ML based Scavenger Hunt:

Scanning Physical elements

Inspiration

𝐔𝐧𝐥𝐢𝐤𝐞 𝐭𝐫𝐚𝐝𝐢𝐭𝐢𝐨𝐧𝐚𝐥 𝐢𝐦𝐚𝐠𝐞 𝐭𝐚𝐫𝐠𝐞𝐭 𝐛𝐚𝐬𝐞𝐝 𝐬𝐜𝐚𝐧𝐧𝐢𝐧𝐠, This experience 𝐥𝐞𝐯𝐞𝐫𝐚𝐠𝐞𝐬 𝐌𝐋 𝐛𝐚𝐬𝐞𝐝 𝐒𝐧𝐚𝐩’𝐬 𝐬𝐜𝐚𝐧 𝐭𝐞𝐜𝐡𝐧𝐨𝐥𝐨𝐠𝐲. So you are not tied to a singular image target. You can use any physical object as a part of game.

What it does

𝐓𝐮𝐫𝐧 𝐚𝐧𝐲 𝐩𝐡𝐲𝐬𝐢𝐜𝐚𝐥 𝐞𝐧𝐯𝐢𝐫𝐨𝐧𝐦𝐞𝐧𝐭 𝐢𝐧𝐭𝐨 𝐚 𝐯𝐢𝐫𝐭𝐮𝐚𝐥 𝐠𝐚𝐦𝐞 𝐚𝐫𝐞𝐧𝐚 𝐚𝐧𝐝 𝐩𝐡𝐲𝐬𝐢𝐜𝐚𝐥 𝐨𝐛𝐣𝐞𝐜𝐭𝐬 𝐚𝐬 𝐠𝐚𝐦𝐞 𝐩𝐫𝐨𝐩𝐬.

𝐔𝐬𝐢𝐧𝐠 𝐬𝐜𝐚𝐧 𝐭𝐞𝐜𝐡𝐧𝐨𝐥𝐨𝐠𝐲 we built this scavenger hunt game. Users will be given clues on items to decode riddles and have to scan appropriate items to unlock effects and move to the next level.

How we built it

This experience 𝐥𝐞𝐯𝐞𝐫𝐚𝐠𝐞𝐬 𝐌𝐋 𝐛𝐚𝐬𝐞𝐝 𝐒𝐧𝐚𝐩’𝐬 𝐬𝐜𝐚𝐧 𝐭𝐞𝐜𝐡𝐧𝐨𝐥𝐨𝐠𝐲. So you are not tied to a singular image target. You can use any physical object as a part of game.

Challenges we ran into

Understanding of scanning mechanism and training ML to leverage it to make them as a scavenger hunt results. VFX activation.

Accomplishments that we're proud of

When the user solves the puzzle (for instance he found out “dog” is the answer for the puzzle, he can scan a dog at his house (pet) or even he can scan some dog toy or poster on the web).

What we learned

Concept of Machine Learning

What's next for ML based Scavenger Hunt:

Dynamic cloud based game and persistent score cards are planned to see who got higher and 𝐩𝐥𝐚𝐧𝐧𝐞𝐝 𝐭𝐨 𝐦𝐚𝐤𝐞 𝐢𝐭 𝐚𝐬 𝐚 𝐜𝐨𝐧𝐧𝐞𝐜𝐭𝐞𝐝 𝐥𝐞𝐧𝐬 𝐚𝐧𝐝 𝐚𝐥𝐬𝐨 𝐚𝐬 𝐚 𝐥𝐞𝐧𝐬 𝐟𝐨𝐫 𝐬𝐩𝐞𝐜𝐭𝐚𝐜𝐥𝐞𝐬