General Program Flow:
The cameras connected to the PC will perpetually scan people walking by the area that the camera is located.
In near Real-Time, the software created will uses Face API or Amazon Face API to determine the results to display on the TV based on the facial characteristics of the person in the area.
The software can detect numerous attributes, including gender, ethnicity, eye color, hair color, facial expressions etc. **The most important attribute for the first application of the FACE experience is Males with a Beard or No-Beard
The APIs does not tell you in Yes or no; rather it gives you a confidence level out of 10. Like a heavy beard will be marked as 9 and lighter beard will result as 4. So you have the control in your programming logic to decide of what is your minimum level which you consider a beard.
The software will process the results to send appropriate video as a UDP command to the defined port. The video playback software will be Brightsign. The software will process the person’s unique attributes and trigger the appropriate video file in near real-time (1-2 seconds maximum per person)...
The API can read the video stream and can detect the human/faces etc. We need the ability to adjust the interval, frames to analyze from the stream, on a per campaign basis.
The average person will not be standing in front of the camera in a stopped position. They will be moving in front for a predefined area...let's assume 10’ x 10’ activation area in front of the camera
Save the results, time, video name, to the system log for statistics purpose
The admin can create the conditions in the control panel to associate various content types to the videos. In this first phase, the conditions can be:
Facial hairs **top Priority**
Gender (eliminate triggers for females in the first application)..
setting for the Brightsign's url/IP and port for UDP communication
A C#/.NET application to be installed on the connected PC which will do the most of running tasks. It will keep on processing the live video input feed (people walking by in front of the camera that’s setup….to trigger new content as frequently as every few seconds….to determine the media to played.
This application will keep track of the video currently playing and will play the next video after completion of currently running video based on the most recently captured photo / or override the existing video every time a new person is identified....
This app will talk to the FACESDK and determine the next video. This app will also be sending the log to the stats database on the cloud. There will be a list of videos to be played in the form of text or XML format.
A PHP/Mysql based application to capture the stats and for reporting purposes.
January 15, 2018
I am looking for a mix of experience and value