Frigate configuration help

Hi,
I just install Frigate and succeed to add my Woox Camera on it.
I need help to add some track objects (cat, dog, person) and bounding box.
Here is my actual conf :

## Connection à MQTT ##
mqtt:
  enabled: false

#cameras:
#  dummy_camera: # <--- this will be changed to your actual camera later
#    enabled: False
#    ffmpeg:
#      inputs:
#        - path: rtsp://127.0.0.1:554/rtsp
#          roles:
#            - detect
# /config/frigate.yaml

## Choix des caméras ##
go2rtc:
  streams:
    r4252_smart_outdoor_camera:
    - echo:bash /config/custom_components/expose_camera_stream_source/get_stream.sh
      camera.r4252_smart_outdoor_camera

## Liste des caméras avec leurs adresses de connection ##
cameras:
  r4252_smart_outdoor_camera:
    ffmpeg:
      inputs:
      - path: rtsp://127.0.0.1:8554/r4252_smart_outdoor_camera?video
        input_args: preset-rtsp-restream-low-latency
        roles:
        - detect
#    motion: ## Création des zones à ignorer ##
#      mask:
#        - 0,461,3,0,1919,0,1919,843,1699,492,1344,458,1346,336,973,317,869,375,866,432
## Choix des options de détection et d'enregistrement ##
    motion:
      mask:
      - 262,0,259,36,32,38,31,0
record:
  enabled: true
  retain:
    days: 7
    mode: motion
  events:
    retain:
      default: 30
#      mode: motion ## Conserve les vidéos avec mouvements seulement ##
      mode: active_objects ## Conserve les vidéos avec détection d'objets ##
      objects:
        dog: 2 ## Durée de conservation par objets ##
        cat: 2
        person: 7
#detect:
#  stationary:
    # Optional: Frequency for confirming stationary objects (default: same as threshold)
    # When set to 1, object detection will run to confirm the object still exists on every frame.
    # If set to 10, object detection will run to confirm the object still exists on every 10th frame.
#    interval: 50
    # Optional: Number of frames without a position change for an object to be considered stationary (default: 10x the frame rate or 10s)
#    threshold: 50
    # Optional: Define a maximum number of frames for tracking a stationary object (default: not set, track forever)
    # This can help with false positives for objects that should only be stationary for a limited amount of time.
    # It can also be used to disable stationary object tracking. For example, you may want to set a value for person, but leave
    # car at the default.
    # WARNING: Setting these values overrides default behavior and disables stationary object tracking.
    #          There are very few situations where you would want it disabled. It is NOT recommended to
    #          copy these values from the example config into your config unless you know they are needed.

# Optional: Object configuration
# NOTE: Can be overridden at the camera level
objects:
  # Optional: list of objects to track from labelmap.txt (default: shown below)
  track:
  - person
  - dog
  - cat
  # Optional: mask to prevent all object types from being detected in certain areas (default: no mask)
  # Checks based on the bottom center of the bounding box of the object.
  # NOTE: This mask is COMBINED with the object type specific mask below
  # mask: 0,0,1000,0,1000,200,0,200
  # Optional: filters to reduce false positives for specific object types
#  filters:
#    person:
      # Optional: minimum width*height of the bounding box for the detected object (default: 0)
#      min_area: 5000
      # Optional: maximum width*height of the bounding box for the detected object (default: 24000000)
#      max_area: 100000
      # Optional: minimum width/height of the bounding box for the detected object (default: 0)
#      min_ratio: 0.5
      # Optional: maximum width/height of the bounding box for the detected object (default: 24000000)
#      max_ratio: 2.0
      # Optional: minimum score for the object to initiate tracking (default: shown below)
#      min_score: 0.5
      # Optional: minimum decimal percentage for tracked object's computed score to be considered a true positive (default: shown below)
#      threshold: 0.7
      # Optional: mask to prevent this object type from being detected in certain areas (default: no mask)
      # Checks based on the bottom center of the bounding box of the object
#      mask: 0,0,1000,0,1000,200,0,200

snapshots:
  enabled: true

But the camera not detect well dog and person, sometimes dog are marked as person, and camera not always detect objects as we can see on history events


The first picture mark dog as person.
And between 9am and 12 noon I make more than 0 move in front of the camera.

Thanks.

it would most likely help if the stream used for detect was upright and not sideways as it appears to be

Yes, with camera upright it’s better.
But still false object detect.

the default model used is trained on the COCO dataset (examples: Imgur: The magic of the Internet)

these images are not from the security camera perspective so false positives are not unlikely. The docs provide ways to filter out false positives and better models will produce more accurate results

My conf :

## Connection à MQTT ##
#mqtt:
#  enabled: false
mqtt:
  host: X.X.X.X
  port: XXX
  client_id: frigate
  topic_prefix: frigate
  user: xxx
  password: XxX

#cameras:
#  dummy_camera: # <--- this will be changed to your actual camera later
#    enabled: False
#    ffmpeg:
#      inputs:
#        - path: rtsp://127.0.0.1:554/rtsp
#          roles:
#            - detect
# /config/frigate.yaml

