Work & Research

Work Interests:

My professional experience is focused on applications of 3D computer vision and multi-view geometry in solving real-world problems such as camera localization, stereo vision and geometry acquisition. More specifically, I have been working on the problem of localizing a single camera to high definition (HD) maps for autonomous vehicles. See below for short videos of our work.

 

 

I enjoy coding in C++ and use it on a daily basis for writing production code for safety-critical and resource constrained systems (read: self-driving cars!). I have extensive prior work experience with C# and Java. More recently, I’ve picked up Python and primarily use it for prototyping ideas and my day-to-day research in deep learning (see below).

Research Interests:

My primary research interests lie in human-computer interaction (HCI) and computer vision. My PhD dissertation is focused on intuitive, yet fast and efficient recurrent neural networks (RNN) for gesture recognition, as well as generative adversarial networks (GANs) for synthetic gesture generation.

Previously, my research at UCF was focused on leveraging sketch-based interfaces to support various robotic manipulation tasks. Our goal was to explore how sketch-based interfaces could be used to assist robots through geometry extraction, behavior modification, and guided autonomy. This research was supported by NSF Award IIS-1638060.

Education:

  • (2019) PhD Candidate, University of Central Florida (Orlando, FL, USA)
  • (2017) Master of Science in Computer Science (MS along the way), University of Central Florida (Orlando, FL, USA)
  • (2014) Master of Science in Computer Science, Brock University (St. Catharines, ON, Canada)
  • (2011) Bachelor of Science in Software Engineering, Sheikhbahaee University (Esfahan, Esfahan, Iran)

Professional Experience:

  • (Since 2017)  Senior DriveWorks Computer Vision Engineer, NVIDIA Corporation
  • (2016)  Computer Vision Developer Technology Intern, NVIDIA Corporation

Patents:

  • (2018) Eugene M. Taranta II, Mehran Maghoumi, Corey Pittman, Joseph J. LaViola Jr. “Synthetic Data Generation of Time Series Data”, US Patent 10,133,949

Publications:

  • (2020) Eugene M. Taranta II, Corey Pittman, Jack Oakley, Mykola Maslych, Mehran Maghoumi, Joseph J. LaViola Jr. “Moving Toward an Ecologically Valid Data Collection Protocol for 2D Gestures In Video Games“, Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems (CHI 2020). (Acceptance rate: 24.3%)
    [PDF]
  • (2019) Fabio Marco Caputo, S Burato, G Pavan, T Voillemin, Hazem Wannous, Jean-Philippe Vandeborre, Mehran Maghoumi, EM Taranta, A Razmjoo, JJ LaViola Jr, F Manganaro, S Pini, G Borghi, R Vezzani, R Cucchiara, H Nguyen, MT Tran, A Giachetti. “SHREC 2019 Track: Online Gesture Recognition“, 12th Eurographics Workshop on 3D Object Retrieval (3DODR 2019)
    Winner of the SHREC 2019 Challenge on Online Gesture Recognition
    [PDF]
  • (2019) Mehran Maghoumi, Joseph J. LaViola Jr. “DeepGRU: Deep Gesture Recognition Utility“, Proceedings of the 14th International Symposium on Visual Computing (ISVC 2019)
    [PDF] [Slides] [Code]
  • (2018) Mehran Maghoumi, Joseph J. LaViola Jr., Karthik Desingh, Odest Chadwicke Jenkins. “GemSketch: Interactive Image-Guided Geometry Extraction from Point Clouds“,  Proceedings of 2018 International Conference on Robotics and Automation (ICRA 2018).
    [PDF] [Supplementary] [Video]
  • (2017) Pooya Khaloo, Mehran Maghoumi, Eugene Taranta II, David Bettner, Joseph J. Laviola Jr. “Code Park: A New 3D Code Visualization Tool“, 2017 IEEE Working Conference on Software Visualization (IEEE VISSOFT 2017).
    [PDF] [Video]
  • (2017) Eugene M. Taranta II, Amirreza Samiei, Mehran Maghoumi, Pooya Khaloo, Corey R. Pittman, Joseph J. LaViola Jr. “Jackknife: A Reliable Recognizer with Few Samples and Many Modalities“, Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems (CHI 2017).  (Acceptance rate: 25.0%)
    Received SIGCHI Best of CHI Honorable Mention Award (top 5%)
    [PDF] [Project Page] [Code]
  • (2016) Eugene M. Taranta II, Mehran Maghoumi, Corey Pittman, Joseph J. LaViola Jr. “A Rapid Prototyping Approach to Synthetic Data Generation For Improved 2D Gesture Recognition“, Proceedings of the 29th Annual Symposium on User Interface Software and Technology (UIST 2016)(Acceptance rate: 20.6%)
    [PDF] [Demo Video] [Presentation Video] [Project Page]
  • (2016) Karthik Desingh, Mehran Maghoumi, Odest Chadwicke Jenkins, Joseph J. LaViola Jr., Lionel Reveret. “Object Manipulation in Cluttered Scenes Informed by Physics and Sketching“, RSS 2016 Workshop: Geometry and Beyond – Representations, Physics, and Scene Understanding for Robotics
    [PDF]
  • (2014) Mehran Maghoumi and Brian J. Ross. “A Comparison of Genetic Programming Feature Extraction Languages for Image Classification“, IEEE SSCI 2014
    [PDF]

Theses:

  • (2014) Real-time Automatic Object Classification and Tracking using Genetic Programming and NVIDIA CUDA, M.Sc. thesis
    [Project Page]
  • (2011) Implementation of a Fuzzy Decision Making Unit for the Nao Robot, undergraduate thesis, supervisor: Dr. Nasser Ghasem-Aghaee

Teaching Experience:

  • (2014) CDA 3103 – Computer Logic and Organization
  • (2014) CNT 3004 – Computer Networks
  • (2014) COSC 1P03 – Data Structures and Abstraction
  • (2013) COSC 3P98 – Computer Graphics
  • (2013) COSC 3P01 – Introduction to Computer Networking
  • (2012) APCO 1P01 – Applied Computing and Fluency with Technology
  • (2012) COSC 1P02 – Introduction to Computer Science

Awards and Honors: