Experience

  1. Machine Learning Intern

    Zenseact
    • Developed an interest point detector and descriptor, based on the SuperPoint model, to excel on fisheye images of specific environments, using Python and PyTorch
    • Trained the model on a mix of COCO images with homographies and fisheye images with fisheye warps
    • The model outperformed SIFT and the offical pretrained SuperPoint, achieving lower reprojection errors when matching points between fisheye images and their warped versions

  2. Algorithms Intern

    Neko health
    • Designed setups and algorithms for calibrating thermal cameras with RGB-D cameras, using Python and OpenCV
    • Captured the same calibration target with both cameras to get 2D-3D correspondences
    • Estimated the rigid transformation between the cameras, enabling fused captures to generate 3D thermal images

Education

  1. PhD Computer Vision

    Lund University

  2. MSc and BSc in Computer Science

    Lund University

    Combined BSc and MSc program. Specialized in computer vision and machine learning for my masters.

    GPA: 4.65/5.0

    Read Master's Thesis