Announcement

Collapse
No announcement yet.

Any interest for a face recognition plugin?

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    Any interest for a face recognition plugin?

    I'm working on a face recognition plugin for HS3, and would like to know if there is a lot of interest for such a thing?

    The plugin would work with most of webcams (and probably netcams too). It would take some pictures of the person you want to be recognized and use them to train the recognizer algorithm.
    Then it would create one device for each person and set the value of this device to "Present" as soon as one person is recognized.


    Click image for larger version

Name:	devices.jpg
Views:	1
Size:	42.2 KB
ID:	1206907

    Click image for larger version

Name:	faces.jpg
Views:	1
Size:	70.0 KB
ID:	1206908
    46
    Yes
    86.96%
    40
    No
    13.04%
    6
    Last edited by spud; September 13, 2014, 06:05 AM.

    #2
    Years ago, I saw a video of a guy that built a pet detector and would only unlock the pet door if the profile of the cat was detected. The idea was that a skunk or possum's profile would be different from his cat. It was very interesting.

    Would you be using an IP camera's snapshot feature to capture the face?
    HS4Pro on a Raspberry Pi4
    54 Z-Wave Nodes / 21 Zigbee Devices / 108 Events / 767 Devices
    Plugins: Z-Wave / Zigbee Plus / EasyTrigger / AK Weather / OMNI

    HSTouch Clients: 1 Android

    Comment


      #3
      It would interest me if I could integrate it into a future install but I would need to think how I could get a camera into a position for as soon as I walked in the door to recognise me and then trigger events. I have used similar applications at work (I think it was something to do with the distance between the eyes our application used to recognise faces) but accuracy was always a bit suspect.

      Comment


        #4
        Years ago there was talk about a facial recognition plug-in but the developer never came through and just disappeared.

        Always wanted to try something like this. Looking forward to it.
        Bryan
        Software/Hardware: Win10 Pro, HS 3 Pro, HS Touch, Echo, Edgeport/4, Z-Net w/88 Devices, Insteon PLM w/19 Devices, Nest, GC-100-6, W800RF32A, WS-2080 Weather Station (KMADRACU10)
        Plug-in/Scripts: Alexa API, BLBackup, BLGData, BLLED, BLLock, BLRF, BLTVGuide, Blue Iris, BLUPS, Current Cost 3P, DirectTV, FitbitSeer, Insteon, Nest, Pushover 3P, Random, Restart, Tasker, UltraGCIR3, UltraWeatherWU, Z-Wave

        Comment


          #5
          will it work with Kinect?

          Comment


            #6
            Originally posted by kideon View Post
            will it work with Kinect?
            yes I will make sure it works with the kinect camera

            Comment


              #7
              Kool... I already have camera's in place
              but.. can it use my feed ???/
              Ubuntu on the Intel NUC ( 8i5BEK ), 32 G, 250G SSD, V4 Pro 😎

              Comment


                #8
                Originally posted by gearyt View Post
                Kool... I already have camera's in place
                but.. can it use my feed ???/
                right now I'm testing it on a USB webcam plug into my HS machine, but ultimately I would like to make it work with any camera feed.

                Comment


                  #9
                  I would appreciate this immensely! A combination of this and the elusive UltraSightHound plugin would revolutionize my home automation.
                  Originally posted by rprade
                  There is no rhyme or reason to the anarchy a defective Z-Wave device can cause

                  Comment


                    #10
                    Originally posted by spud View Post
                    right now I'm testing it on a USB webcam plug into my HS machine, but ultimately I would like to make it work with any camera feed.
                    Let me know when you are ready, I will get you in like last time
                    Ubuntu on the Intel NUC ( 8i5BEK ), 32 G, 250G SSD, V4 Pro 😎

                    Comment


                      #11
                      UltraSightHound plugin?

                      Comment


                        #12
                        Originally posted by kideon View Post
                        UltraSightHound plugin?
                        http://board.homeseer.com/showthread.php?t=169683

                        If you find it interesting please comment and maybe he'll be more motivated to work on it!
                        Originally posted by rprade
                        There is no rhyme or reason to the anarchy a defective Z-Wave device can cause

                        Comment


                          #13
                          Definitely interested, there is an immense value if this plugin is successful.

                          Comment


                            #14
                            +1


                            Tom
                            Tom
                            baby steps...starting again with HS3
                            HS3Pro: Z-NET & 80 Z wave Devices,
                            HSTouch: 4 Joggler (Android Kitkat), 2 iPhone, 3 iPads
                            Whole House Audio: 5 SqueezePlay Jogglers w Bose Speakers
                            In The Works: 10 Cameras Geovision, new Adecmo/Envisalink Alarm, Arduinos
                            System: XP on Fanless Mini-ITX w/ SSD

                            Comment


                              #15
                              I think such a plug-in would be very interesting, as long as it has good performance.

                              For purposes related to FR, I presently use facial recognition hardware. It works very well but :

                              A- it is more expensive (200-300) per hardware unit.
                              B- it is not as flexible as a plug-in could be.
                              C- It is basically ON-OFF because, in my current setup, it doesn't speak to my HS Server to tell it which user it is recognizing.

                              On the upside, the hardware deals surprisingly well with different lighting conditions and partial facial changes such as beard vs no beard, glasses vs no glasses, hat vs no hat, etc. The hardware is also difficult to fool and has virtually no false positives...

                              Comment

                              Working...
                              X