IPCam

Camera integration into your home automation

9 Topics 122 Posts
  • This topic is deleted!

    1
    0 Votes
    1 Posts
    25 Views
    No one has replied
  • Video Doorbells

    38
    1 Votes
    38 Posts
    2k Views

    BTW I found out that the Google also a new API for their cameras. Has anyone tried to integrate the Nest Doorbell (or another Nest cam) in openluup?

  • Object Recognition

    20
    2 Votes
    20 Posts
    1k Views

    Yeah a bit far 👅 Though I am sure I would enjoy it.

  • Facial recognition triggering automation

    20
    2 Votes
    20 Posts
    988 Views

    I have taken another look at this and have now drastically modified my fork of homeassistant and deviated from the main branch:

    I added a state variable to the image processing component so as to turn on and off the processing from openLuup. (and giving the ability to turn on only upon camera movement detection.) Massively optimized the video frame handling so that it doesn't get converted into a variety of formats. It's still not perfect since it still goes from the GPU decoder to the CPU, runs some processing (resizing mostly) and gets sent back to the GPU for model inference but I will get there eventually. Huge improvement in the face detection model where I refactored the model to take out some constants, preventing them from getting recalculated for every frame and more than doubled the inference speed. Updated my object detection model to this improved version of Yolo which is more accurate and only cost a tiny bit more:
    GitHub - WongKinYiu/ScaledYOLOv4: Scaled-YOLOv4: Scaling Cross Stage Partial Network GitHub - WongKinYiu/ScaledYOLOv4: Scaled-YOLOv4: Scaling Cross Stage Partial Network

    Scaled-YOLOv4: Scaling Cross Stage Partial Network - GitHub - WongKinYiu/ScaledYOLOv4: Scaled-YOLOv4: Scaling Cross Stage Partial Network

    Note: I am heavily relying on pytorch as my neural network framework (supported by facebook but initiated as a lua project before it moved to python) and opencv for video and image processing (supported by intel).

  • CCTV on Openluup

    5
    0 Votes
    5 Posts
    263 Views

    Cool. Thanks, gents

    C

  • 2 Votes
    7 Posts
    1k Views

    I have been optimizing the code further to have less dependencies (relying directly on the dlib library instead of a wrapper for it) and learned some more about convoluted neural network inferences... fascinating field. I also changed the jitter parameter from 1 to 10 which should help the accuracy.

    https://github.com/rafale77/core/blob/live/homeassistant/components/dlib_face_identify/image_processing.py

    The original dlib model has scored 99.17% at the LFW benchmark and my mods sped up the face detection and probably improved accuracy by using a classifier rather than just an Euclidean distance, enlarging the face detection and training set, and the increase of the jitter parameter. Looking around for what is available, the only thing that is potentially better is the retinaface detection + arcface encoding which potentially could improve the angled recognition. I am almost done implementing it for testing but it probably is overkill for my doorbell... 🙂

  • opensource NVR

    23
  • License Plate recognition

    7
    0 Votes
    7 Posts
    902 Views

    Hmm not quite. I don't use supervisor and am running the core directly. I actually don't need any special support at all as I am running the most basic from source, similar to the python virtual environment except that I install mine without it. This means that for my setup to not be supported, they would have to stop support home assistant source...

  • Monocle on Alexa

    1
    1 Votes
    1 Posts
    178 Views
    No one has replied

Recent Topics