Open Access
Accurate pupil center detection in off-the-shelf eye tracking systems using convolutional neural networks
(MDPI, 2021) Larumbe Bergera, Andoni; Garde Lecumberri, Gonzalo; Porta Cuéllar, Sonia; Cabeza Laguna, Rafael; Villanueva Larre, Arantxa; Ingeniería Eléctrica, Electrónica y de Comunicación; Ingeniaritza Elektrikoa, Elektronikoaren eta Telekomunikazio Ingeniaritzaren; Universidad Pública de Navarra / Nafarroako Unibertsitate Publikoa
Remote eye tracking technology has suffered an increasing growth in recent years due to its applicability in many research areas. In this paper, a video-oculography method based on convolutional neural networks (CNNs) for pupil center detection over webcam images is proposed. As the first contribution of this work and in order to train the model, a pupil center manual labeling procedure of a facial landmark dataset has been performed. The model has been tested over both real and synthetic databases and outperforms state-of-the-art methods, achieving pupil center estimation errors below the size of a constricted pupil in more than 95% of the images, while reducing computing time by a 8 factor. Results show the importance of use high quality training data and well-known architectures to achieve an outstanding performance.
Open Access
Robust and accurate 2D-tracking-based 3D positioning method: application to head pose estimation
(Elsevier, 2019) Ariz Galilea, Mikel; Villanueva Larre, Arantxa; Cabeza Laguna, Rafael; Ingeniería Eléctrica, Electrónica y de Comunicación; Ingeniaritza Elektrikoa, Elektronikoaren eta Telekomunikazio Ingeniaritzaren
Head pose estimation (HPE) is currently a growing research field, mainly because of the proliferation of human–computer interfaces (HCI) in the last decade. It offers a wide variety of applications, including human behavior analysis, driver assistance systems or gaze estimation systems. This article aims to contribute to the development of robust and accurate HPE methods based on 2D tracking of the face, enhancing performance of both 2D point tracking and 3D pose estimation. We start with a baseline method for pose estimation based on POSIT algorithm. A novel weighted variant of POSIT is then proposed, together with a methodology to estimate weights for the 2D–3D point correspondences. Further, outlier detection and correction methods are also proposed in order to enhance both point tracking and pose estimation. With the aim of achieving a wider impact, the problem is addressed using a global approach: all the methods proposed are generalizable to any kind of object for which an approximate 3D model is available. These methods have been evaluated for the specific task of HPE using two different head pose video databases; a recently published one that reflects the expected performance of the system in current technological conditions, and an older one that allows an extensive comparison with state-of-the-art HPE methods. Results show that the proposed enhancements improve the accuracy of both 2D facial point tracking and 3D HPE, with respect to the implemented baseline method, by over 15% in normal tracking conditions and over 30% in noisy tracking conditions. Moreover, the proposed HPE system outperforms the state of the art on the two databases.
Open Access
Development of a prediction protocol for the screening of metabolic associated fatty liver disease in children with overweight or obesity
(Wiley, 2022) Osés Recalde, Maddi; Cadenas-Sánchez, Cristina; Medrano Echeverría, María; Galbete Jiménez, Arkaitz; Miranda Ferrúa, Emiliano; Ruiz, Jonatan R.; Sánchez-Valverde, Félix; Ortega, Francisco B.; Cabeza Laguna, Rafael; Villanueva Larre, Arantxa; Idoate, Fernando; Labayen Goñi, Idoia; Osasun Zientziak; Institute of Smart Cities - ISC; Institute on Innovation and Sustainable Development in Food Chain - ISFOOD; Ciencias de la Salud; Gobierno de Navarra / Nafarroako Gobernua
Background: the early detection and management of children with metabolic associ-ated fatty liver disease (MAFLD) is challenging. Objective: to develop a non-invasive and accurate prediction protocol for the identi-fication of MAFLD among children with overweight/obesity candidates to confirma-tory diagnosis. Methods: a total of 115 children aged 8–12 years with overweight/obesity, rec-ruited at a primary care, were enrolled in this cross-sectional study. The external vali-dation was performed using a cohort of children with overweight/obesity (N=46)aged 8.5–14.0 years. MAFLD (≥5.5% hepatic fat) was diagnosed by magnetic reso-nance imaging (MRI). Fasting blood biochemical parameters were measured, and25 candidates’ single nucleotide polymorphisms (SNPs) were determined. Variablespotentially associated with the presence of MAFLD were included in a multivariatelogistic regression. Results: children with MAFLD (36%) showed higher plasma triglycerides (TG),insulin, homeostasis model assessment ofinsulin resistance (HOMA-IR), alanineaminotransferase (ALT), aspartate transaminase (AST), glutamyl-transferase (GGT)and ferritin (p< 0.05). The distribution of the risk-alleles of PPARGrs13081389, PPARGrs1801282, HFErs1800562 and PNLPLA3rs4823173 was significantly different between children with and without MAFLD (p<0.05). Threebiochemical- and/or SNPs-based predictive models were developed, showingstrong discriminatory capacity (AUC-ROC: 0.708–0.888) but limited diagnosticperformance (sensitivity 67%–82% and specificity 63%–69%). A prediction proto-col with elevated sensitivity (72%) and specificity (84%) based on two consecutive steps was developed. The external validation showed similar results: sensitivity of 70% and specificity of 85%. Conclusions: the HEPAKID prediction protocol is an accurate, easy to implant, minimally invasive and low economic cost tool useful for the early identification and management of paediatric MAFLD in primary care.
