• A deep learning framework to reduce video compression effects on rPPG signals. • Selectively enhances rPPG-relevant info, unlike standard video quality methods. • Combines an rPPG Estimator and a magnifier for pulse signals in compressed video. • Validated on four public datasets, showing strong robustness and generalization. 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 investigate the impact of video compression on the recovery of physiological signals from camera-based recordings. To mitigate the negative effects of compression on rPPG estimation, we propose a novel two-stage training strategy. This approach incorporates a pulse-signal magnification transformation, which adapts compressed video data into an uncompressed domain, where the rPPG signal is amplified. We validate the effectiveness of our model through comprehensive evaluations on two publicly available datasets, UCLA-rPPG and UBFC-rPPG, assessing both intra- and cross-database performance across various 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.

Recent advances in remote heart rate measure-ment, 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 inher-ently 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.

Current methods based on Neural Radiance Fields (NeRF) significantly lack the capacity to quantify uncertainty in their predictions, particularly on the unseen space including the occluded and outside scene content. This limitation hinders their extensive applications in robotics, where the reliability of model predictions has to be considered for tasks such as robotic exploration and planning in unknown environments. To address this, we propose a novel approach to estimate a 3D Uncertainty Field based on the learned incomplete scene geometry, which explicitly identifies these unseen regions. By considering the accumulated transmittance along each camera ray, our Uncertainty Field infers 2D pixel-wise uncertainty, exhibiting high values for rays directly casting towards occluded or outside the scene content. To quantify the uncertainty on the learned surface, we model a stochastic radiance field. Our experiments demonstrate that our approach is the only one that can explicitly reason about high uncertainty both on 3D unseen regions and its involved 2D rendered pixels, compared with recent methods. Furthermore, we illustrate that our designed uncertainty field is ideally suited for real-world robotics tasks, such as next-best-view selection.

Neural Radiance Fields (NeRF) has become a popular framework for learning implicit 3D representations and addressing different tasks such as novel-view synthesis or depth-map estimation. However, in downstream applications where decisions need to be made based on automatic predictions, it is critical to leverage the confidence associated with the model estimations. Whereas uncertainty quantification is a long-standing problem in Machine Learning, it has been largely overlooked in the recent NeRF literature. In this context, we propose Stochastic Neural Radiance Fields (S-NeRF), a generalization of standard NeRF that learns a probability distribution over all the possible radiance fields modeling the scene. This distribution allows to quantify the uncertainty associated with the scene information provided by the model. S-NeRF optimization is posed as a Bayesian learning problem that is efficiently addressed using the Variational Inference framework. Exhaustive experiments over benchmark datasets demonstrate that S-NeRF is able to provide more reliable predictions and confidence values than generic approaches previously proposed for uncertainty estimation in other domains.

We address the problem of multi-person 3D body pose and shape estimation from a single image. While this problem can be addressed by applying single-person approaches multiple times for the same scene, recent works have shown the advantages of building upon deep architectures that simultaneously reason about all people in the scene in a holistic manner by enforcing, e.g., depth order constraints or minimizing interpenetration among reconstructed bodies. However, existing approaches are still unable to capture the size variability of people caused by the inherent body scale and depth ambiguity. In this work, we tackle this challenge by devising a novel optimization scheme that learns the appropriate body scale and relative camera pose, by enforcing the feet of all people to remain on the ground floor. A thorough evaluation on MuPoTS- 3D and 3DPW datasets demonstrates that our approach is able to robustly estimate the body translation and shape of multiple people while retrieving their spatial arrangement, consistently improving current state-of-the-art, especially in scenes with people of very different heights. Code can be found at: https://github.com/nicolasugrinovic/size_depth_disambiguation

Inference in deep neural networks can be computationally expensive, and networks capable of anytime inference are important in scenarios where the amount of compute or input data varies over time. In such networks the inference process can interrupted to provide a result faster, or continued to obtain a more accurate result. We propose Hierarchical Neural Ensembles (HNE), a novel framework to embed an ensemble of multiple networks in a hierarchical tree structure, sharing intermediate layers. In HNE we control the complexity of inference on-the-fly by evaluating more or less models in the ensemble. Our second contribution is a novel hierarchical distillation method to boost the predictions of small ensembles. This approach leverages the nested structure of our ensembles, to optimally allocate accuracy and diversity across the individual models. Our experiments show that, compared to previous anytime inference models, HNE provides state-of-the-art accuracy-computation trade-offs on the CIFAR-10/100 and ImageNet datasets.

Attribution map visualization has arisen as one of the most effective techniques to understand the underlying inference process of Convolutional Neural Networks. In this task, the goal is to compute an score for each image pixel related to its contribution to the network output. In this paper, we introduce Disentangled Masked Backpropagation (DMBP), a novel gradient-based method that leverages on the piecewise linear nature of ReLU networks to decompose the model function into different linear mappings. This decomposition aims to disentangle the attribution maps into positive, negative and nuisance factors by learning a set of variables masking the contribution of each filter during back-propagation. A thorough evaluation over standard architectures (ResNet50 and VGG16) and benchmark datasets (PASCAL VOC and ImageNet) demonstrates that DMBP generates more visually interpretable attribution maps than previous approaches. Additionally, we quantitatively show that the maps produced by our method are more consistent with the true contribution of each pixel to the final network output.

