Face and Barcode Recognition on Android
- Android device with at least an Amlogic S912 CPU (or equivalent), 3 GB ram, and Google Play Services installed on the device.
- A USB camera with autofocus (important for Barcode recognition) and with at least HD (1280×720) resolution
- Make sure that the Google Play Services application is installed on the device. It can be found in the device’s settings. (Settings – Apps – Usually on the top left corner in the 3 dot menu click show system – Find Google Play Services in the list.)
- Install the update for your signage app. (The lowest version which has this feature is ‘2.20.03.0’)
- Plug the camera into your device.
- Create a sensor configuration for recognition. (You can find more information about the Face recognition’s sensor configuration below.)
- Apply the sensor configuration on your device.
- Sometimes the camera won’t start on the first try, it can happen because the Google Play Services application has to download some libraries needed for face recognition. Please be patient as it can take some time.
- Create a content that can react to the events coming from the recognition. (You can find more information about the content setup below.)
1. Open the Wallboard Sensor designer app. (You can find more information about the Sensor Designer here.)
2. Select Android USB as a connected device.
3. Select Port USB, and then USB Camera.
4. Now you can set the settings for the camera:
ID: This is the ID of the sensor. You can set it to anything you want.
Camera Resolution: This is the resolution of your camera. (Do not set to a too high resolution because it can heavily affect performance.)
Camera FPS: This determines how many camera frames will be sent to the processing algorithm each second.
Report URL: Here a custom URL can be set. The device will send the statistics to this custom URL if needed. (You have to only fill this field if you have a custom site that you want to send the statistics to.)
Show Preview: If you check this then you can see a preview of the camera on the device.
Tracker Mode: Here you can set the mode for the recognition. It can be set to Faces, Barcode, or both. If you set it to Faces it will only recognize faces, if you set it to Barcode then it will only recognize barcodes, or if you set it to both then it will look for both of them, but it will affect performance heavily!
Highlight Faces/Barcode: If you check this then the device will draw a circle or a rectangle around the detected objects.
Delay of Send To Content: This will determine the delay between two messages sent to the content.
5. Next, you can set the presence settings, this will determine the events depending on how many faces are detected.
Range: These are the ranges. Here you can set the intervals for the presence events.
Message: Here you can set the events that you want to receive from the device in your content.
6. Next, you can set the events for the Face Analysis. Here you have to enable the events one-by-one. Only enable the ones that you will use!
Gender analysis: Here you can set the messages for gender events.
When there are more males: This event will be sent to the content if mostly males are in front of the camera.
When there are more females: This event will be sent to the content if mostly females are in front of the camera.
When they are equal: This event will be sent to the content if the number of males and females is equal.
Age analysis: In this part, you can set the messages for the age events.
When the most are children: This event will be sent to the content if mostly children are in front of the camera. (0-15 years)
When the most are juniors: This event will be sent to the content if mostly juniors are in front of the camera. (16-25 years)
When the most are young: This event will be sent to the content if mostly young people are in front of the camera. (26-40 years)
When the most are matures: This event will be sent to the content if mostly mature people are in front of the camera. (41-64 years)
When the most are seniors: This event will be sent to the content if mostly senior people are in front of the camera. (64+ years)
7. After you set these values and you are connected to your device, then you can hit the “Send configuration to screen over the cloud” button on the bottom left corner. If everything is set correctly then the recognition will start to work.
In this example, a content will be set up which will jump to another page if 1 person is in front of the camera. The process is the same for every event.
1. Create new content for recognition.
2. Create a second page. This is where the content will navigate if a person is standing in front of the camera. It is important that the page you create for this should not be in the Main Playlist, add it as an Other Page instead.
3. Select Content Settings on the right side, and scroll down to the External events part, and select Edit Sensor events.
4. Here at the React to sensor events click Add action.
5. Now set your sensor event (in this case some-one-here), set the action (Jump to), and set the target (Second Page).
6. Now set up an event that will go back to the first (Default) page, so after the person leaves the content will reset.
7. Click save at the bottom, and now the content will react to the events coming from the Face recognition.
8. Save your content, and assign it to your device.