👀 A proof-of-concept gaze detection Android app
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
app
clm
demo
dlib
gradle/wrapper
libs
openCVLibrary310
.gitignore
.gitmodules
LICENSE
README.md
build.gradle
gradle.properties
gradlew
gradlew.bat
settings.gradle

README.md

About this repository

This repository contains a proof of concept gaze detection Android app.

This work integrates code from the following open-source projects:

  • OpenCV the major open-source computer vision library.
  • Cambridge face tracker: a Constrained Local Model (CLM) framework developed by Tadas Baltrušaitis from the University of Cambridge.
  • dlib: a C++ library for machine learning.
  • dlib-android-app integrates dlib library in Android.
  • Boost: a collection of C++ libraries.
  • Weka: the main library for machine learning algorithms.

Here be dragons

This project is in a highly experimental state. We are not responsible for any damage that can happen to your device. Currently the app has been tested in the following devices:

  • Nexus 10
  • Samsung Galaxy Note 10.1

Download

Note: Currently the APK only works for devices with ARMv7-a processor.

gaze-detection.apk 8.4MB v1.0 md5sum: a5bef6fb6e54e56a1786b9f77688abce

Additional required files: Gazer.zip 21M
md5sum: a6122e017046bf927a148f8f3927a152

Install

Enable installation of apps from unknown sources in the settings of your device. Then open the APK file from your device to install it.

Extract Gazer.zip into your /sdcard folder.

From command line: $ adb install demo/gaze-detection.apk

Get the sources

$ git clone https://github.com/wearebase/gaze-detection-android-app

$ git submodule update --init --recursive

Development

The project is ready to work in Android studio 1.5.

Requirements:

  • Android SDK 23 (min API level is 18)
  • Android NDK 10e

License

See LICENSE file.

External libraries used in this project have their own licenses and copyrights.

Acknowledgement

This work has been possible thanks to the Knowledge Transfer Partnership (KTP) between We Are Base and Bournemouth University.