Despite the outstanding performance of convolutional neural networks (CNNs) for many vision tasks, the required computational cost during inference is problematic when resources are limited. In this context, we propose Convolutional Neural Mixture Models (CNMMs), a probabilistic model embedding a large number of CNNs that can be jointly trained and evaluated in an efficient manner. Within the proposed framework, we present different mechanisms to prune subsets of CNNs from the mixture, allowing to easily adapt the computational cost required for inference. Image classification and semantic segmentation experiments show that our method achieve excellent accuracy-compute trade-offs. Moreover, unlike most of previous approaches, a single CNMM provides a large range of operating points along this trade-off, without any re-training.

Head pose estimation is a challenging computer vision problem with important applications in different scenarios such as human-computer interaction or face recognition. In this paper, we present an algorithm for 3D head pose estimation using only depth information from Kinect sensors. A key feature of the proposed approach is that it allows modeling the underlying 3D manifold that results from the combination of pitch, yaw and roll variations. To do so, we use tensor decomposition to generate separate subspaces for each variation factor and show that each of them has a clear structure that can be modeled with cosine functions from a unique shared parameter per angle. Such representation provides a deep understanding of data behavior and angle estimations can be performed by optimizing combination of these cosine functions. We evaluate our approach on two publicly available databases, and achieve top state-of-the-art performance.

We propose a Multi-Instance-Learning (MIL) approach for weakly-supervised learning problems, where a training set is formed by bags (sets of feature vectors or instances) and only labels at bag-level are provided. Specifically, we consider the Multi-Instance Dynamic-Ordinal-Regression (MI-DOR) setting, where the instance labels are naturally represented as ordinal variables and bags are structured as temporal sequences. To this end, we propose Multi-Instance Dynamic Ordinal Random Fields (MI-DORF). In this framework, we treat instance-labels as temporally-dependent latent variables in an Undirected Graphical Model. Different MIL assumptions are modelled via newly introduced high-order potentials relating bag and instance-labels within the energy function of the model. We also extend our framework to address the Partially-Observed MI-DOR problems, where a subset of instance labels are available during training.We show on the tasks of weakly-supervised facial behavior analysis, Facial Action Unit (DISFA dataset) and Pain (UNBC dataset) Intensity estimation, that the proposed framework outperforms alternative learning approaches. Furthermore, we show that MIDORF can be employed to reduce the data annotation efforts in this context by large-scale.

In this paper we present a system that is able to estimate head pose using only depth information from consumer RGB-D cameras such as Kinect 2. In contrast to most approaches addressing this problem, we do not rely on tracking and produce pose estimation in terms of pitch, yaw and roll angles using single depth frames as input. Our system combines three different methods for pose estimation: two of them are based on state-of-the-art landmark detection and the third one is a dictionarybased approach that is able to work in especially challenging scans where landmarks or mesh correspondences are too difficult to obtain. We evaluated our system on the SASE database, which consists of ~30K frames from 50 subjects. We obtained average pose estimation errors between 5 and 8 degrees per angle, achieving the best performance in the FG2017 Head Pose Estimation Challenge. Full code of the developed system is available on-line.

An essential issue when training and validating computer vision systems for affect analysis is how to obtain reliable ground-truth labels from a pool of subjective annotations. In this paper, we address this problem when labels are given in an ordinal scale and annotated items are structured as temporal sequences. This problem is of special importance in affective computing, where collected data is typically formed by videos of human interactions annotated according to the Valence and Arousal (V-A) dimensions. Moreover, recent works have shown that inter-observer agreement of V-A annotations can be considerably improved if these are given in a discrete ordinal scale. In this context, we propose a novel framework which explicitly introduces ordinal constraints to model the subjective perception of annotators. We also incorporate dynamic information to take into account temporal correlations between ground-truth labels. In our experiments over synthetic and real data with V-A annotations, we show that the proposed method outperforms alternative approaches which do not take into account either the ordinal structure of labels or their temporal correlation.

Limited annotated training data is a challenging problem in Action Unit recognition. In this paper, we investigate how the use of large databases labelled according to the 6 universal facial expressions can increase the generalization ability of Action Unit classifiers. For this purpose, we propose a novel learning framework: Hidden-Task Learning. HTL aims to learn a set of Hidden-Tasks (Action Units) for which samples are not available but, in contrast, training data is easier to obtain from a set of related Visible-Tasks (Facial Expressions). To that end, HTL is able to exploit prior knowledge about the relation between Hidden and Visible-Tasks. In our case, we base this prior knowledge on empirical psychological studies providing statistical correlations between Action Units and universal facial expressions. Additionally, we extend HTL to Semi-Hidden Task Learning (SHTL) assuming that Action Unit training samples are also provided. Performing exhaustive experiments over four different datasets, we show that HTL and SHTL improve the generalization ability of AU classifiers by training them with additional facial expression data. Additionally, we show that SHTL achieves competitive performance compared with state-of-the-art Transductive Learning approaches which face the problem of limited training data by using unlabelled test samples during training.