdomains of slotless brushless machines with any possible number of rotor pole pairs of U-shaped IPMs. A slightly modified algorithm also applies to V-shaped IPMs.
COBISS.SI-ID: 12682068
The paper presents a novel measuring system for detecting air gap anomalies in direct-drive electrical motors. A novel measuring method is proposed using a sensory system integrated in a motor air gap to measure its width directly in either a static or dynamic state. It uses an optical sensory (OS) system to measure reflection of the infrared radiation between the rotor and stator. The method is validated by using a parallel measuring system employing an analog Hall sensory (AHS) system that measures the change in the magnetic flux density. The two measuring systems are calibrated by comparing them to a reference laboratory measuring system. Using different calibration techniques, the optimal system accuracy with a maximum permissible error of 0,15 mm is obtained, in the measuring range between 0 and 2 mm. It covers most applications of the direct-drive electrical motors. The new measuring method is validated by using an experimental setup consisting of the presented OS system and validation system consisting of an AHS system integrated on a testing platform (Smart Fortwo car).
COBISS.SI-ID: 12414548
The purpose of this study was to evaluate whether combined preoperative non-invasive determination of cardiac autonomic regulation and PR interval allows for the identification of patients at risk of new-onset atrial fibrillation after cardiac surgery. RR, PR and QT intervals, and linear and non-linear heart rate variability parameters from 20?min high-resolution electrocardiographic recordings were determined one day before surgery in 150 patients on chronic beta blockers undergoing elective coronary artery bypass grafting, aortic valve replacement, or both, electively. Thirty-one patients (21%) developed postoperative atrial fibrillation. In the atrial fibrillation group, more arterial hypertension, a greater age, a higher EuroSCORE II, a higher heart rate variability index (pNN50: 9?±?20 vs. 4?±?10, p?=?0.050), a short PR interval (156?±?23 vs. 173?±?31?ms; p?=?0.011), and a reduced short-term scaling exponent of the detrended fluctuation analysis (DFA1, 0.96?±?0.36 vs. 1.11?±?0.30?ms; p?=?0.032) were found compared to the sinus rhythm group. Logistic regression modeling confirmed PR interval, DFA1 and age as the strongest preoperative predictors of postoperative atrial fibrillation (area under the receiver operating characteristic curve?=?0.804). Patients developing atrial fibrillation after cardiac surgery presented with severe cardiac autonomic derangement and a short PR interval preoperatively. The observed state characterizes both altered heart rate regulation and arrhythmic substrate and is strongly related to an increased risk of postoperative atrial fibrillation.
COBISS.SI-ID: 5627243