DoodlAR
Hyper personalized interior design through Augmented Reality
Every brilliant idea starts with pen and paper – it's simple and accessible. But what comes after? How do we bridge the gap between hand-drawn sketches and the digital realm? Enter DoodlAR, offering a seamless solution to transform your interior design concepts into digital formats, both 2D and 3D.
Client
Massachusetts Institute of Technology (MIT) Hacking Arts Hackathon - Wayfair Challenge
Responsibility:
Product Design, UX/UI Design, Visual Design
Tools used
Figma, Unity, IOS AR Kit
“There’s nothing like drawing a thing to make you really see it”
- Margaret Atwood
Context
When you're settling into a new space or looking to rearrange furniture, having a solid plan is essential. However, this process can be time-consuming, exhausting, and demanding. Quick sketches with pen and paper can be immensely helpful for planning and minimizing effort in such situations.

But the challenge arises when you need to translate those sketches into practical applications from mere pieces of paper. Your drawings may differ significantly from reality; they're two-dimensional and lack the context provided by the actual space's atmosphere. Understanding furniture arrangement requires context, viewed from various perspectives.
The MIT Hacking Arts is an annual conference that sparks cross-disciplinary innovation at MIT. Over the course of two days, artists, engineers, and entrepreneurs collaborate in a hackathon to tackle challenges at the intersection of the arts and technology.

This event is organized by the MIT Sloan School of Management Entertainment, Media & Sports Club in collaboration with MIT’s Center for Art, Science & Technology and the Martin Trust Center for MIT Entrepreneurship.
Image source: https://arts.mit.edu/start/entrepreneurship/hacking-arts/
Solution
Utilize augmented reality to seamlessly visualize authentic Wayfair furniture within your space. Effortlessly maneuver furniture and experiment with optimal arrangements.
Explore and add items to your cart within the app, with the added convenience of saving scenes for future reference.
Transform hand-drawn sketches into intricate 3D furniture models through advanced computer vision technology. Doodlar continuously enhances its capabilities through machine learning algorithms, ensuring smarter outcomes with each use.
Technology
Machine Learning; CNN image classifier with ml.js
To decipher human drawings, our team employed DoodleNet from ml.js in conjunction with a CNN image classifier for training purposes. Leveraging this machine learning technology, DoodlAR can accurately identify the user's sketches. Over time and with continued usage, DoodlAR will progressively improve its accuracy, becoming more adept at recognizing drawings.
Augmented Reality; Image target recognition with 3D models sets
Leveraging image target recognition to seamlessly integrate 3D models into the user's environment. Doodlar can identify specific images or markers in the real world and overlay corresponding 3D models. This capability enhances user experience by providing interactive and immersive content that appears to seamlessly blend with the physical surroundings.
User Journey & UX Construct
Journey
UX Construct
Screen Design & Demo
AR Experiences - Detect sketches, view 3D models
AR Experiences - Detect sketches, view 3D models
Product Demo
Results
We've gained invaluable insights from the brief yet intensive hackathon sprint. Here are key action items we're eager to pursue moving forward:
• Introduce VR mode and enrich its social capabilities
• Provide curated search results tailored to users' doodles
• Develop educational content focusing on interior design
• Expand into social features to broaden the platform's appeal
DoodlAR
Hyper personalized interior design through Augmented Reality
Every brilliant idea starts with pen and paper – it's simple and accessible. But what comes after? How do we bridge the gap between hand-drawn sketches and the digital realm? Enter DoodlAR, offering a seamless solution to transform your interior design concepts into digital formats, both 2D and 3D.
Client
Massachusetts Institute of Technology (MIT) Hacking Arts Hackathon - Wayfair Challenge
Responsibility:
Product Design, UX/UI Design, Visual Design
Tools used
Figma, Unity, IOS AR Kit
“There’s nothing like drawing a thing to make you really see it”
- Margaret Atwood
Context
When you're settling into a new space or looking to rearrange furniture, having a solid plan is essential. However, this process can be time-consuming, exhausting, and demanding. Quick sketches with pen and paper can be immensely helpful for planning and minimizing effort in such situations.

But the challenge arises when you need to translate those sketches into practical applications from mere pieces of paper. Your drawings may differ significantly from reality; they're two-dimensional and lack the context provided by the actual space's atmosphere. Understanding furniture arrangement requires context, viewed from various perspectives.
Image source: https://arts.mit.edu/start/entrepreneurship/hacking-arts/
The MIT Hacking Arts is an annual conference that sparks cross-disciplinary innovation at MIT. Over the course of two days, artists, engineers, and entrepreneurs collaborate in a hackathon to tackle challenges at the intersection of the arts and technology.

This event is organized by the MIT Sloan School of Management Entertainment, Media & Sports Club in collaboration with MIT’s Center for Art, Science & Technology and the Martin Trust Center for MIT Entrepreneurship.
Solution
Utilize augmented reality to seamlessly visualize authentic Wayfair furniture within your space. Effortlessly maneuver furniture and experiment with optimal arrangements.
Explore and add items to your cart within the app, with the added convenience of saving scenes for future reference.
Transform hand-drawn sketches into intricate 3D furniture models through advanced computer vision technology. Doodlar continuously enhances its capabilities through machine learning algorithms, ensuring smarter outcomes with each use.
Technology
Machine Learning; CNN image classifier with ml.js
To decipher human drawings, our team employed DoodleNet from ml.js in conjunction with a CNN image classifier for training purposes. Leveraging this machine learning technology, DoodlAR can accurately identify the user's sketches. Over time and with continued usage, DoodlAR will progressively improve its accuracy, becoming more adept at recognizing drawings.
Augmented Reality; Image target recognition with 3D models sets
Leveraging image target recognition to seamlessly integrate 3D models into the user's environment. Doodlar can identify specific images or markers in the real world and overlay corresponding 3D models. This capability enhances user experience by providing interactive and immersive content that appears to seamlessly blend with the physical surroundings.
User Journey & UX Construct
Journey
UX Construct
Screen Design & Demo
AR Experiences - Detect sketches, view 3D models
AR Experiences - Detect sketches, view 3D models
Product Demo
Results
We've gained invaluable insights from the brief yet intensive hackathon sprint. Here are key action items we're eager to pursue moving forward:
• Introduce VR mode and enrich its social capabilities
• Provide curated search results tailored to users' doodles
• Develop educational content focusing on interior design
• Expand into social features to broaden the platform's appeal