Computer-Vision-Based Multicamera Person Reidentification (2021)

In this project, we worked with N camera views of the same area from different angles, where some cameras captured information others didn’t. The objective of the person re-identification task was to link images of the same individual captured by different cameras and assign them the same ID. Due to the dataset's unique characteristics, where everyone wore identical clothing, deep-learning approaches were not viable, so we relied on intrinsic and extrinsic camera parameters.