We had an idea to make a Web Application to enter the types of alcohol you have and be presented with a list of cocktails/drinks you can make with the inputted alcohol. Coming into the Hackathon we noticed the weekly challenge of using clarifai's API, and came up with the idea to use image recognition to predict a cocktail.
What it does:
Step 1: User Takes/Uploads a picture from their mobile device and Submits the image Step 2: The image the goes to are server which uses are custom model we created on clarifai to predict the image Step 3: Then the server gets the most similar image from are SQL Database to get the information about the cocktail Step 4: The result is displayed on the Web App, Results include the name of the drink, an image, and a recipe.
Challenges we ran into:
Training the model, was the most time consuming task, since it only allowed 128 image to be bulk loaded at a time. Are solution was to using a python script to bulk upload in batches, in order to upload are 528 images into the model.
Accomplishments that we're proud of
Setting up a full stack Web Application using image recognition, with a 2 man team.
What we learned
Creating a Web Application, image recognition, and Google cloud service
What's next for Tell me the drink
AI to create new drinks/cocktails and add them to are database and model. As well as creating a model with a lot more images then 528.
Log in or sign up for Devpost to join the conversation.