A doctoral student from our department, Jakub Pokrywka, M.Sc., took high places in the competitions at this year's IEEE International Conference on Big Data (IEEE BigData 2022) held in Japan.
Jakub proposed his solutions in the competitions: Privacy-preserving Matching of Encrypted Images and Vehicle Class and Orientation Detection Challenge in which he took second, third place respectively. In the third competition, the Crowdsensing-based Road Damage Detection Challenge, Jakub received an honorable mention in the form of an invitation to present his solution.
The first of the aforementioned competitions involved verifying a pair of images (source and encrypted) and determining whether or not it represented the same image. The competition was to verify image encryption mechanisms to ensure the privacy of information in data collected from the environment.
The Vehicle Class and Orientation Detection competition involved creating a model to detect and classify vehicles in an image. The model had to be trained on synthetic data (from a simulator), but its evaluation was carried out on real images.
The last of the competitions with the highest number of participants from around the world was the Crowdsensing-based Road Damage Detection Challenge. It involved creating a model for detecting and classifying road surface damage based on smartphone photos. The system under development is intended to replace, currently used, more expensive systems based on specialized vehicles with multiple sensors.