A5041283253- Catalysis and Hydrodesulfurization Studies
- Catalytic Processes in Materials Science
- Advanced Vision and Imaging
- Optical measurement and interference techniques
- Robotics and Sensor-Based Localization
- Human Pose and Action Recognition
- Nanomaterials for catalytic reactions
- Catalysis and Oxidation Reactions
- Zeolite Catalysis and Synthesis
- 3D Shape Modeling and Analysis
- Petroleum Processing and Analysis
- Generative Adversarial Networks and Image Synthesis
- Computer Graphics and Visualization Techniques
- Advanced Neural Network Applications
- Metal Extraction and Bioleaching
- Face recognition and analysis
- Image and Object Detection Techniques
- Video Surveillance and Tracking Methods
- Robot Manipulation and Learning
- Human Motion and Animation
- Chemical Synthesis and Characterization
- Remote Sensing and LiDAR Applications
- Catalysts for Methane Reforming
- Corrosion Behavior and Inhibition
- Video Analysis and Summarization
Universitat Politècnica de Catalunya
2016-2025
Institut de Robòtica i Informàtica Industrial
2016-2025
Consejo Superior de Investigaciones Científicas
1991-2023
Universidad de Zaragoza
1970-2015
Instituto de Catálisis y Petroleoquímica
1998-2009
Universidad Autónoma de Madrid
1995-2003
University of Concepción
1993
Low textured scenes are well known to be one of the main Achilles heels geometric computer vision algorithms relying on point correspondences, and in particular for visual SLAM. Yet, there many environments which, despite being low textured, can still reliably estimate line-based primitives, instance city indoor scenes, or so-called "Manhattan worlds", where structured edges predominant. In this paper we propose a solution handle these situations. Specifically, build upon ORB-SLAM,...
We present a novel approach for synthesizing photorealistic images of people in arbitrary poses using generative adversarial learning. Given an input image person and desired pose represented by 2D skeleton, our model renders the same under new pose, views parts visible hallucinating those that are not seen. This problem has recently been addressed supervised manner [16, 35], i.e., during training ground truth given to network. go beyond these approaches proposing fully unsupervised...
We propose a new approach to simultaneously recover camera pose and 3D shape of non-rigid potentially extensible surfaces from monocular image sequence. For this purpose, we make use the Extended Kalman Filter based Simultaneous Localization And Mapping (EKF-SLAM) formulation, Bayesian optimization framework traditionally used in mobile robotics for estimating reconstructing rigid scenarios. In order extend problem deformable domain represent object's surface mechanics by means Navier's...
We propose an online solution to non-rigid structure from motion that performs camera pose and 3D shape estimation of highly deformable surfaces on a frame-by-frame basis. Our method models deformations as linear combination some mode shapes obtained using modal analysis continuum mechanics. The is first discretized into elastic triangles, modelled by means finite elements, which are used the force balance equations for undamped free vibrations model. basis computation comes down solving...
We propose a method for predicting the 3D shape of deformable surface from single view. By contrast with previous approaches, we do not need pre-registered template surface, and our is robust to lack texture partial occlusions. At core approach geometry-aware deep architecture that tackles problem as usually done in analytic solutions: first perform 2D detection mesh then estimate geometrically consistent image. train this an end-to-end manner using large dataset synthetic renderings shapes...
We present a comprehensive framework for studying and leveraging morphological symmetries in robotic systems. These are intrinsic properties of the robot’s morphology, frequently observed animal biology robotics, which stem from replication kinematic structures symmetrical distribution mass. illustrate how these extend to state space both proprioceptive exteroceptive sensor measurements, resulting equivariance equations motion optimal control policies. Thus, we recognize as relevant...
In this paper, we propose a sequential solution to simultaneously estimate camera pose and non-rigid 3D shape from monocular video. contrast most existing approaches that rely on global representations of the shape, model object at local level, as an ensemble particles, each ruled by linear equation Newton's second law motion. This dynamic is incorporated into bundle adjustment framework, in combination with simple regularization components ensure temporal spatial consistency estimated...
© 2014. The copyright of this document resides with its authors. This paper describes a sequential solution to dense non-rigid structure from motion that recovers the camera and 3D shape objects by processing monocular image sequence as data arrives. We propose model time-varying probabilistic linear subspace mode shapes obtained continuum mechanics. To efficiently encode deformations contain large number mesh vertexes, we compute deformation modes on down sampled rest using finite element...
This paper introduces an approach to simultaneously estimate 3D shape, camera pose, and object type of deformation clustering, from partial 2D annotations in a multi-instance collection images. Furthermore, we can indistinctly process rigid non-rigid categories. advances existing work, which only addresses the problem for one single or, if multiple objects are considered, they assumed be clustered priori. To handle this broader version problem, model using formulation based on unions...