About
We develop JavaScript libraries to bring real-time computer vision to web developers. They are powered by the fastest deep learning engine running in the web browser. It is so fast that it can analyze a video stream in real-time even on mid range mobile devices.
We don't rely on any mainstream deep learning framework or computer vision library like OpenCV or Tensorflow. Our software is home-made, lightweight, standalone, efficient and our license fees are affordable.
You can browse our libraries in our Github page.
WebAR.rocks.hand
This is our hand detection and tracking computer vision solution.
- It is fully web based and mobile friendly,
- It is featured with these demos:
- Hand based navigation (as on this website),
- 3D Object manipulation in the palm of your hand,
- Shoes virtual try-on (both barefoot and shoes on),
- Wristwatch and ring virtual try-on
- Virtual try-on demos are provided using THREE.js as rendering engine,
- Integration demos are provided both in static JavaScript and in a React / Three fiber / NPM / ES6 environment.
Source code and demos are hosted on Github here:
github.com/WebAR-rocks/WebAR.rocks.hand.
WebAR.rocks.image
This is our image detection and tracking solution.
- It is fully web based and mobile friendly,
- THREE.js is used as 3D rendering engine,
- Can track an unlimited number of target images,
- Specific neural network needs to be trained in some cases,
- It can be used to track generic plane objects like QRCodes, screens with web content, identity documents, ...
- Integration demos are provided both in static JavaScript and in a React / Three fiber / NPM / ES6 environment.
The source code and demos are hosted in a private Github repository. To access it, please contact us at
[email protected].
WebAR.rocks.face
This is our face detection and tracking solution.
- It is fully web based and mobile friendly,
- THREE.js is used as 3D rendering engine,
- It detects specific facial landmarks,
- It is featured with these demos:
- Flexible mask face filter,
- Accessories VTO: earrings, hats, glasses, motorcycle helmet/headphones,
- Makeup VTO: sport makeup, lipstick, full makeup,
- Face replacement.
- Integration demos are provided both in static JavaScript and in a React / Three fiber / NPM / ES6 environment,
- We provide support and upgrades through a yearly licencing plan.
Source code and demos are hosted on Github here:
github.com/WebAR-rocks/WebAR.rocks.face.
WebAR.rocks.object
This is our generalistic object detection and tracking solution.
- It is fully web based and mobile friendly,
- THREE.js is used as 3D rendering engine,
- It detects object orientation to overlay 3D content with 6DoF,
- It can also be used with SLAM engines (8th Wall, WebXR) to handle the tracking once object is detected,
- Neural network training is done from a 3D model of the object,
- We can train a neural network to detect a general concept (like a car or a cup) or a specific 3D object (like a specific can model),
- Integration demos are provided both in static JavaScript and in a React / Three fiber / NPM / ES6 environment,
Source code and demos are hosted on Github here:
github.com/WebAR-rocks/WebAR.rocks.object.