The largest challenge-driven hackathon in the U.S. commenced over the weekend of February 23rd - 24th in San Francisco. Beginning on Saturday morning, it continued for a mere twenty-six hours before concluding with rapid-fire project pitches and awards Sunday afternoon. Hundreds of developers, designers, and entrepreneurs converged at Galvanize SF to form teams and work to solve technical challenges for prizes and glory!
Hackathon teams were invited to work on any type of project (web apps, mobile apps, wearable solutions, data visualizations, algorithms, even video presentations, or visual mock-ups) but all of the hackathon prizes were sponsored prizes for teams that best utilize a sponsor technology. Sponsoring companies provided access to their APIs and hardware and incentivized hacker teams with branded prizes. Plenty of food, snacks, and caffeine fueled the hackers throughout the event.
Clarifai was an official sponsor and we gave open access to our easy-to-use computer vision APIs. Sponsoring a hackathon gives us a great opportunity to engage with our users directly to get their feedback in real-time. Over 30 teams used Clarifai in their final applications!
Dianna (Backend Engineer @ Clarifai) and I had fun staffing the Clarifai table during the event, helping attendees use Clarifai and troubleshooting their technical challenges.
DeveloperWeek SF Hackathon -- The Rules
- Teams built apps from scratch (Mobile App/Consumer Web App/Business Web App). Teams were allowed to solve no challenges or submit to as many Accelerate.im challenges as they want.
- Judges heard the pitches from all teams building on their technology and filled out a scoring sheet on each team.
- Judging was based on 1) How much progress the team made 2) Concept – does it solve a real problem?, and 3) Feasibility – could this become a startup or company?
- The top 5 selected project teams pitched on the main stage.
Over 150 teams convened at the event and dozens of hack projects were submitted to sponsor challenges.
Clarifai was used in three of the top five finalist projects selected by an independent panel of judges! Clarisign (2nd overall winner), FairLegal, and You Snooze, You Lose.
In addition, other hackathon sponsors awarded their prizes to projects that incorporated Clarifai. Below is a list of projects that used Clarifai that won awards from other sponsors:
- Photobot (Wowza Media Systems winner) (DocuSign winner)
- Revision (Wowza Media Systems winner )
- ROADLIFE (Wowza Media Systems winner)
- WeSign (Agora.io winner)
- Canonify (Canon winner)
- ClariSign (2nd Place Overall winner)
Here is a sample of the 30+ projects that used Clarifai computer vision in their unique applications:
"ROADLIFE analyzes video data and alerts relevant parties & authorities about local road hazards & incidents in real time. During the hackathon, we used Clarifai to analyze video data from traffic cameras to automatically alert authorities and improve their response time."
"RoboSitter is a content mediator chrome extension and web app that allows parents to whitelist internet content to help limit exposure to unsafe content. DocuSign: handles legal liabilities for capturing user images using Embedded Signing, Composite Template. Agora.io: streaming service to capture video for facial authentication. Clarifai: Train "Admin" model to recognize authorized parent and identify who has authority to edit whitelist content. Yubico: U2F authentication with YubiKey."
"Your second eyes anywhere, a photo streaming app. Connect any two points on earth in an instant and on-demand with reVision."
"Leveraging Clarifai to tag inventory, geofencing, and RTC to build the ultimate preventative solution to lost & found problems."
"Mobile app for identifying clothes from images with Clarifai."
You Snooze, You Lose
"You Snooze, You Lose is a mobile alarm app that helps you to wake up in the morning by making it literally costly to hit snooze. The alarm will continue ringing until you are out of bed to take a picture of a specified household item (checked using Clarifai). If snooze is chosen, you will be directed to sign a document legally binding you to pay a friend or donate to an organization before the alarm is disabled (via DocuSign)."
"Influencer AI allows anyone to easily become an influencer. Additionally, brands can easily identify and develop an influencer network. Influencer AI automatically detects an existing influencer relationship to a brand's items or logo using Clarifai, and identifies the need for legal documentation to be drafted between influencer and brand. The tokenization of the image on the Near protocol allows control of the licensing of the IP."
Diseased Crop Identifier
"Crop Scanner is an app that makes it easy to identify diseased crops from healthy crops and helps guide a farmer about the prevention methods and cure for the given disease."
"A decentralized application that facilitates interactions between clients and attorneys."
"Combining Canon API and Clarifai to identify good vs spoiled fruit at scale."
"RentrLoc adds multi-factor security to leases and short-term rentals. Renters upload an image of themselves, sign and pay the agreement with DocuSign and authenticate themselves with their Yubikey, which becomes their unique identifier, and key to the apartment. The user also uploads one or more pictures of themselves. When the renters show up to access the apartment, they verify their identity using their Yubikey and their face using Clarifai."
"PhotoBot is an automated photographer for events. It consists of a Canon camera and Raspberry Pi mounted on a motorized robot, and an App for entering events. The Pi controls the camera, converts media, broadcasts on Wowza, moves the robot, and provides an API. Photos are processed with Clarifai to identify people. The App scans for local geofence regions using TomTom. You can enter an associated event, sign a model release with DocuSign, watch a Wowza feed, snap photos, and view your pics."
Has seeing these projects got you thinking about your own use case for computer vision? Clarifai has a free tier! It's easy to get started. The projects you see above were built in just two days.