Initially, our idea was to use Rekognition, an amazon web service's AI that could recognize objects and describe their confidence as a percentage. Being committed to making AWS's AI work on our computer, we spent 10 hours troubleshooting just to run into a series of problems. We sat with the members of AWS for 2 hours just to see if there was anything we could do. Seeing as there was no way to get this running, we eventually had to move on. Determined to stick to the idea, we ended up using an AWS server to train our own algorithm to recognize objects.
The program currently has a login page that, when completed, directs you to the application where you can enter an image and see whether the AI can recognize it.
We used OpenCV as inspiration for how to train the AI. We're proud of the app that we were able to create and how much we learned. We have a better understanding of AWS and can definitely implement it in the workforce when we can properly access it.
For the next steps, we want to properly implement the AWS AI into the application in order to have functional recognition software. From there, we will use the object it recognizes to add additional functionality; for example, estimating the price of your groceries depending on the image that you sent.
Log in or sign up for Devpost to join the conversation.