I attended the On Deck Cup hackathon in NYC a couple of weeks ago and saw some of the most creative sports hacks. Two crowd favorites: a video app that would sync all the crowd sourced videos from a particular game, so any play can be seen from multiple different angles; an MMA betting application that would give you odds of winning through unusual attributes like who was smiling more during the weigh in. However, my personal favorite has to go to Jason Massih for his app Sports Recog.
Jason wanted to create an educational tool that would allow other people like himself, who aren’t the biggest golf fans to learn more about the players and game. So he decided to make a mashup of SendGrid, SeatGeek, ParkWhiz and Rekognition, a facial recognition API. He trained the recognition algorithms to understand most photos of Tiger Woods, Phil Mickelson and Henrik Stenson. He did this by using different angles of their faces for the training dataset.
He then needed a way to send this photo on his phone to the facial recognotion he made and get data back. So he decided to take advantage of SendGrid’s Inbound Parse Webhook to send the photo from his phone to server. His server would then recognize that the person is Tiger Woods and get upcoming games/stats and available parking that can be booked now and emails that info back to you.
Currently the SportsRecog hack only recognizes photos of Tiger Woods, Phil Mickelson and Henrik Stenson. Jason does have plans to train it for other players as well. However, in the meantime feel free to try it yourself:
- Obtain a photo of Mr. Woods, Mickelson or Stenson
- Email that photo to firstname.lastname@example.org
- Make sure to attach the image of the golfer!
Photo by Flickr user familymwr