After seeing what Clarifai's deep learning API was able to do, we really wanted to use its features to create something. So we thought of creating a website where it could tell us if our food is healthy.
What it does
By giving an input picture of your food, the website would request Clarifai to analyze the image and it would return the list of ingredients.
How we built it
We used the Clarifai API with python to act as the back-end of the webpage. The back-end filtered all the useful information and kept the most accurate ones to send to the front-end. The front-end of the webpage receives a picture and calls the python code to identify the image using Clarifai's deep learning API.
Challenges we ran into
The main challenges we ran into was understanding the Clarifai API and retrieving useful information from the large amount of data received. Also, we tried to train the Clarifai API for a specific topic, however we ran into an expiring key issue therefore we had to give up the training of the API. We initially wanted to prototype the back-end in python and use Node.js as the final product. Unfortunately we encountered problems when we tried to install the Clarifai API and we chose to changed our tactics to keep python as the official back-end and switch to HTML, Ajax and bootstrap for the front-end.
Accomplishments that we're proud of
While encountering many obstacles, bugs, and errors, being able to solve these problems is be the best accomplishment we are proud of. Finishing it and being able to demo it in time is definitely something we are proud of as well.
What We learned
What's next for Foodifai
Use Train function to distinguish dishes instead of just the ingredients and find the restaurants around the user that offer those dishes by using the YellowPages API