Woosh. Accelerated Novelty Detection

Dear readers,

Thanks for your interest in my submission to the Xilinx x hackster.io adaptive compute competition 2020.

This being my first hackster.io project, I was unaware that this form is the desired way of telling the story of my project, hence all my project files are packaged in the file nd_zcu104.zip in the attachments. Instructions on how to build this are enclosed. I hope this is fine.

This file is my starting point. If you are mostly interested in results, simply play the videos in the demo folder. One of these is shown below.

They show objects on streets that don't belong there, and my pipeline (mostly) finds them in an unsupervised manner.

If you are interested in my project description (story-style), head over to xlnx_report.pdf

This is not very technical, I put all those details into the README in the top level. It will guide you through all my code.If you are interested in simply copying the stuff onto the ZCU104 and run it, the target folder is where you're after.

Thanks and have a nice day.