Abstract
This work presents a novel multimodal database comprising 3D face, 2D face, thermal face, visible iris, finger and hand veins, voice and anthropometrics. This dataset will constitute a valuable resource to the field with its number and variety of biometric traits. Acquired in the context of the EU PROTECT project, the dataset allows several combinations of biometric traits and envisages applications such as border control. Based upon the results of the unimodal data, a fusion scheme was applied to ascertain the recognition potential of combining these biometric traits in a multimodal approach. Due to the variability on the discriminative power of the traits, a leave the n-best out fusion technique was applied to obtain different recognition results. © 2018 Gesellschaft fuer Informatik.

Abstract

Abstract
Surgical training for minimal invasive kidney interventions (MIKI) has huge importance within the urology field. Within this topic, simulate MIKI in a patient-specific virtual environment can be used for pre-operative planning using the real patient's anatomy, possibly resulting in a reduction of intra-operative medical complications. However, the validated VR simulators perform the training in a group of standard models and do not allow patient-specific training. For a patient-specific training, the standard simulator would need to be adapted using personalized models, which can be extracted from pre-operative images using segmentation strategies. To date, several methods have already been proposed to accurately segment the kidney in computed tomography (CT) images. However, most of these works focused on kidney segmentation only, neglecting the extraction of its internal compartments. In this work, we propose to adapt a coupled formulation of the B-Spline Explicit Active Surfaces (BEAS) framework to simultaneously segment the kidney and the renal collecting system (CS) from CT images. Moreover, from the difference of both kidney and CS segmentations, one is able to extract the renal parenchyma also. The segmentation process is guided by a new energy functional that combines both gradient and region-based energies. The method was evaluated in 10 kidneys from 5 CT datasets, with different image properties. Overall, the results demonstrate the accuracy of the proposed strategy, with a Dice overlap of 92.5%, 86.9% and 63.5%, and a point-to-surface error around 1.6 mm, 1.9 mm and 4 mm for the kidney, renal parenchyma and CS, respectively. © 2018 IEEE.

Abstract
Data imbalance is characterized by a discrepancy in the number of examples per class of a dataset. This phenomenon is known to deteriorate the performance of classifiers, since they are less able to learn the characteristics of the less represented classes. For most imbalanced datasets, the application of sampling techniques improves the classifier's performance. For small datasets, oversampling has been shown to be the most appropriate strategy since it augments the original set of samples. Although several oversampling strategies have been proposed and tested over the years, the work has mostly focused on binary or multi-class tasks. Motivated by medical applications, where there is often an order associated with the classes (increasing likelihood of malignancy, for instance), the present work tests some existing oversampling techniques in ordinal contexts. Moreover, four new oversampling techniques are proposed. Experiments were made both on private and public datasets. Private datasets concern the assessment of response to treatment on oncologic diseases. The 15 public datasets were chosen since they are widely used in the literature. Results show that data balance techniques improve classification results on ordinal imbalanced datasets, even when these techniques are not specifically designed for ordinal problems. With our pipeline, better or equal to published results were obtained for 10 out of the 15 public datasets with improvements upon a decrease of 0.43 on MMAE.

Abstract

Abstract
While the popularity of virtual reality (VR) grows in a wide range of application contexts - e.g. entertainment, training, cultural heritage and medicine -, its economic impact is expected to reach around 15bn USD, by the year of 2020. Within VR field, 360video has been sparking the interest of development and research communities. However, editing tools supporting 360panoramas are usually expensive and/or demand programming skills and/or advanced user knowledge. Besides, application approaches to quickly and intuitively set up such 360video-based VR environments complemented with diverse types of parameterizable virtual assets and multimedia elements are still hard to find. Thereby, this paper aims to propose a system specification to simply and rapidly configure immersive VR environments composed of surrounding 360video spheres that can be complemented with parameterizable multimedia contents - namely 3D models, text and spatial sound -, whose behavior can be either time-range or user-interaction dependent. Moreover, a preliminary prototype that follows a substantial part of the previously mentioned specification and implements the enhancement of 360videos with time-range dependent virtual assets is presented. Preliminary tests evaluating usability and user satisfaction were also carried out with 30 participants, from which encouraging results were achieved. © 2018 The Authors. Published by Elsevier Ltd..