Search

Understanding and predicting pedestrian crossing behavioral intention is crucial for autonomous vehicles driving safety. Nonetheless, challenges emerge when using promising images or environmental context masks to extract various factors for time-series network modeling, causing pre-processing errors or a loss in efficiency. Typically, pedestrian positions captured by onboard cameras are often distorted and do not accurately reflect their actual movements. To address these issues, GTransPDM -- a Graph-embedded Transformer with a Position Decoupling Module -- was developed for pedestrian crossing intention prediction by leveraging multi-modal features. First, a positional decoupling module was proposed to decompose the pedestrian lateral movement and simulate depth variations in the image view. Then, a graph-embedded Transformer was designed to capture the spatial-temporal dynamics of human pose skeletons, integrating essential factors such as position, skeleton, and ego-vehicle motion. Experimental results indicate that the proposed method achieves 92% accuracy on the PIE dataset and 87% accuracy on the JAAD dataset, with a processing speed of 0.05ms. It outperforms the state-of-the-art in comparison.

The last mile of unsupervised domain adaptation (UDA) for semantic segmentation is the challenge of solving the syn-to-real domain gap. Recent UDA methods have progressed significantly, yet they often rely on strategies customized for synthetic single-source datasets (e.g., GTA5), which limits their generalisation to multi-source datasets. Conversely, synthetic multi-source datasets hold promise for advancing the last mile of UDA but remain underutilized in current research. Thus, we propose DEC, a flexible UDA framework for multi-source datasets. Following a divide-and-conquer strategy, DEC simplifies the task by categorizing semantic classes, training models for each category, and fusing their outputs by an ensemble model trained exclusively on synthetic datasets to obtain the final segmentation mask. DEC can integrate with existing UDA methods, achieving state-of-the-art performance on Cityscapes, BDD100K, and Mapillary Vistas, significantly narrowing the syn-to-real domain gap.

The development of Autonomous Driving (AD) systems in simulated environments like CARLA is crucial for advancing real-world automotive technologies. To drive innovation, CARLA introduced Leaderboard 2.0, significantly more challenging than its predecessor. However, current AD methods have struggled to achieve satisfactory outcomes due to a lack of sufficient ground truth data. Human driving logs provided by CARLA are insufficient, and previously successful expert agents like Autopilot and Roach, used for collecting datasets, have seen reduced effectiveness under these more demanding conditions. To overcome these data limitations, we introduce PRIBOOT, an expert agent that leverages limited human logs with privileged information. We have developed a novel BEV representation specifically tailored to meet the demands of this new benchmark and processed it as an RGB image to facilitate the application of transfer learning techniques, instead of using a set of masks. Additionally, we propose the Infraction Rate Score (IRS), a new evaluation metric designed to provide a more balanced assessment of driving performance over extended routes. PRIBOOT is the first model to achieve a Route Completion (RC) of 75% in Leaderboard 2.0, along with a Driving Score (DS) and IRS of 20% and 45%, respectively. With PRIBOOT, researchers can now generate extensive datasets, potentially solving the data availability issues that have hindered progress in this benchmark.

Virtual engines can generate dense depth maps for various synthetic scenes, making them invaluable for training depth estimation models. However, discrepancies between synthetic and real-world colors pose significant challenges for depth estimation in real-world scenes, especially in complex and uncertain environments encountered in unsupervised monocular depth estimation tasks. To address this issue, we propose Back2Color, a framework that predicts realistic colors from depth using a model trained on real-world data, thus transforming synthetic colors into their real-world counterparts. Additionally, we introduce the Syn-Real CutMix method for joint training with both real-world unsupervised and synthetic supervised depth samples, enhancing monocular depth estimation performance in real-world scenes. Furthermore, to mitigate the impact of non-rigid motions on depth estimation, we present an auto-learning uncertainty temporal-spatial fusion method (Auto-UTSF), which leverages the strengths of unsupervised learning in both temporal and spatial dimensions. We also designed VADepth, based on the Vision Attention Network, which offers lower computational complexity and higher accuracy than transformers. Our Back2Color framework achieves state-of-the-art performance on the Kitti dataset, as evidenced by improvements in performance metrics and the production of fine-grained details. This is particularly evident on more challenging datasets such as Cityscapes for unsupervised depth estimation.

