Federico M. Sukno

Joaquim Comas, Adria Ruiz, Federico Sukno

We present a lightweight neural model for remote heart rate estimation focused on the efficient spatio-temporal learning of facial photoplethysmography (PPG) based on i) modelling of PPG dynamics by combinations of multiple convolutional derivatives, and ii) increased flexibility of the model to learn possible offsets between the facial video PPG and the ground truth. PPG dynamics are modelled by a Temporal Derivative Module (TDM) constructed by the incremental aggregation of multiple convolutional derivatives, emulating a Taylor series expansion up to the desired order. Robustness to ground truth offsets is handled by the introduction of TALOS (Temporal Adaptive LOcation Shift), a new temporal loss to train learning-based models. We verify the effectiveness of our model by reporting accuracy and efficiency metrics on the public PURE and UBFC-rPPG datasets. Compared to existing models, our approach shows competitive heart rate estimation accuracy with a much lower number of parameters and lower computational cost.

Araceli Morales, Antonio R. Porras, Marius George Linguraru et al.

We present a 3D face reconstruction system that aims at recovering the 3D facial geometry of babies from uncalibrated photographs, BabyNet. Since the 3D facial geometry of babies differs substantially from that of adults, baby-specific facial reconstruction systems are needed. BabyNet consists of two stages: 1) a 3D graph convolutional autoencoder learns a latent space of the baby 3D facial shape; and 2) a 2D encoder that maps photographs to the 3D latent space based on representative features extracted using transfer learning. In this way, using the pre-trained 3D decoder, we can recover a 3D face from 2D images. We evaluate BabyNet and show that 1) methods based on adult datasets cannot model the 3D facial geometry of babies, which proves the need for a baby-specific method, and 2) BabyNet outperforms classical model-fitting methods even when a baby-specific 3D morphable model, such as BabyFM, is used.

Antonia Alomar, Mireia Masias, Marius George Linguraru et al.

Early detection of developmental disorders can be aided by analyzing infant craniofacial morphology, but modeling infant faces is challenging due to limited data and frequent spontaneous expressions. We introduce BabyFlow, a generative AI model that disentangles facial identity and expression, enabling independent control over both. Using normalizing flows, BabyFlow learns flexible, probabilistic representations that capture the complex, non-linear variability of expressive infant faces without restrictive linear assumptions. To address scarce and uncontrolled expressive data, we perform cross-age expression transfer, adapting expressions from adult 3D scans to enrich infant datasets with realistic and systematic expressive variants. As a result, BabyFlow improves 3D reconstruction accuracy, particularly in highly expressive regions such as the mouth, eyes, and nose, and supports synthesis and modification of infant expressions while preserving identity. Additionally, by integrating with diffusion models, BabyFlow generates high-fidelity 2D infant images with consistent 3D geometry, providing powerful tools for data augmentation and early facial analysis.

Joaquim Comas, Antonia Alomar, Adria Ruiz et al.

In recent years, deep learning methods have shown impressive results for camera-based remote physiological signal estimation, clearly surpassing traditional methods. However, the performance and generalization ability of Deep Neural Networks heavily depends on rich training data truly representing different factors of variation encountered in real applications. Unfortunately, many current remote photoplethysmography (rPPG) datasets lack diversity, particularly in darker skin tones, leading to biased performance of existing rPPG approaches. To mitigate this bias, we introduce PhysFlow, a novel method for augmenting skin diversity in remote heart rate estimation using conditional normalizing flows. PhysFlow adopts end-to-end training optimization, enabling simultaneous training of supervised rPPG approaches on both original and generated data. Additionally, we condition our model using CIELAB color space skin features directly extracted from the facial videos without the need for skin-tone labels. We validate PhysFlow on publicly available datasets, UCLA-rPPG and MMPD, demonstrating reduced heart rate error, particularly in dark skin tones. Furthermore, we demonstrate its versatility and adaptability across different data-driven rPPG methods.

Mahdi Ghafourian, Federico M. Sukno

Head pose estimation (HPE) plays a critical role in various computer vision applications such as human-computer interaction and facial recognition. In this paper, we propose a novel deep learning approach for head pose estimation with limited training data via non-linear manifold learning called NLML-HPE. This method is based on the combination of tensor decomposition (i.e., Tucker decomposition) and feed forward neural networks. Unlike traditional classification-based approaches, our method formulates head pose estimation as a regression problem, mapping input landmarks into a continuous representation of pose angles. To this end, our method uses tensor decomposition to split each Euler angle (yaw, pitch, roll) to separate subspaces and models each dimension of the underlying manifold as a cosine curve. We address two key challenges: 1. Almost all HPE datasets suffer from incorrect and inaccurate pose annotations. Hence, we generated a precise and consistent 2D head pose dataset for our training set by rotating 3D head models for a fixed set of poses and rendering the corresponding 2D images. 2. We achieved real-time performance with limited training data as our method accurately captures the nature of rotation of an object from facial landmarks. Once the underlying manifold for rotation around each axis is learned, the model is very fast in predicting unseen data. Our training and testing code is available online along with our trained models: https: //github.com/MahdiGhafoorian/NLML_HPE.