Open Access
A deep learning image analysis method for renal perfusion estimation in pseudo-continuous arterial spin labelling MRI
(Elsevier, 2023) Oyarzun Domeño, Anne; Cia Alonso, Izaskun; Echeverría Chasco, Rebeca; Fernández Seara, María A.; Martín Moreno, Paloma L.; Bastarrika, Gorka; Navallas Irujo, Javier; Villanueva Larre, Arantxa; Ingeniería Eléctrica, Electrónica y de Comunicación; Institute of Smart Cities - ISC; Ingeniaritza Elektrikoa, Elektronikoaren eta Telekomunikazio Ingeniaritzaren; Universidad Pública de Navarra / Nafarroako Unibertsitate Publikoa
Accurate segmentation of renal tissues is an essential step for renal perfusion estimation and postoperative assessment of the allograft. Images are usually manually labeled, which is tedious and prone to human error. We present an image analysis method for the automatic estimation of renal perfusion based on perfusion magnetic resonance imaging. Specifically, non-contrasted pseudo-continuous arterial spin labeling (PCASL) images are used for kidney transplant evaluation and perfusion estimation, as a biomarker of the status of the allograft. The proposed method uses machine/deep learning tools for the segmentation and classification of renal cortical and medullary tissues and automates the estimation of perfusion values. Data from 16 transplant patients has been used for the experiments. The automatic analysis of differentiated tissues within the kidney, such as cortex and medulla, is performed by employing the time-intensity-curves of non-contrasted T1-weighted MRI series. Specifically, using the Dice similarity coefficient as a figure of merit, results above 93%, 92% and 82% are obtained for whole kidney, cortex, and medulla, respectively. Besides, estimated cortical and medullary perfusion values are considered to be within the acceptable ranges within clinical practice.
Open Access
Attention to online channels across the path to purchase: an eye-tracking study
(Elsevier, 2019) Cortiñas Ugalde, Mónica; Cabeza Laguna, Rafael; Chocarro Eguaras, Raquel; Villanueva Larre, Arantxa; Ingeniaritza Elektrikoa, Elektronikoaren eta Telekomunikazio Ingeniaritzaren; Enpresen Kudeaketa; Institute for Advanced Research in Business and Economics - INARBE; Ingeniería Eléctrica, Electrónica y de Comunicación; Gestión de Empresas
Currently, consumers display what is known as omnichannel behavior: the combined use of digital and physical channels providing them with multiple points of contact with firms. We combine the stimulus-organism-response (S-O-R) model and visual attention theory to study how customers’ attention to digital channels varies across different purchasing tasks. We use eye-tracking techniques to observe attention in an experimental setting. The experimental design is composed of four purchasing tasks in four different product categories and measures the attention to the website and time spent on each task in addition to several control variables. The results show that shoppers attend to more areas of the website for purposes of website exploration than for performing purchase tasks. The most complex and time-consuming task for shoppers is the assessment of purchase options. The actual purchase and post-purchase tasks require less time and the inspection of fewer areas of interest. Personal involvement also plays a role in determining these patterns by increasing attention to the product area.
Open Access
Evaluation of accurate eye corner detection methods for gaze estimation
(Bern Open Publishing, 2014) Bengoechea Irañeta, José Javier; Cerrolaza Martínez, Juan José; Villanueva Larre, Arantxa; Cabeza Laguna, Rafael; Ingeniería Eléctrica, Electrónica y de Comunicación; Ingeniaritza Elektrikoa, Elektronikoaren eta Telekomunikazio Ingeniaritzaren
Accurate detection of iris center and eye corners appears to be a promising approach for low cost gaze estimation. In this paper we propose novel eye inner corner detection methods. Appearance and feature based segmentation approaches are suggested. All these methods are exhaustively tested on a realistic dataset containing images of subjects gazing at different points on a screen. We have demonstrated that a method based on a neural network presents the best performance even in light changing scenarios. In addition to this method, algorithms based on AAM and Harris corner detector present better accuracies than recent high performance face points tracking methods such as Intraface.
