The SF AppWorks Blog

Smart Vision - the 2019 SFAW Hackathon Results | SF AppWorks

Written by SF AppWorks | Jun 10, 2019 9:34:00 AM

A Weekly Snapshot of Life-Changing Technology

 

Happy Friday!

Every year the team at SF AppWorks participates in an internal hackathon. For those unfamiliar with what a hackathon is – it’s a competition where developers have to ‘hack’ an app or digital experience together in a short period of time. It’s a great exercise in innovation and creativity. Teams have access to the newest platforms and try to build proofs of concept utilizing the latest technologies.

 

Sometimes we go on the road and compete in public hackathons. We even won the TechCrunch Disrupt London Hackathon with our AI-powered Emotion Journal.

(Curious? Check out the short teaser video or watch the documentary we made.)

This week we are devoting the Wonder to the seven ideas that were brought to life over a 24-hour period.

Which is your favorite?

-Andrew

 

Now on to the Wonder, HACKATHON EDITION.

 

WINNER: Smart Vision

 

Smart Vision, the winner

of the 2019 SF AppWorks Hackathon, built a digital walking stick by using object detection and voice recognition.

Watch the video>>

 

How it works
Users can launch the app using the Google Voice Assistant. As the viewer picks up objects, it announces them. If the user gets too close to an object, the companion watch app vibrates.

 

Why it matters
Tech for the visually impaired is due for an upgrade. Walking sticks are useful, but with the help of computers and sensors, a blind person could experience their surroundings in a much richer way.

 

Coolest part
The app is actually a progressive web app, meaning it can launch from a browser and requires no installation. That means it could also be launched from a QR code positioned at the entrance of a building if, say, a company wanted to provide a custom map of their layout. 

 

RUNNER UP: Happy Feet

 

 

Happy Feet, the runner up of the 2019 SF AppWorks Hackathon, used home-made sensors and a Raspberry Pi to measure and analyze foot pressure.

Watch the video>>

 

How it works
You could line a custom shoe with the sensors or even create insertable soles. As a user walks, the sensors could provide realtime feedback on pressure, pronation, or cadence.

 

Why it matters
We put a lot of pressure on our feet. Understanding when improper form could cause damage would help people prevent injuries and stay mobile later into their lives.

 

Coolest part
The sensors, which use very little energy, could be designed to be self-powered by the foot pressure.

 

Life In A Box

 

 

Life In A Box uses blockchain technology to make the food distribution network more transparent and efficient.

Watch the video>>

 

How it works
Users can snap a photo of a QR code affixed to a perishable good and trace its entire journey.

 

Why it matters
When the last E. coli outbreak occurred in California last month, it took weeks for the CDC to track down the source. Better supply chain management can lead to quicker containment.

 

Coolest part
When blockchain is used in supply chain management, activity is stored in a shared ledger that’s updated and validated instantaneously with each network participant.

 

Hello Again

 

 

Hello Again created a facial recognition tool to help people with memory disorders recognize loved ones.

Watch the video>>

 

How it works
Family members or caretakers can create a portfolio of pictures and information for each person that a memory-impaired person knows. They only have to open the app and scan a visitor to see the portfolio.

 

Why it matters
We’re living longer, but unfortunately that means the number of people affected my memory impairment is growing. Technology can be a subtle and helpful way to remind people of things that slipped their memory.

 

Coolest part
The system processes in real time, meaning you could move the phone from person to person to quickly pull up contextually relevant data.

 

Squat God

 

 

Squat God analyzes your body position while working out and alerts you when your form is incorrect.

Watch the video>>

 

How it works
Object recognition tracks your movements and overlays a model on top of you. When your movements break from the model, the lines turn red.

 

Why it matters
Home workouts are becoming increasingly popular. Unlike in a gym, there are no trained professionals to supervise, which could lead to injury.

 

Coolest part
The software is highly customizable, meaning you could create models for every type of movement. You could even use it to train dancers.

 

SportMe Optical Recognition-Based Calorie Counter and Step Tracker

 

 

The SportMe Optical Recognition-Based Calorie Counter and Step Tracker is two hacks in one. The calorie counter uses image recognition to determine the calories in a picture of food. The step tracker uses optical recognition to measure movement and calculate steps.

Watch the video>>

 

How it works
Snap a photo of your food and the app will identify it and compare it to a database of food, then estimate the nutritional value. Pull out your camera and start working out. The app will measure your movements and provide an estimate of calories burned.

 

Why it matters
Over 30% of Americans are obese. Making it easier to determine the nutritional value of your food and providing better exercise tracking are two steps in solving the problem.

 

Coolest part
The two hacks work together to tell users how much they can eat after a workout, or how hard they should workout out after a big meal.

 

Codex Machine Learning to Predict When Users Will Quit

 

 

The Codex tool uses Microsoft Azure to analyze user data and predict at what point that user will quit.

Watch the video>>

 

How it works
Upload user activity data and analyze it to see when users are predicted to quit. Reengage them with notifications and offers.

 

Why it matters
Whatever kind of business you are managing, understanding why customers leave is an important part of providing better customer experiences. Machine learning can see patterns that would go unnoticed by people.

 

Coolest part
The model predicted with 95% accuracy when a current user of the SportMe app would quit.

 

JOKE HACK: The MetronoBOOM. Time, Controlled.

 

 

You gotta see it to believe it.
Watch the Video>>>

 

BONUS: The demo from the the winning team

 

 

Thanks for reading! We’ll see you next week.

-Andrew and Darius