This project synthetically generates a blurred image dataset using the MS COCO dataset as a source. It then attempts to perform semantic segmentation on all 80 classes of the newly generated dataset.

Run the IPYNB file named "Final Dataset Generation Using Annotations (1)" in a Jupyter notebook or similar environment.

This will guide you in separating the foregrounds from the image backgrounds, applying the motion blur algorithm on the separated foreground, and then matching the new foreground with its corresponding background.