HuntMaster

HuntMaster is a modular framework made to create puzzle hunts at the intersection of reality and technology.

Comment

Inspiration

The creation of HuntMaster came about through an interest in puzzle/scavenger hunts. Our team desired to bring digital media into these kinds of experiences and to make it easy for anyone else to do so.

What it does

With HuntMaster, it's easy to create custom experiences known as hunts. You provide the story and tasks, HuntMaster does the heavy lifting.

Want to create a guided tour of your campus? Make a hunt that directs players to various parts of your campus and automatically proceeds to the next part of the tour when they reach them by using their GPS. All you need to do is write a few lines of python, the instructions/descriptions, and deploy using HuntMaster.

Want to play a cross between Bingo and a scavenger hunt where players are tasked with taking pictures of randomized items (e.g. basketballs, muffins) to cross them off their Bingo card? If you write the Bingo game logic, it's just a matter of interfacing with HuntMaster and then deploying.

Most setup for a hunt is done in python, although if you want your resulting page to look nice it's worth writing some HTML/CSS. Currently HuntMaster supports feature detection of images (e.g. determining whether a picture for the hypothetical Bingo hunt is valid), GPS location tracking, image comparison, QR code generation, as well as basic forms of input (text, buttons, etc.). Interfacing with these functions is often a matter of a single call.

How we built it

HuntMaster is built for the most part on python 2.7. It hosts a Flask server with a Bootstrap front-end. When a player in the hunt submits something, it's relayed to a python function which connects to whichever API is needed for that particular part of the hunt. If, for example, the player is tasked with finding an image of a dog, when she finds it and sends it to the server, the server then queries Google's Vision API to determine the contents of the picture. If it is of a dog, she'll receive the next part of the hunt.

Regarding modules,

  • Feature detection is done through Google's Vision API
  • GPS-based location tracking is done through HTML5
  • Image comparison is done through OpenCV (specifically feature tracking with SIFT)
  • QR Code generation is done through the qrcode python module

Challenges we ran into

Making HuntMaster required a lot of learning on our behalf. There was no single component which was overwhelmingly difficult, but having to juggle the creation of a centralized framework along with individual modules (each of which required figuring out a different API) kept us on our toes throughout the hackathon. And each API seemed to have its own difficulties (OpenCV was by far the most arduous). If we had to choose a biggest challenge, it would be web development; we were much more comfortable delving into the mathematics of GPS calculations or reading research papers on OpenCV than we were figuring out how to make our demo look nice.

Accomplishments that we're proud of

Though HuntMaster is a minimum viable product, a lot of its functionalities are more than minimally sufficient. We're very proud with the level of polish we've been able to give HuntMaster while still making sure it is rife with high-tech features.

What we learned

We've learned many things at BeachHacks. We learned how to architect a framework, design a web app, and many, many APIs.

What's next for HuntMaster

More modules, for sure. We are planning support for modules less related to traveling to a physical location and more to puzzles that can be done on the spot. Think things related to gyroscope/acceleration tracking, I-Spy using Google Vision's feature tracking (Google identifies parts of a picture and you have to click on them), and many more.

Share this project:

Updates