Antonia Alomar, Ricardo Rubio, Laura Salort et al.

Craniofacial anomalies indicate early developmental disturbances and are usually linked to many genetic syndromes. Early diagnosis is critical, yet ultrasound (US) examinations often fail to identify these features. This study presents an AI-driven tool to assist clinicians in standardizing fetal facial axes/planes in 3D US, reducing sonographer workload and facilitating the facial evaluation. Our network, structured into three blocks-feature extractor, rotation and translation regression, and spatial transformer-processes three orthogonal 2D slices to estimate the necessary transformations for standardizing the facial planes in the 3D US. These transformations are applied to the original 3D US using a differentiable module (the spatial transformer block), yielding a standardized 3D US and the corresponding 2D facial standard planes. The dataset used consists of 1180 fetal facial 3D US images acquired between weeks 20 and 35 of gestation. Results show that our network considerably reduces inter-observer rotation variability in the test set, with a mean geodesic angle difference of 14.12$^{\circ}$ $\pm$ 18.27$^{\circ}$ and an Euclidean angle error of 7.45$^{\circ}$ $\pm$ 14.88$^{\circ}$. These findings demonstrate the network's ability to effectively standardize facial axes, crucial for consistent fetal facial assessments. In conclusion, the proposed network demonstrates potential for improving the consistency and accuracy of fetal facial assessments in clinical settings, facilitating early evaluation of craniofacial anomalies.

Joaquim Comas, Adria Ruiz, Federico Sukno

Recent advancements in data-driven approaches for remote photoplethysmography (rPPG) have significantly improved the accuracy of remote heart rate estimation. However, the performance of such approaches worsens considerably under video compression, which is nevertheless necessary to store and transmit video data efficiently. In this paper, we present a novel approach to address the impact of video compression on rPPG estimation, which leverages a pulse-signal magnification transformation to adapt compressed videos to an uncompressed data domain in which the rPPG signal is magnified. We validate the effectiveness of our model by exhaustive evaluations on two publicly available datasets, UCLA-rPPG and UBFC-rPPG, employing both intra- and cross-database performance at several compression rates. Additionally, we assess the robustness of our approach on two additional highly compressed and widely-used datasets, MAHNOB-HCI and COHFACE, which reveal outstanding heart rate estimation results.

Joaquim Comas, Adria Ruiz, Federico Sukno

Recent advances in remote heart rate measurement, motivated by data-driven approaches, have notably enhanced accuracy. However, these improvements primarily focus on recovering the rPPG signal, overlooking the implicit challenges of estimating the heart rate (HR) from the derived signal. While many methods employ the Fast Fourier Transform (FFT) for HR estimation, the performance of the FFT is inherently affected by a limited frequency resolution. In contrast, the Chirp-Z Transform (CZT), a generalization form of FFT, can refine the spectrum to the narrow-band range of interest for heart rate, providing improved frequential resolution and, consequently, more accurate estimation. This paper presents the advantages of employing the CZT for remote HR estimation and introduces a novel data-driven adaptive CZT estimator. The objective of our proposed model is to tailor the CZT to match the characteristics of each specific dataset sensor, facilitating a more optimal and accurate estimation of HR from the rPPG signal without compromising generalization across diverse datasets. This is achieved through a Sparse Matrix Optimization (SMO). We validate the effectiveness of our model through exhaustive evaluations on three publicly available datasets UCLA-rPPG, PURE, and UBFC-rPPG employing both intra- and cross-database performance metrics. The results reveal outstanding heart rate estimation capabilities, establishing the proposed approach as a robust and versatile estimator for any rPPG method.

