Inspiration
As someone who recently became a father I wanted to create something that parents and children could do together that is playful, encourages exploration, and can be used to teach children words.
What it does
This app asks the user to take a photo of a common object in the world, a cat, a tree, a flower, etc. and uses the Google Vision API to check if the picture contains the requested subject. If it does then the user is rewarded with money being added to their Square gift card. The aim of the app is to let parents and children search out the subject for their photo together and for the children to learn the names of things.
How we built it
The app is built in Unity and uses the Google Vision API object detection function to establish if a photo contains a requested object.
The app contains a list of words that are detectable by the Google Vision API and selects them in a semi-random order for the user to find and take a photo of. The app takes an image from the Camera in Unity and uploads it to the Google Vision API where a list of objects in the photo are returned.
If the Google Vision API returns the requested object then a bonus is added to the user’s Square gift card using the Square API.
Whether the response is positive or negative the app also uses the Google Text to Speech AI API to create a vocal response which congratulates the user, or asks them to try again depending on the results.
Challenges we ran into
Narrowing down the vision api objects to something that would reliably return results was a challenge. Finding objects that the Vision API would consistently identify in the same way, were also quite accessible to most people, and without reducing the list too much to prevent repetition took a lot of testing.
Accomplishments that we're proud of
The simple but engaging user experience, especially when playing along with a child is a very enjoyable experience.
What we learned
Very simple interactions and challenges can still be engaging.
What's next for Little AI Photographer
It would be nice to use AI to identify the location/environment of the user so that the object requests could be better suited to the user’s specific surroundings.
Log in or sign up for Devpost to join the conversation.