Azin Jahedi

I designed and developed multiple novel state-of-the-art deep learning methods for motion estimation in both supervised and unsupervised settings, achieving up to 50% accuracy improvement over baseline methods through innovative model design, training strategies, and large-scale experimentation leading to publications in top-tier computer vision conferences and journals.

-Reseach in visual odometry -Software development

Lab assistant for an automated plenoptic recording.

I develop pieces of software to solve some computer vision tasks. I was given synthesised image sequences of evolving objects. I had to track the objects and detect the important events regarding the objects. (using OpenCV in C++) At the end, I had to visualize the detected events using d3js.

-Perusing papers in Computer Vision: Face recognition and facial expression recognition -Analyzing the requirements of the applications -Developing pieces of software

Tutor and invigilator of the course Operating System Principle and Design for four semesters. Grading the assignments and the mid-term exams.

Tutor and invigilator of the course Artificial Intelligence for one semester. Grading the assignments.

Doctoral researcher in optical flow estimation.

Computer Vision Imaging Science, Detection and Pattern Recognition, Machine Learning & Multimodal Interaction for Ubiquitous Computers. Research Project and Thesis (Improved Descriptor Learning for Correspondence Problems) in Computer Vision & Deep Learning