AR Face Detection

A face detection app that tracks the face and displays any picture above the user’s head using both front and rear cameras.

1 week
1 member
Augmented Reality
Open Source

AR Face Detection

AR Face Detection was created by our Android developers to contribute to the world of open-source AR projects.

An image

Project idea

The main idea was to share our expertise with the community of open-source enthusiasts with this AR library.

An image

Target audience

The audience of AR face detection is mostly AR enthusiasts and ordinary people who are curious about new technologies and use cases in this field.

Project challenges
Implementation of different toolkits to the app
Use of both frontal and rear cameras
Complicated mathematical calculations of the head slope degrees
Our Android team built an open-source Android app that precisely tracks a user's movements and displays an image above the user’s head.
How does face detection work?An image
Face detection recognizes faces using four patterns:
  • Knowledge-based. Describes the face based on rules.
  • Feature invariant. Detects faces based on face features (nose, eyes)
  • Template-matching. Compares input with previously-stored face samples.
  • Appearance-based. Uses machine learning and analysis to detect faces.
We did a thorough analysis:
  • Removing background. Algorithms remove background to simplify the search process.
  • Skin color. In color images, skin color and tone help searching for faces.
  • Motion analysis. In videos, algorithms compare moving faces with a static background.
  • Pose analysis. Analyzing sets of poses, algorithms can detect faces better.

Main features of the project

With the help of ARFace Detection Library, our developers managed to develop the following features.

Displaying Image
The main part of the app puts any image above the user’s head. With the help of face tracking, the image rotates together with head movements. The app processes several faces simultaneously.
An image
Toolkit Selection
AR Face Detection offers users two augmented reality tools. Both of them display the image in unique ways. ML Kit can use both rear and front camera, and ARCore uses the front camera only.
An imageML KitAn imageARCore
An image
Image Selection
Our application allows users to display any image of their choice. Developers can manipulate the open-source code of ARFace Detection Library and insert any picture they want.
An image

Tech stack

The following tools were used to develop Face Detection App:

An image
An image

Visit our GitHub account to get the source code of this project and get an inside look at our application.

An image

Check this article for more information about the ARFace Detection Library and technical peculiarities.

Our Cases

An image


Video-On-Demand Service

Explore Case
An image

Social Media for Dancers

Social Media for Dancers

Explore Case
Back to top