Open Access
Effects of time-restricted eating and resistance training on skeletal muscle tissue quantity, quality and function in postmenopausal women with overweight or obesity: a study protocol
(Elsevier, 2024-12-30) Alfaro-Magallanes, Víctor Manuel; Medrano Echeverría, María; Echarte Medina, Jon; Osés Recalde, Maddi; Izquierdo Rodríguez, Claudia; Concepción Álvarez, Mara de la Caridad; Galbete Jiménez, Arkaitz; Idoate, Fernando; Zugasti Murillo, Ana; Petrina Jáuregui, María Estrella; Goñi Gironés, María Elena; Ribelles, María Jesús; Amasene, María; Arenaza Etxeberría, Lide; Tejada Garrido, Clara Isabel; Elejalde, E.; Azcárate Jiménez, Unai Xabier; Ruiz Sarrias, Oskitz; Sayar-Beristain, Onintza; García-Ramos, Amador; Martínez Labari, Cristina; Armendáriz Brugos, Cristina; Villanueva Larre, Arantxa; Ruiz, Jonatan R.; Cabeza Laguna, Rafael; Labayen Goñi, Idoia; Ciencias de la Salud; Osasun Zientziak; Institute of Smart Cities - ISC; Ingeniería Eléctrica, Electrónica y de Comunicación; Ingeniaritza Elektrikoa, Elektronikoa eta Telekomunikazio Ingeniaritza; Estadística, Informática y Matemáticas; Estatistika, Informatika eta Matematika; Institute on Innovation and Sustainable Development in Food Chain - ISFOOD
Background & aims: time-restricted eating (TRE) shows promise for weight loss and improving menopauserelated body composition and cardiometabolic health, but its effects on skeletal muscle tissue (SMT) in postmenopausal women are unknown. This study investigates the effects of three weight loss interventions over 12 weeks on SMT quantity, quality, function, and cardiometabolic health in postmenopausal women with overweight/obesity, with effects persistence evaluated at a 12-month follow-up. Methods and results: in this randomized controlled trial, 78 postmenopausal women (50–65 years; BMI 25–40 kg/m2; sedentary lifestyle; eating window ≥12 h/day; no severe metabolic impairments) will be recruited. Participants will be randomly assigned to one of three groups for 12 weeks: TRE, TRE + resistance training, or CR + resistance training. The TRE groups will reduce their eating window to 8 h and receive nutritional advice to adhere to a Mediterranean diet. The CR group will follow a personalized hypocaloric diet (− 500 kcal/day). Resistance training groups will perform supervised resistance training 3 times/week. Primary Outcome: Change in SMT quantity measured by MRI at baseline and after 12 weeks. Secondary Outcomes: intermuscular adipose tissue (IMAT), strength, power, body weight and composition, and cardiometabolic risk factors. Conclusion: this study will illustrate the effects of TRE and TRE combined with resistance exercise compared with the currently recommended obesity-lifestyle treatment on SMT quantity, quality, function, and cardiometabolic markers. The results will offer insights into dietary strategies to combat obesity and metabolic diseases without increasing sarcopenia risk in postmenopausal women, a sparsely studied and particularly affected population.
