Intention prediction (IP) is a challenging task for intelligent vehicle’s perception systems. IP provides the likelihood, or probability, of a target vehicle to perform a maneuver subjected to a finite set of possibi...
Intention prediction (IP) is a challenging task for intelligent vehicle’s perception systems. IP provides the likelihood, or probability, of a target vehicle to perform a maneuver subjected to a finite set of possibilities. There are many factors that influence the decision-making process of a driver, which should be considered in a prediction framework. In addition, the lack of labeled large-scale dataset with maneuver annotation imposes another challenge to the task. This paper proposes an Interaction-aware Maneuver Prediction framework, called IAMP, using interaction graphs to extract complex interaction features from traffic scenes. In addition, we explored a semi-supervised approach called Noisy Student to take advantage of unlabeled data in the training step. Experimental results show relevant improvement when using unlabeled data, increasing the average performance of a classifier by 7.17% of accuracy. Moreover, this approach also made it possible to obtain an intention predictor with similar results to a classifier., even when using a shorter observation horizon.
It is important to figure out the patterns of woven fabrics before producing woven fabric with a machine. Recognition of woven fabric pattern usually with the help of the human eye can understand the fabric pattern. H...
详细信息
Hyperbolic machine learning is an emerging field aimed at representing data with a hierarchical structure. However, there is a lack of tools for evaluation and analysis of the resulting hyperbolic data representations...
详细信息
We address the problem of (a) predicting the trajectory of an arm reaching motion, based on a few seconds of the motion’s onset, and (b) leveraging this predictor to facilitate shared-control manipulation tasks, easi...
详细信息
As dynamical systems equipped with neural network controllers (neural feedback systems) become increasingly prevalent, it is critical to develop methods to ensure their safe operation. Verifying safety requires extend...
详细信息
Learning to assemble geometric shapes into a larger target structure is a pivotal task in various practical applications. In this work, we tackle this problem by establishing local correspondences between point clouds...
Learning to assemble geometric shapes into a larger target structure is a pivotal task in various practical applications. In this work, we tackle this problem by establishing local correspondences between point clouds of part shapes in both coarse- and fine-levels. To this end, we introduce Proxy Match Transform (PMT), an approximate high-order feature transform layer that enables reliable matching between mating surfaces of parts while incurring low costs in memory and compute. Building upon PMT, we introduce a new framework, dubbed Proxy Match TransformeR (PMTR), for the geometric assembly task. We evaluate the proposed PMTR on the large-scale 3D geometric shape assembly benchmark dataset of Breaking Bad and demonstrate its superior performance and efficiency compared to state-of-the-art methods. Project page: https://***/pmtr.
Beneficence is a social phenomenon that has rarely been modeled computationally so far. In this paper, we propose to study the beneficence of online opinions and comments published on social media on essential topics ...
Beneficence is a social phenomenon that has rarely been modeled computationally so far. In this paper, we propose to study the beneficence of online opinions and comments published on social media on essential topics for society. Our computational approach is based on measuring semantic similarity. We apply three measures to assess the beneficence of $\sim 41 K$ social media users: average Confidence, Normalized Google Distance, and Pointwise Mutual Information. As a use case, we analyze opinions on the topic of vaccinations on Facebook, where two distinct groups (Pro-Vax vs. Anti-Vax) are present. The results reveal a shared connection to beneficence among social media users, with both groups exhibiting similar levels of similarity and no significant clustering into echo chambers.
Partial discharge (PD) is a widespread phenomenon instigated in power transformer (PT) insulation systems. PDs are triggered by voids that vary in size and position within the PT insulation. The electrical characteris...
详细信息
This paper addresses the problem of detecting humans in RGB and Thermal (long-wave IR) images taken by cameras mounted onboard a mobile robot. Human/Pedestrian detection is currently one of the most pertinent object d...
This paper addresses the problem of detecting humans in RGB and Thermal (long-wave IR) images taken by cameras mounted onboard a mobile robot. Human/Pedestrian detection is currently one of the most pertinent object detection problems, mainly due to safety concerns in autonomous vehicles. The majority of approaches apply deep-learning techniques based solely on RGB images. However, they have a few shortcomings, namely that during foggy weather, nighttime, and low-light scenarios, these images may not contain sufficient information. To address these issues, this work studies the use of thermal cameras as a complementary source of information for human detection in indoor and outdoor environments. The proposed approach uses YOLOv5 to detect pedestrians in both thermal and RGB images. Moreover, the different modalities are combined using early and late fusion techniques. Evaluation of the proposed approach is carried out in the FLIR Aligned dataset and in a new in-house dataset. Results indicate that the use of fusion techniques highlights a promising way to improve the overall performance in this application domain.
This paper addresses the task of predicting the behavior of traffic participants, which involves complexities such as road geometry and agent interactions. To overcome these challenges, this paper presents a novel fra...
This paper addresses the task of predicting the behavior of traffic participants, which involves complexities such as road geometry and agent interactions. To overcome these challenges, this paper presents a novel framework called AIMP (Attention-based Interaction-aware Maneuver Prediction). AIMP utilizes interaction graphs to extract intricate interaction features from traffic scenes. The framework incorporates a Gated Mixture-of-Experts Attention Mechanism, which combines information from road geometry, interaction patterns, and motion dynamics. This fusion process also considers prior maneuver intention estimations, enhancing both explainability and informativeness. Experimental results highlight a performance enhancement (approximately 2% ∼ 9% of accuracy) of the proposed AIMP framework compared to alternative fusion methods.
暂无评论