In this article, we propose the detection of crowd anomalies through the extraction of information in the form of time series from video format using a multimodal approach. Through pattern recognition algorithms and segmentation, informative measures of the number of people and image occupancy are extracted at regular intervals, which are then analyzed to obtain trends and anomalous behaviors. Specifically, through temporal decomposition and residual analysis, intervals or specific situations of unusual behaviors are identified, which can be used in decision-making and improvement of actions in sectors related to human movement such as tourism or security. The application of this methodology on the webcam of Turisme Comunitat Valenciana in the town of Morella (Comunitat Valenciana, Spain) has provided excellent results. It is shown to correctly detect specific anomalous situations and unusual overall increases during the previous weekend and during the festivities in October 2023. These results have been obtained while preserving the confidentiality of individuals at all times by using measures that maximize anonymity, without trajectory recording or person recognition.

Predictive maintenance in manufacturing industry applications is a challenging research field. Packaging machines are widely used in a large number of logistic companies' warehouses and must be working uninterruptedly. Traditionally, preventive maintenance strategies have been carried out to improve the performance of these machines. However, this kind of policies does not take into account the information provided by the sensors implemented in the machines. This paper presents an expert system for the automatic estimation of work orders to implement predictive maintenance policies for packaging machines. The key idea is that, from a set of alarms related to sensors implemented in the machine, the expert system should take a maintenance action while optimizing the response time. The work order estimator will act as a classifier, yielding a binary decision of whether a machine must undergo a maintenance action by a technician or not, followed by an unsupervised anomaly detection-based filtering stage to audit the classifier's output. The methods used for anomaly detection were: One-Class Support Vector Machine (OCSVM), Minimum Covariance Determinant (MCD) and a majority (hard) voting ensemble of them. All anomaly detection methods improve the performance of the baseline classifer but the best performance in terms of F1 score was obtained by the majority voting ensemble.

Background: Calculating the characteristics of radiation emerging from radiation shielding to determine the level of safety in protected areas has been a widely studied topic in Radiation Protection due to its importance for the protection of the public and professionals. Methods: This work provides spatial distributions of energy spectra of emergent photons after shielding barriers. These distributions are obtained through Monte Carlo simulations using monoenergetic pencil photon beams incident on shielding barriers of different materials and thicknesses. The obtained results are applied to calculate the spectra of transmitted radiation in cases of polyenergetic beams similar to those encountered in clinical practice. Results: The significant contribution of secondary radiation in the radiation found behind shields is demonstrated across a wide range of energies, barrier thicknesses and barrier materials. Conclusion: It is shown that characterizing shields using their response to monoenergetic pencil beams provides a useful tool for characterizing radiation behind shields in a wide range of practical situations.

Unsupervised Domain Adaptation (UDA) aims to transfer knowledge learned from a labeled source domain to an unlabeled target domain. While UDA methods for synthetic to real-world domains (synth-to-real) show remarkable performance in tasks such as semantic segmentation and object detection, very few were proposed for instance segmentation in the field of vision-based autonomous driving, and the existing ones are based on a suboptimal baseline, which severely limits the performance. In this paper, we introduce UDA4Inst, a strong baseline of synth-to-real UDA for instance segmentation. UDA4Inst adopts cross-domain bidirectional data mixing at the instance level to effectively utilize data from both source and target domains. Rare-class balancing and category module training are also employed to further improve the performance. It is worth noting that we are the first to demonstrate results on two new synth-to-real instance segmentation benchmarks, with 39.0 mAP on UrbanSyn->Cityscapes and 35.7 mAP on Synscapes->Cityscapes. Our method outperforms the source-only Mask2Former model by +7 mAP and +7.6 mAP, respectively. On SYNTHIA->Cityscapes, our method improves the source-only Mask2Former by +6.7 mAP, achieving state-of-the-art results.Our code will be released soon.