Joaquim Comas, Federico Sukno

Remote photoplethysmography (rPPG) captures cardiac signals from facial videos and is gaining attention for its diverse applications. While deep learning has advanced rPPG estimation, it relies on large, diverse datasets for effective generalization. In contrast, handcrafted methods utilize physiological priors for better generalization in unseen scenarios like motion while maintaining computational efficiency. However, their linear assumptions limit performance in complex conditions, where deep learning provides superior pulsatile information extraction. This highlights the need for hybrid approaches that combine the strengths of both methods. To address this, we present BeatFormer, a lightweight spectral attention model for rPPG estimation, which integrates zoomed orthonormal complex attention and frequency-domain energy measurement, enabling a highly efficient model. Additionally, we introduce Spectral Contrastive Learning (SCL), which allows BeatFormer to be trained without any PPG or HR labels. We validate BeatFormer on the PURE, UBFC-rPPG, and MMPD datasets, demonstrating its robustness and performance, particularly in cross-dataset evaluations under motion scenarios.

Joaquim Comas, Alexander Joel Vera, Xavier Vives et al.

In recent years, affective computing and its applications have become a fast-growing research topic. Despite significant advancements, the lack of affective multi-modal datasets remains a major bottleneck in developing accurate emotion recognition systems. Furthermore, the use of contact-based devices during emotion elicitation often unintentionally influences the emotional experience, reducing or altering the genuine spontaneous emotional response. This limitation highlights the need for methods capable of extracting affective cues from multiple modalities without physical contact, such as remote physiological emotion recognition. To address this, we present the Contactless Affective States Through Physiological Signals Database (CAST-Phys), a novel high-quality dataset explicitly designed for multi-modal remote physiological emotion recognition using facial and physiological cues. The dataset includes diverse physiological signals, such as photoplethysmography (PPG), electrodermal activity (EDA), and respiration rate (RR), alongside high-resolution uncompressed facial video recordings, enabling the potential for remote signal recovery. Our analysis highlights the crucial role of physiological signals in realistic scenarios where facial expressions alone may not provide sufficient emotional information. Furthermore, we demonstrate the potential of remote multi-modal emotion recognition by evaluating the impact of individual and fused modalities, showcasing its effectiveness in advancing contactless emotion recognition technologies.

Javier Ortega-Garcia, Julian Fierrez, Fernando Alonso-Fernandez et al.

A new multimodal biometric database designed and acquired within the framework of the European BioSecure Network of Excellence is presented. It is comprised of more than 600 individuals acquired simultaneously in three scenarios: 1) over the Internet, 2) in an office environment with desktop PC, and 3) in indoor/outdoor environments with mobile portable hardware. The three scenarios include a common part of audio/video data. Also, signature and fingerprint data have been acquired both with desktop PC and mobile portable hardware. Additionally, hand and iris data were acquired in the second scenario using desktop PC. Acquisition has been conducted by 11 European institutions. Additional features of the BioSecure Multimodal Database (BMDB) are: two acquisition sessions, several sensors in certain modalities, balanced gender and age distributions, multimodal realistic scenarios with simple and quick tasks per modality, cross-European diversity, availability of demographic data, and compatibility with other multimodal databases. The novel acquisition conditions of the BMDB allow us to perform new challenging research and evaluation of either monomodal or multimodal biometric systems, as in the recent BioSecure Multimodal Evaluation campaign. A description of this campaign including baseline results of individual modalities from the new database is also given. The database is expected to be available for research purposes through the BioSecure Association during 2008

Nuria Rodriguez-Diaz, Decky Aspandi, Federico Sukno et al.

Lie detection is considered a concern for everyone in their day to day life given its impact on human interactions. Thus, people normally pay attention to both what their interlocutors are saying and also to their visual appearances, including faces, to try to find any signs that indicate whether the person is telling the truth or not. While automatic lie detection may help us to understand this lying characteristics, current systems are still fairly limited, partly due to lack of adequate datasets to evaluate their performance in realistic scenarios. In this work, we have collected an annotated dataset of facial images, comprising both 2D and 3D information of several participants during a card game that encourages players to lie. Using our collected dataset, We evaluated several types of machine learning-based lie detectors in terms of their generalization, person-specific and cross-domain experiments. Our results show that models based on deep learning achieve the best accuracy, reaching up to 57\% for the generalization task and 63\% when dealing with a single participant. Finally, we also highlight the limitation of the deep learning based lie detector when dealing with cross-domain lie detection tasks.

