The problem CrimeClix solves

It solves the following problems:
It can be used to identify and catch criminals given their textual descriptions
It can be used TO INVESTIGATE MISSING PEOPLE CASES
It will help investigation agencies to get geographic location of criminals based on which camera detected them.
It will give alerts based on criminal detection in real time
It can be used to convert criminal sketches into approximate face pictures using machine learning algorithms
It uses decentralized blockchain technology to store face images of criminals and therefore is highly secure.
It uses state of the art YOLO V4 algorithms for fast face detection in real time.

Challenges we ran into

The first challenge that we ran into was deciding a sweet spot between accuracy and time required to train the machine learning model for text to image synthesis. We have to pick a model architecture that was fairly accurate while having low training and testing time.

The second challenge that we ran into was integrating different subparts of the software together. It required all the subparts to meet some commom architecture so that integration could be done with ease. We used several API's for information exchange inside the software. Several endpoints were created for data transfer.

We also had some trouble with uploading images to IPFS. We had to go through documentation extensively to resolve all the errors and dependencies issues. We explored a lot of solutions before finally getting onto an optimal one.