Vision-based end-to-end driving models trained by imitation learning can lead to affordable solutions for autonomous driving. However, training these well-performing models usually requires a huge amount of data, while still lacking explicit and intuitive activation maps to reveal the inner workings of these models while driving. In this paper, we study how to guide the attention of these models to improve their driving quality and obtain more intuitive activation maps by adding a loss term during training using salient semantic maps. In contrast to previous work, our method does not require these salient semantic maps to be available during testing time, as well as removing the need to modify the model's architecture to which it is applied. We perform tests using perfect and noisy salient semantic maps with encouraging results in both, the latter of which is inspired by possible errors encountered with real data. Using CIL++ as a representative state-of-the-art model and the CARLA simulator with its standard benchmarks, we conduct experiments that show the effectiveness of our method in training better autonomous driving models, especially when data and computational resources are scarce., Comment: Accepted for publication at the 35th IEEE Intelligent Vehicles Symposium (IV 2024)

Recent methods using Reinforcement Learning (RL) have proven to be successful for training intelligent agents in unknown environments. However, RL has not been applied widely in real-world robotics scenarios. This is because current state-of-the-art RL methods require large amounts of data to learn a specific task, leading to unreasonable costs when deploying the agent to collect data in real-world applications. In this paper, we build from existing work that reshapes the reward function in RL by introducing a Control Lyapunov Function (CLF), which is demonstrated to reduce the sample complexity. Still, this formulation requires knowing a CLF of the system, but due to the lack of a general method, it is often a challenge to identify a suitable CLF. Existing work can compute low-dimensional CLFs via a Hamilton-Jacobi reachability procedure. However, this class of methods becomes intractable on high-dimensional systems, a problem that we address by using a system decomposition technique to compute what we call Decomposed Control Lyapunov Functions (DCLFs). We use the computed DCLF for reward shaping, which we show improves RL performance. Through multiple examples, we demonstrate the effectiveness of this approach, where our method finds a policy to successfully land a quadcopter in less than half the amount of real-world data required by the state-of-the-art Soft-Actor Critic algorithm.

We revisit the concepts of acyclic orderings and number of acyclic orderings of acyclic digraphs in terms of dispositions and counters for arbitrary multidigraphs. We prove that when we add a sequence of nested directed paths to a directed graph there is a unique polynomial such that the generatrix function of the family of counters is the product of the polynomial and the exponential function. We give an application, by considering a kind of digraphs arranged in rows introduced by the authors in a previous paper, called dispositional digraphs, in the particular case in which the digraph has two rows, to obtain new families of linear differential equations of small order whose coefficients are polynomials of small degree which admit polynomial solutions. In particular, we obtain a new differential equation associated to Catalan numbers, and the corresponding associated polynomials, which are solution of this differential equation; we term them Catalan differencial equation and Catalan polynomials, respectively. We prove that the Catalan polynomials obtained when we connect the directed path to the second vertex of the lower row of the digraph are orthogonal polynomials for an appropriate weight function. We characterize the digraphs that maximize the counter of connected dispositional digraphs and we find a new differential equation associated to these digraphs. We introduce also dispositions and counters in any multidigraph with non-strict inequalities in the dispositions, and we find new differential equations associated to some of them.