Open Access
Advancing ASL kidney image registration: a tailored pipeline with VoxelMorph
(Springer, 2025-01-31) Oyarzun Domeño, Anne; Cia Alonso, Izaskun; Echeverría Chasco, Rebeca; Fernández Seara, María A.; Martín Moreno, Paloma L.; García Fernández, Nuria; Bastarrika, Gorka; Navallas Irujo, Javier; Villanueva Larre, Arantxa; Ingeniería Eléctrica, Electrónica y de Comunicación; Ingeniaritza Elektrikoa, Elektronikoa eta Telekomunikazio Ingeniaritza; Institute of Smart Cities - ISC; Universidad Pública de Navarra / Nafarroako Unibertsitate Publikoa; Gobierno de Navarra / Nafarroako Gobernua
In clinical renal assessment, image registration plays a pivotal role, as patient movement during data acquisition can significantly impede image post-processing and the accurate estimation of hemodynamic parameters. This study introduces a deep learning-based image registration framework specifically for arterial spin labeling (ASL) imaging. ASL is a magnetic resonance imaging technique that modifies the longitudinal magnetization of blood perfusing the kidney using a series of radiofrequency pulses combined with slice-selective gradients. After tagging the arterial blood, label images are captured following a delay, allowing the tagged blood bolus to enter the renal tissue, while control images are acquired without tagging the arterial spins. Given that perfusion maps are generated at the pixel level by subtracting control images from label images and considering the relatively small signal intensity difference, precise alignment of these images is crucial to minimize motion artefacts and prevent significant errors in perfusion calculations. Moreover, due to the extended ASL acquisition times and the anatomical location of the kidneys, renal images are often susceptible to pulsation, peristalsis, and breathing motion. These motion-induced noises and other instabilities can adversely affect ASL imaging outcomes, making image registration essential. However, research on renal MRI registration, particularly with respect to learning-based techniques, remains limited, with even less focus on renal ASL. Our study proposes a learning-based image registration approach that builds upon VoxelMorph and introduces groupwise inference as a key enhancement. The dataset includes 2448 images of transplanted kidneys (TK) and 2456 images of healthy kidneys (HK). We compared the automatic image registration results with the widely recognized optimization method Elastix. The model’s performance was evaluated using the mean structural similarity index (MSSIM), normalized correlation coefficient (NCC), temporal signal-tonoise ratio (TSNR) of the samples, and the mean cortical signal (CSIM) in perfusion-weighted images, thereby extending the evaluation beyond traditional similarity-based metrics. Our method achieved superior image registration performance, with peak NCC (0.987 ± 0.006) and MSSIM (0.869 ± 0.048) values in the kidney region, significantly surpassing Elastix and the unregistered series (p\ 0.05) on TK and HK datasets. Regularization analysis showed that higher k values (1, 2) produced smoother deformation fields, while moderate k values (0.5, 0.9) balanced smoothness and detail, maintaining low non-positive Jacobian percentages (\1%) comparable to Elastix. Additionally, our method improved CSIM by 14.3% (2.304 ± 1.167) and TSNR by 13.1% (3.888 ± 2.170) in TK, and achieved up to 13.2% (CSIM) and 29.8% (TSNR) enhancements in HK, demonstrating robustness and improved signal quality across datasets and acquisition techniques.
Open Access
Improved strategies for HPE employing learning-by-synthesis approaches
(IEEE, 2018) Larumbe Bergera, Andoni; Ariz Galilea, Mikel; Bengoechea Irañeta, José Javier; Segura, Rubén; Cabeza Laguna, Rafael; Villanueva Larre, Arantxa; Ingeniería Eléctrica, Electrónica y de Comunicación; Ingeniaritza Elektrikoa, Elektronikoaren eta Telekomunikazio Ingeniaritzaren
The first contribution of this paper is the presentation of a synthetic video database where the groundtruth of 2D facial landmarks and 3D head poses is available to be used for training and evaluating Head Pose Estimation (HPE) methods. The database is publicly available and contains videos of users performing guided and natural movements. The second and main contribution is the submission of a hybrid method for HPE based on Pose from Ortography and Scaling by Iterations (POSIT). The 2D landmark detection is performed using Random Cascaded-Regression Copse (R-CR-C). For the training stage we use, state of the art labeled databases. Learning-by-synthesis approach has been also used to augment the size of the database employing the synthetic database. HPE accuracy is tested by using two literature 3D head models. The tracking method proposed has been compared with state of the art methods using Supervised Descent Regressors (SDR) in terms of accuracy, achieving an improvement of 60%.
Open Access
SeTA: semiautomatic tool for annotation of eye tracking images
(ACM, 2019) Larumbe Bergera, Andoni; Porta Cuéllar, Sonia; Cabeza Laguna, Rafael; Villanueva Larre, Arantxa; Ingeniería Eléctrica, Electrónica y de Comunicación; Ingeniaritza Elektrikoa, Elektronikoaren eta Telekomunikazio Ingeniaritzaren
Availability of large scale tagged datasets is a must in the field of deep learning applied to the eye tracking challenge. In this paper, the potential of Supervised-Descent-Method (SDM) as a semiautomatic labelling tool for eye tracking images is shown. The objective of the paper is to evidence how the human effort needed for manually labelling large eye tracking datasets can be radically reduced by the use of cascaded regressors. Different applications are provided in the fields of high and low resolution systems. An iris/pupil center labelling is shown as example for low resolution images while a pupil contour points detection is demonstrated in high resolution. In both cases manual annotation requirements are drastically reduced.