Bachelor in Computer Science, Master in Artificial Intelligence (AI), and PhD in AI. I am a senior researcher in Computer Vision (CV). I have managed several CV and robotic projects that have been presented in symposiums and journals. I also contribute to important libraries such as OpenCV.
Among a large list of projects I would like to highlight my contributions in The Graffter, where I developed Augmented Reality technology for urban environment and XOIA Soft. Dev., a company that I have co-founded and nowadays has more than 10 employees. I led the CV team of New Horizon Technologies, developing several video surveillance, and medical image projects.
At the moment, I am a Senior Machine Learning Engineer at Qualcomm XR Labs Europe, where we are creating a new generation of AR/VR headsets by using cutting-edge Deep Learning technology to improve SLAM and SfM.
GlueStick: Robust Image Matching by Sticking Points and Lines Together
*Equal contribution. ICCV 2023.
Paper | Code | Project page | Colab Demo
ELSED: Enhanced Line SEgment Drawing
Pattern Recognition, 127, 108619.
Paper | Code (C++) | DOI | Project page
Revisiting Binary Local Image Description for Resource Limited Devices
IEEE Robotics and Automation Letters, 6(4), 8317-8324.
Paper | Code (C++) | DOI | Project page | Video
BEBLID: Boosted efficient binary local image descriptor
Pattern Recognition Letters, 133, 366-372.
FSG: A statistical approach to line detection via fast segments grouping
International Conference on Intelligent Robots and Systems (IROS) (pp. 97-102).
Paper | Code (C++) | DOI | Dataset | Video
During three amazing months I have been working in CVG group of ETH under the supervision of Prof. Marc Pollefeys. I developed a project about geometric-aware line segment matching for SLAM systems.
In my PhD I created and improved the basics of Augmented Reality to move it towards Urban Environments. My research is centered in efficient detectors and descriptors for local image features such as corners or line segments where we have already published two articles. We are also working with RANSAC based geometric techniques to improve the estimation of the device pose in the world.
thesis, slidesEmail: iagosuarz@gmail.com