Pedro Miraldo MERL Website

Pedro Miraldo

Pedro Miraldo
Mitsubishi Electric Research Labs
201 Broadway, Cambridge, MA
miraldo (at) merl (dot) com

I am a Senior Principal Research Scientist at MERL Mitsubishi Electric Research Laboratories. My research focuses on computer vision, robotics, and artificial intelligence, with a particular focus on 3D reconstruction, localization, and mapping for autonomous systems. Before joining MERL, I was a second-stage Researcher (comparable to an Assistant Research Professor) at the Institute for Systems & Robotics and the Department of Electrical & Computer Engineering, IST Instituto Superior Técnico, Lisboa. From 2018 to 2019, I was a postdoctoral associate at KTH Royal Institute of Technology. Previously, I held an FCT postdoctoral researcher grant (a highly competitive individual research grant) at IST.

I received my Master's and Ph.D. degrees in Electrical and Computer Engineering from the Faculty of Sciences and Technology, University of Coimbra, Portugal.

News:

  • Three papers accepted to IEEE/CVF Conference on Computer Vision and Pattern Recognition 2026:
    • - Revisiting Monocular SLAM with Spatio-Temporal Scene Modeling
    • - Point4Cast: Streaming Dynamic Scene Reconstruction and Forecasting (Highlight)
    • - LASER: Layer-wise Scale Alignment for Training-Free Streaming 4D Reconstruction
  • Paper accepted to the Annual Conference on Neural Information Processing Systems (NeurIPS), 2025.
  • Paper accepted to the IEEE/CVF International Conference on Computer Vision (ICCV), 2025 as Highlight.
  • Paper accepted to the International Conference on 3D Vision (3DV), 2025.

Selected Projects and Publications:

Illustration for Graduated Non-Convexity
Simultaneous Localization and Mapping [project]:
  • Revisiting Monocular SLAM with Spatio-Temporal Scene Modeling, from
    Valter Piedade, Lalit Manam, Masashi Yamazaki, and P. Miraldo
    IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026
    [arXiv, project, code, doi] (code will be released before the conference)


Illustration for Graduated Non-Convexity
Graduated Non-Convexity [project]:
  • SAC-GNC: SAmple Consensus for adaptive Graduated Non-Convexity, from
    V. Piedade, C. Sidhartha, J. Gaspar, V. M. Govindu, and P. Miraldo
    IEEE/CVF Int'l Conf. Computer Vision (ICCV), 2025
    [doi, paper link, video, code (soon)]

    Highlight paper (2.3%)


Illustration for Neural Implicit Surface Rendering
Neural Implicit Surface Rendering [project]:
  • A Probability-guided Sampler for Neural Implicit Surface Rendering, from
    G. Dias Pais, Valter Piedade, Moitreya Chatterjee, Marcus Greiff, and Pedro Miraldo
    European Conference on Computer Vision (ECCV), 2024
    [doi, merl-tr, project, video, code]



Illustration for Neural Radiance Fields for Dynamic Scenes
Neural Radiance Fields for Dynamic Scenes [project]:
  • Gear-NeRF: Free-Viewpoint Rendering and Tracking with Motion-aware Spatio-Temporal Sampling, from
    X. Liu, Y-W Tai, C-K Tang, P. Miraldo, S. Lohit and M. Chatterjee,
    IEEE/CVF Conf. Computer Vision and Pattern Recognition (CVPR), 2024
    [arXiv, merl-tr, video, code, doi]

    Highlight paper (2.8%)


Illustration for Adaptive Sample Consensus
Adaptive Sample Consensus [project]:
  • BANSAC: A dynamic BAyesian Network for adaptive SAmple Consensus, from
    V. Piedade and P. Miraldo,
    IEEE/CVF Int'l Conf. Computer Vision (ICCV), 2023.
    [arXiv, code, doi]


Illustration for Frame-to-Frame Rotation Estimation in Crowded Scenes
Frame-to-Frame Rotation Estimation in Crowded Scenes [project]:
  • Robust Frame-to-Frame Camera Rotation Estimation in Crowded Scenes, from
    F. Delattre, D. Dirnfeld, P. Nguyen, S. Scarano, M. J. Jones, P. Miraldo, and E. Learned-Miller,
    IEEE/CVF Int'l Conf. Computer Vision (ICCV), 2023.
    [arXiv, code, dataset, doi]


Illustration for Intersecting Lines for 3D Registration
Intersecting Lines for 3D Registration [project]:
  • Fast and Accurate 3D Registration from Line Intersection Constraints, from
    A. Mateus, S. Ranade, S. Ramalingam, and P. Miraldo,
    International Journal of Computer Vision (IJCV), 2023.
    [doi, code]
  • Minimal Solvers for 3D Scan Alignment with Pairs of Intersecting Lines, from
    A. Mateus, S. Ramalingam, and P. Miraldo,
    IEEE/CVF Conf. Computer Vision and Pattern Recognition (CVPR), 2020.
    [doi, video, code]
  • Mapping of Sparse 3D Data using Alternating Projection, from
    S. Ranade, X. Yu, S. Kakkar, P. Miraldo, and S. Ramalingam,
    Asian Conf. Computer Vision (ACCV), 2020.
    [arXiv, video, doi]




Illustration for Line Projections in Catadioptric Cameras
Line Projections in Catadioptric Cameras:
  • A Unified Model for Line Projections in Catadioptric Cameras with Rotationally Symmetric Mirrors, from
    P. Miraldo and Jose Pedro Iglesias
    IEEE/CVF Conf. Computer Vision and Pattern Recognition (CVPR), 2022.
    [pdf, doi, code]
  • Analytical Modeling of Vanishing Points and Curves in Catadioptric Cameras, from
    Pedro Miraldo, Francisco Eiras, and Srikumar Ramalingam
    IEEE/CVF Conf. Computer Vision and Pattern Recognition (CVPR), 2020.
    [arXiv:1804.09460, doi];


Illustration for Active Structure-from-Motion using Lines
Active Structure-from-Motion using Lines:
  • An observer cascade for velocity and multiple line estimation, from
    A. Mateus, P. U. Lima, and P. Miraldo,
    IEEE Int'l Conf. Robotics and Automation (ICRA), 2022.
    [arXiv, doi]
  • On Incremental Structure-from-Motion using Lines, from
    A. Mateus, O. Tahri, A. P. Aguiar, P. U. Lima, and P. Miraldo,
    Transactions on Robotics (T-RO), 2021. [arXiv, doi]
  • Active Estimation of 3D Lines in Spherical Coordinates, from
    A. Mateus, O. Tahri, and P. Miraldo,
    American Control Conference (ACC), 2019. [arXiv, doi]
  • Active Structure-from-Motion for 3D Straight Lines, from
    A. Mateus, O. Tahri, and P. Miraldo,
    IEEE/RSJ Int'l Conf. Intelligent Robots and Systems (IROS), 2018. [link, doi]


Illustration for 3D Registration using Deep Learning
3D Registration using Deep Learning [project]:
  • 3DRegNet: A Deep Neural Network for 3D Point Registration from
    G. D. Pais, S. Ramalingam, V. M. Govindu, J. C. Nascimento, R. Chellappa, and P. Miraldo,
    IEEE/CVF Conf. Computer Vision and Pattern Recognition (CVPR), 2020.
    [arXiv, doi, code]


Illustration for Generalized Essential Matrix
Generalized Essential Matrix
  • On the Generalized Essential Matrix Correction: An efficient solution to the problem and its applications, from
    Pedro Miraldo and Joao R. Cardoso (2020),
    Journal of Mathematical Imaging and Vision (JMIV).
    [arXiv:1709.06328, doi]
  • Generalized Essential Matrix: Properties of the Singular Value Decomposition, from
    P. Miraldo and H. Araujo (2015),
    Image and Vision Computing (IVC).
    [pdf, doi]


Illustration for General Camera Pose Estimation using Lines and Points
Using Lines and Points for General Camera Pose Estimation:
  • A Minimal Closed-Form Solution for Multi-Perspective Pose Estimation using Points and Lines, from
    P. Miraldo, T. Dias, S. Ramalingam,
    European Conf. Computer Vision (ECCV), 2018.
    [link, video]
  • Pose Estimation for General Cameras using Lines, from
    P. Miraldo, H. Araujo and N. Gonçalves,
    IEEE Trans. Cybernetics (Systems, Man, and Cybernetics, Part B), 2015.
    [pdf, doi, video]
  • Planar Pose Estimation for General Cameras using Known 3D Lines, from
    P. Miraldo and H. Araujo,
    IEEE/RSJ Int'l Conf. Intelligent Robots and Systems (IROS), 2014.
    [pdf, doi, video]


Illustration for General Camera Calibration
Smooth Camera Models: Modeling and Calibration
  • Calibration of Smooth Camera Models, from
    P. Miraldo and H. Araujo (2013),
    IEEE Trans. Pattern Analysis and Machine Intelligence (T-PAMI).
    [pdf, appendix, doi]
  • Point-based Calibration Using a Parametric Representation of General Imaging Models, from
    P. Miraldo, H. Araujo, and J. Queiro (2011),
    IEEE Int'l Conf. Computer Vision (ICCV).
    [pdf, appendix, doi]

Last updated: Feb 1, 2026