Neural ordinary differential equations (neural ODEs) have emerged as a natural tool for supervised learning from a control perspective, yet a complete understanding of their optimal architecture remains elusive. In this work, we examine the interplay between their width $p$ and number of layer transitions $L$ (effectively the depth $L+1$). Specifically, we assess the model expressivity in terms of its capacity to interpolate either a finite dataset $D$ comprising $N$ pairs of points or two probability measures in $\mathbb{R}^d$ within a Wasserstein error margin $\varepsilon>0$. Our findings reveal a balancing trade-off between $p$ and $L$, with $L$ scaling as $O(1+N/p)$ for dataset interpolation, and $L=O\left(1+(p\varepsilon^d)^{-1}\right)$ for measure interpolation. In the autonomous case, where $L=0$, a separate study is required, which we undertake focusing on dataset interpolation. We address the relaxed problem of $\varepsilon$-approximate controllability and establish an error decay of $\varepsilon\sim O(\log(p)p^{-1/d})$. This decay rate is a consequence of applying a universal approximation theorem to a custom-built Lipschitz vector field that interpolates $D$. In the high-dimensional setting, we further demonstrate that $p=O(N)$ neurons are likely sufficient to achieve exact control., Comment: 16 pages, 10 figures, double column

Pedestrian intention prediction is crucial for autonomous driving. In particular, knowing if pedestrians are going to cross in front of the ego-vehicle is core to performing safe and comfortable maneuvers. Creating accurate and fast models that predict such intentions from sequential images is challenging. A factor contributing to this is the lack of datasets with diverse crossing and non-crossing (C/NC) scenarios. We address this scarceness by introducing a framework, named ARCANE, which allows programmatically generating synthetic datasets consisting of C/NC video clip samples. As an example, we use ARCANE to generate a large and diverse dataset named PedSynth. We will show how PedSynth complements widely used real-world datasets such as JAAD and PIE, so enabling more accurate models for C/NC prediction. Considering the onboard deployment of C/NC prediction models, we also propose a deep model named PedGNN, which is fast and has a very low memory footprint. PedGNN is based on a GNN-GRU architecture that takes a sequence of pedestrian skeletons as input to predict crossing intentions.

It is well known that the visual information represented in the simple cells of the primary visual cortex V1 is spatially localized, orientation-sensitive and bandpass-filtered. In addition, the visual information represented is subsampled. On the other hand, the methods used to assess image quality of x-ray systems performing detectability tasks are usually carried out in the spatial or image domain. This study examines the behavior of two observers, IBO and NPW, employed in the evaluation of imaging systems, in a domain that simulates the representation of visual information in V1 to gain a better understanding of the degree of optimization in human observer detectability tasks. The comparison of the two observers is conducted on images of a contrast-detail phantom and is performed in both the spatial and the wavelet domains. Furthermore, the study examines the covariance matrices of the noise. The strong covariance exhibited in the noise among neighboring pixels in the spatial domain results in covariance matrices with relatively high values outside the main diagonal, leading to significant differences between the NPW and IBO observers. In the transformed domain, the band-pass decimated filters have substantially reduced the covariance between neighboring pixels, resulting in the convergence of the detectability indices of both observers. The vast amount of information captured by the retina necessitates an optimization of informational nature. This strategy eliminates low-frequency correlations in the image, after bandpass filtering carried out by simple cells in V1. This loss of information could contribute to the suboptimal performance of the human observer. Interestingly, in V1 the differences between IBO and NPW observers are very small. Therefore, either NPW does not accurately represent the human observer in V1, or the human observer is close to optimal in this domain., Comment: 17 pages, 8 figures, submitted to Medical Physics

Classification of $N$ points becomes a simultaneous control problem when viewed through the lens of neural ordinary differential equations (neural ODEs), which represent the time-continuous limit of residual networks. For the narrow model, with one neuron per hidden layer, it has been shown that the task can be achieved using $O(N)$ neurons. In this study, we focus on estimating the number of neurons required for efficient cluster-based classification, particularly in the worst-case scenario where points are independently and uniformly distributed in $[0,1]^d$. Our analysis provides a novel method for quantifying the probability of requiring fewer than $O(N)$ neurons, emphasizing the asymptotic behavior as both $d$ and $N$ increase. Additionally, under the sole assumption that the data are in general position, we propose a new constructive algorithm that simultaneously classifies clusters of $d$ points from any initial configuration, effectively reducing the maximal complexity to $O(N/d)$ neurons., Comment: 26 pages, 10 figures

