How many times have you had this conversation?...
"Do you want to go out for dinner?" "Sounds great, what do you FEEL like eating?...what are you in the MOOD for?..."
What it does
- access a set of images containing food & people by URL and tag them by mood -- using Clarifai
- the user can then filter the URLs by specific mood (eg, happiness)
- this subset of images are then tagged with food types (eg, pizza), helping the user identify "Happy Meals"
How we built it
Incorporated a user-defined probability/ confidence threshhold to restrict to the tags that Clarifai had greater confidence in its Neural Net-based classification.
Challenges we ran into
Accomplishments that we're proud of
Implementing Clarifai to tag images with moods and food types.
What we learned
Matt & Keeyon at Clarifai are super helpful!
What's next for Eat What You Feel
- Scale up
- Inform marketing efforts by - adding search for brand hashtags (eg, #Subway #McDonalds ), locating twitter images with hashtags, and classifying for food & mood
- Competitor Analysis: monitor frequencies of food type images that are being posted by guests about competitors