This app replaces the earlier version which has now been removed from play.google.
Larger screens (eg tablets) may be slow due to the amount of data being processed.
Read more about the app at http://www.barrythomas.co.uk/machinevision.html - Source code of the key sections of the frame processing loop are on the web site.
The app was written mainly as a learning exercise and is provided with no guarantee of road worthiness or fitness for purpose. Let me know if you want a copy of the source code.
- Set minimum face size threshold with the scroll bar on side
- Press menu button and select "Change Camera" to change camera
- Switch between native and Java face detection algorithms
The application works in two phases:
1. training: the user has to train the software by taking pictures of people (at least 20 photos of each person)
2. recognition: the user can point the device to any (known) person and get a positive response about that person's identity
Remember this is a simple OpenCV-Android face recognition project mainly for demonstration of the power of OpenCV.
You can either ring the buzzer sound when it detects, shoot the photos.
The number of detections so far (Detects) is displayed on the screen.
I can adjust the sensitivity slider bar (SensorLevel).
I can specify the presence or absence of the buzzer sound sound button (Sound).
I can specify the presence or absence of photography photo button (Photo).
It is stored in the photo folder of default photos taken.
Adjust the volume button on the device body volume of the buzzer sound.
Installation of OpenCV Manager is required to run.
If OpenCV Manager is not installed on this application run time
Please install because it jump to OpenCV Manager to Google Play of automatically.