Inspiration
My team and I have been interested in Artificial Intelligence and Computer Vision from the start. We were experimenting with Object Recognition software such as OpenCV and Tesseract. Unfortunately, neither of the two played nice; it was fate that Clarifai was there to save the day.
What it does
When the user opens the app, he or she has the choice to upload a picture from the gallery or take their own picture. After they have chosen the respective picture, the app calls upon the Clarifai API to analyze the content of the picture. The user is greeted with 20 keywords that match to the content of the picture, in addition to the probability of correctness.
How we built it
We spent a lot of time tinkering with the Android SDK and messing with dependancies. We had spent many hours trying out OpenCV with no luck, but the dependancies for Clarifai didn't give us any troubles. It was simple and easy to use. We installed the Java API Client and found the plethora of methods to use in our Android app.
Challenges we ran into
We had a huge problem setting up the back end of the app. We wanted to make this a journal app, which requires that data from the past be stored. HashSets and HashMaps are very useful, but implementing them required us to mess with many Listeners and OnClick objects. Ultimately, those data structures became the bane of our existence.
Accomplishments that we're proud of
We feel that not many things have been implemented with the Clarifai API, and we are proud to say that we were able to create a resourceful app that even we would use in our daily lives.
What we learned
During this weekend, every one of my team members can safely say that our development skills have been honed. We had to mess with many things we haven't seen before: half of us hadn't worked with Android Studio, and none of us had ever seen JSON. We were able to work together and persevere. Overall, it was a great weekend to not only create something, but learn.
Log in or sign up for Devpost to join the conversation.