Araceli Morales, Gemma Piella, Federico M. Sukno

Recently, a lot of attention has been focused on the incorporation of 3D data into face analysis and its applications. Despite providing a more accurate representation of the face, 3D facial images are more complex to acquire than 2D pictures. As a consequence, great effort has been invested in developing systems that reconstruct 3D faces from an uncalibrated 2D image. However, the 3D-from-2D face reconstruction problem is ill-posed, thus prior knowledge is needed to restrict the solutions space. In this work, we review 3D face reconstruction methods proposed in the last decade, focusing on those that only use 2D pictures captured under uncontrolled conditions. We present a classification of the proposed methods based on the technique used to add prior knowledge, considering three main strategies, namely, statistical model fitting, photometry, and deep learning, and reviewing each of them separately. In addition, given the relevance of statistical 3D facial models as prior knowledge, we explain the construction procedure and provide a list of the most popular publicly available 3D facial models. After the exhaustive study of 3D-from-2D face reconstruction approaches, we observe that the deep learning strategy is rapidly growing since the last few years, becoming the standard choice in replacement of the widespread statistical model fitting. Unlike the other two strategies, photometry-based methods have decreased in number due to the need for strong underlying assumptions that limit the quality of their reconstructions compared to statistical model fitting and deep learning methods. The review also identifies current challenges and suggests avenues for future research.

Dmytro Derkach, Federico M. Sukno

We investigate the problem of facial expression recognition using 3D data. Building from one of the most successful frameworks for facial analysis using exclusively 3D geometry, we extend the analysis from a curve-based representation into a spectral representation, which allows a complete description of the underlying surface that can be further tuned to the desired level of detail. Spectral representations are based on the decomposition of the geometry in its spatial frequency components, much like a Fourier transform, which are related to intrinsic characteristics of the surface. In this work, we propose the use of Graph Laplacian Features (GLF), which results from the projection of local surface patches into a common basis obtained from the Graph Laplacian eigenspace. We test the proposed approach in the BU-3DFE database in terms of expressions and Action Units recognition. Our results confirm that the proposed GLF produces consistently higher recognition rates than the curves-based approach, thanks to a more complete description of the surface, while requiring a lower computational complexity. We also show that the GLF outperform the most popular alternative approach for spectral representation, Shape- DNA, which is based on the Laplace Beltrami Operator and cannot provide a stable basis that guarantee that the extracted signatures for the different patches are directly comparable.

Adriana Fernandez-Lopez, Federico M. Sukno

Speech is the most common communication method between humans and involves the perception of both auditory and visual channels. Automatic speech recognition focuses on interpreting the audio signals, but it has been demonstrated that video can provide information that is complementary to the audio. Thus, the study of automatic lip-reading is important and is still an open problem. One of the key challenges is the definition of the visual elementary units (the visemes) and their vocabulary. Many researchers have analyzed the importance of the phoneme to viseme mapping and have proposed viseme vocabularies with lengths between 11 and 15 visemes. These viseme vocabularies have usually been manually defined by their linguistic properties and in some cases using decision trees or clustering techniques. In this work, we focus on the automatic construction of an optimal viseme vocabulary based on the association of phonemes with similar appearance. To this end, we construct an automatic system that uses local appearance descriptors to extract the main characteristics of the mouth region and HMMs to model the statistic relations of both viseme and phoneme sequences. To compare the performance of the system different descriptors (PCA, DCT and SIFT) are analyzed. We test our system in a Spanish corpus of continuous speech. Our results indicate that we are able to recognize approximately 58% of the visemes, 47% of the phonemes and 23% of the words in a continuous speech scenario and that the optimal viseme vocabulary for Spanish is composed by 20 visemes.

Adriana Fernandez-Lopez, Oriol Martinez, Federico M. Sukno

Speech is the most used communication method between humans and it involves the perception of auditory and visual channels. Automatic speech recognition focuses on interpreting the audio signals, although the video can provide information that is complementary to the audio. Exploiting the visual information, however, has proven challenging. On one hand, researchers have reported that the mapping between phonemes and visemes (visual units) is one-to-many because there are phonemes which are visually similar and indistinguishable between them. On the other hand, it is known that some people are very good lip-readers (e.g: deaf people). We study the limit of visual only speech recognition in controlled conditions. With this goal, we designed a new database in which the speakers are aware of being read and aim to facilitate lip-reading. In the literature, there are discrepancies on whether hearing-impaired people are better lip-readers than normal-hearing people. Then, we analyze if there are differences between the lip-reading abilities of 9 hearing-impaired and 15 normal-hearing people. Finally, human abilities are compared with the performance of a visual automatic speech recognition system. In our tests, hearing-impaired participants outperformed the normal-hearing participants but without reaching statistical significance. Human observers were able to decode 44% of the spoken message. In contrast, the visual only automatic system achieved 20% of word recognition rate. However, if we repeat the comparison in terms of phonemes both obtained very similar recognition rates, just above 50%. This suggests that the gap between human lip-reading and automatic speech-reading might be more related to the use of context than to the ability to interpret mouth appearance.