We introduce UrbanSyn, a photorealistic dataset acquired through semi-procedurally generated synthetic urban driving scenarios. Developed using high-quality geometry and materials, UrbanSyn provides pixel-level ground truth, including depth, semantic segmentation, and instance segmentation with object bounding boxes and occlusion degree. It complements GTAV and Synscapes datasets to form what we coin as the 'Three Musketeers'. We demonstrate the value of the Three Musketeers in unsupervised domain adaptation for image semantic segmentation. Results on real-world datasets, Cityscapes, Mapillary Vistas, and BDD100K, establish new benchmarks, largely attributed to UrbanSyn. We make UrbanSyn openly and freely accessible (www.urbansyn.org)., Comment: The UrbanSyn Dataset is available in http://urbansyn.org/

We present a sample dataset featuring pedestrians generated using the ARCANE framework, a new framework for generating datasets in CARLA (0.9.13). We provide use cases for pedestrian detection, autoencoding, pose estimation, and pose lifting. We also showcase baseline results. For more information, visit https://project-arcane.eu/.

Monocular Depth Estimation (MDE) is performed to produce 3D information that can be used in downstream tasks such as those related to on-board perception for Autonomous Vehicles (AVs) or driver assistance. Therefore, a relevant arising question is whether the standard metrics for MDE assessment are a good indicator of the accuracy of future MDE-based driving-related perception tasks. We address this question in this paper. In particular, we take the task of 3D object detection on point clouds as a proxy of on-board perception. We train and test state-of-the-art 3D object detectors using 3D point clouds coming from MDE models. We confront the ranking of object detection results with the ranking given by the depth estimation metrics of the MDE models. We conclude that, indeed, MDE evaluation metrics give rise to a ranking of methods that reflects relatively well the 3D object detection results we may expect. Among the different metrics, the absolute relative (abs-rel) error seems to be the best for that purpose., Comment: 11 pages, 8 figures

On end-to-end driving, human driving demonstrations are used to train perception-based driving models by imitation learning. This process is supervised on vehicle signals (e.g., steering angle, acceleration) but does not require extra costly supervision (human labeling of sensor data). As a representative of such vision-based end-to-end driving models, CILRS is commonly used as a baseline to compare with new driving models. So far, some latest models achieve better performance than CILRS by using expensive sensor suites and/or by using large amounts of human-labeled data for training. Given the difference in performance, one may think that it is not worth pursuing vision-based pure end-to-end driving. However, we argue that this approach still has great value and potential considering cost and maintenance. In this paper, we present CIL++, which improves on CILRS by both processing higher-resolution images using a human-inspired HFOV as an inductive bias and incorporating a proper attention mechanism. CIL++ achieves competitive performance compared to models which are more costly to develop. We propose to replace CILRS with CIL++ as a strong vision-based pure end-to-end driving baseline supervised by only vehicle signals and trained by conditional imitation learning., Comment: This paper has been accepted to the 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2023)

In this work we propose a new non-monotonic activation function: the modulus. The majority of the reported research on nonlinearities is focused on monotonic functions. We empirically demonstrate how by using the modulus activation function on computer vision tasks the models generalize better than with other nonlinearities - up to a 15% accuracy increase in CIFAR100 and 4% in CIFAR10, relative to the best of the benchmark activations tested. With the proposed activation function the vanishing gradient and dying neurons problems disappear, because the derivative of the activation function is always 1 or -1. The simplicity of the proposed function and its derivative make this solution specially suitable for TinyML and hardware applications., Comment: Accepted at Engineering Applications of AI