Facial recognition & room presence using Double Take & Frigate

Hi All,

I have a question about the files listed below, I noticed that backups were quite a bit larger and found that a couple of files located in /image/blobs/sha256 were quite large. They may have been created when I was setting up the system however my current configuration appears not to be creating any more. My question is whether these files can be safely removed?

Thanks,
Paul

Hi!

I have this set up on a camera in my hallway, pointed at the door. The use-case is to recognize people as they arrive so that my smart home knows who’s visiting.

The challenge is to get a decent snapshot to run the recognition on. Frigate sends Doubletake one or two snapshots from the video clip for each event (person arriving). In these snapshots odds are that the person is looking away, and not at the camera. One way to mitigate this would be to take multiple snapshots on demand in stead of relying on only the snapshots from the video. Is there any way of accomplishing this?

I know I can take snapshots with the camera via the “camera.snapshot” service, but how can this snapshot be used by Doubletake and forwarded to Compreface?

Could someone help me setting up frigate, deepstack and doubletake?? I have an esp32 cam that is working well with frigate but in doubletake I can’t see anything in my matches.

Here is my dockercompose.yaml:

  deepstack:
    container_name: deepstack
    restart: unless-stopped
    image: deepquestai/deepstack
    ports:
      - "4000:5000"
    environment:
      - VISION-DETECTION=True
      - VISION-FACE=True
      - API-KEY=********
    volumes:
      - ./deepstack/pytorch-models:/modelstore/detection

  frigate:
    container_name: frigate
    privileged: true
    restart: unless-stopped
    image: ghcr.io/blakeblackshear/frigate:stable
    shm_size: "64mb"
    volumes:
      - /etc/localtime:/etc/localtime:ro
      - /home/matteo/dockercompose/frigate/config/config.yaml:/config/config.yaml
      - /home/matteo/dockercompose/frigate/storage:/media/frigate
      - type: tmpfs # 1GB of memory
        target: /tmp/cache
        tmpfs:
          size: 1000000000
    ports:
      - "5000:5000" # Port used by the Web UI

  double-take:
    container_name: double-take
    image: jakowenko/double-take
    restart: unless-stopped
    volumes:
      - ./double-take:/.storage
    ports:
      - 3000:3000

This is my frigate config:

mqtt:
  enabled: false

cameras:
  esp32_camera:
    ffmpeg:
      inputs:
        - path: *mypath*
          roles:
            - detect
            - record
      input_args: -avoid_negative_ts make_zero -fflags nobuffer -flags low_delay -strict experimental -fflags +genpts+discardcorrupt -use_wallclock_as_timestamps 1 -c:v mjpeg
      output_args:
        record: -f segment -segment_time 10 -segment_format mp4 -reset_timestamps 1 -strftime 1 -c:v libx264 -an
        rtmp: -c:v libx264 -f flv
    rtmp:
      enabled: false 
    objects:
      track:
        - person
    detect:
      width: 800
      height: 600
      fps: 3
    record:
      enabled: True
      retain:
        days: 3
        mode: motion
      events:
        retain:
          default: 30
          mode: motion

I also tried with mqtt enabled, but nothing appeared in doubletake UI.

Here my doubletake config:

The UI shows that everything is well connected but I’m not able to use doubletake.

1 Like