FlorAI

Inspiration

We were also inspired by the game Pokémon, a game that encourages players to explore the world and find new creatures. Players carry around a device called a Pokédex, a device that keeps a catalog of all the creatures you have encountered. We took this idea an applied it as an app for users, encouraging players to explore their environment outside and discover flowers and plants.

What it does

Users are able to use the app to identify flowers they encounter in nature. The app keeps track of a library of all the different kinds of flowers they have encountered.

How we built it

We trained an AI model to identify various flowers and built our app around it. We trained the model using TensorFlow and then converted to Apple's CoreML model to integrate with Swift. for the front end app design.

Challenges we ran into

In early stages of training the model, it made inaccurate predictions and took a significant time to train. We retrained our model using TensorFlow and the transfer learning technology. We used a pre-trained model and customised for our use. This increased our model's accuracy and overall performance.

Another challenge was communicating progress on each task with each other. Although we tried to divide-and-conquer, we weren't able to explain our progress to each other. This caused us to undo certain things and then redo them together.

Accomplishments that we're proud of

Our model achieved an overall 85 % accuracy and works well.

What we learned

Communication is key. As a team, we need to communicate and make sure we're on the same page.

What's next for FlowAI

Identifying sounds of various birds and insects in the wild to help backpackers identify the wilderness around them.

Built With

• coreml
• swift
• tensorflow