This project was part of the Technology Design Project unit in the field of computer vision at my university. It served as a starting point for live image segmentation tasks across various domains, pending future training or fine-tuning on domain-specific datasets.
I developed U-Net model trained on Cityscapes Dataset using TensorFlow, starting from data preparation, model creation, training, evaluation, static prediction, and live prediction.