## Choix des caméras ##
go2rtc:
  streams:
    r4252_smart_outdoor_camera:
    - echo:bash /config/custom_components/expose_camera_stream_source/get_stream.sh
      camera.r4252_smart_outdoor_camera

## Liste des caméras avec leurs adresses de connection ##
cameras:
  r4252_smart_outdoor_camera:
    ffmpeg:
      inputs:
      - path: rtsp://127.0.0.1:8554/r4252_smart_outdoor_camera?video
        input_args: preset-rtsp-restream-low-latency
        roles:
        - detect
#    motion: ## Création des zones à ignorer ##
#      mask:
#        - 0,461,3,0,1919,0,1919,843,1699,492,1344,458,1346,336,973,317,869,375,866,432
## Choix des options de détection et d'enregistrement ##
    motion:
      mask:
      - 262,0,259,36,32,38,31,0
record:
  enabled: true
  retain:
    days: 7
    mode: motion
  events:
    retain:
      default: 15
#      mode: motion ## Conserve les vidéos avec mouvements seulement ##
      mode: active_objects ## Conserve les vidéos avec détection d'objets ##
      objects:
        dog: 2 ## Durée de conservation par objets ##
        cat: 2
        person: 7
#detect:
#  stationary:
    # Optional: Frequency for confirming stationary objects (default: same as threshold)
    # When set to 1, object detection will run to confirm the object still exists on every frame.
    # If set to 10, object detection will run to confirm the object still exists on every 10th frame.
#    interval: 50
    # Optional: Number of frames without a position change for an object to be considered stationary (default: 10x the frame rate or 10s)
#    threshold: 50
    # Optional: Define a maximum number of frames for tracking a stationary object (default: not set, track forever)
    # This can help with false positives for objects that should only be stationary for a limited amount of time.
    # It can also be used to disable stationary object tracking. For example, you may want to set a value for person, but leave
    # car at the default.
    # WARNING: Setting these values overrides default behavior and disables stationary object tracking.
    #          There are very few situations where you would want it disabled. It is NOT recommended to
    #          copy these values from the example config into your config unless you know they are needed.

# Optional: Object configuration
# NOTE: Can be overridden at the camera level
objects:
  # Optional: list of objects to track from labelmap.txt (default: shown below)
  track:
  - person
  - dog
  - cat
  # Optional: mask to prevent all object types from being detected in certain areas (default: no mask)
  # Checks based on the bottom center of the bounding box of the object.
  # NOTE: This mask is COMBINED with the object type specific mask below
  # mask: 0,0,1000,0,1000,200,0,200
  # Optional: filters to reduce false positives for specific object types
  filters:
    person:
      # Optional: minimum width*height of the bounding box for the detected object (default: 0)
#      min_area: 5000
      # Optional: maximum width*height of the bounding box for the detected object (default: 24000000)
#      max_area: 100000
      # Optional: minimum width/height of the bounding box for the detected object (default: 0)
#      min_ratio: 0.5
      # Optional: maximum width/height of the bounding box for the detected object (default: 24000000)
#      max_ratio: 2.0
      # Optional: minimum score for the object to initiate tracking (default: shown below)
      min_score: 0.5
      # Optional: minimum decimal percentage for tracked object's computed score to be considered a true positive (default: shown below)
      threshold: 0.7
    dog:
      # Optional: minimum width*height of the bounding box for the detected object (default: 0)
#      min_area: 5000
      # Optional: maximum width*height of the bounding box for the detected object (default: 24000000)
#      max_area: 100000
      # Optional: minimum width/height of the bounding box for the detected object (default: 0)
#      min_ratio: 0.5
      # Optional: maximum width/height of the bounding box for the detected object (default: 24000000)
#      max_ratio: 2.0
      # Optional: minimum score for the object to initiate tracking (default: shown below)
      min_score: 0.2
      # Optional: minimum decimal percentage for tracked object's computed score to be considered a true positive (default: shown below)
      threshold: 0.5
      # Optional: mask to prevent this object type from being detected in certain areas (default: no mask)
      # Checks based on the bottom center of the bounding box of the object
#      mask: 0,0,1000,0,1000,200,0,200
    cat:
      # Optional: minimum width*height of the bounding box for the detected object (default: 0)
#      min_area: 5000
      # Optional: maximum width*height of the bounding box for the detected object (default: 24000000)
#      max_area: 100000
      # Optional: minimum width/height of the bounding box for the detected object (default: 0)
#      min_ratio: 0.5
      # Optional: maximum width/height of the bounding box for the detected object (default: 24000000)
#      max_ratio: 2.0
      # Optional: minimum score for the object to initiate tracking (default: shown below)
      min_score: 0.2
      # Optional: minimum decimal percentage for tracked object's computed score to be considered a true positive (default: shown below)
      threshold: 0.5

snapshots:
  enabled: true

Once a week, I go on Frigate’s events to Edit in Frigate+ false result like my dog seen as a cat.
But still wrong result.

It’s normal?

Hello,
I have a problem with Frigate configuration and object detection.

I was able to integrate Frigate and my camera. But objects were poorly detected, so I took a Coral USB that I just received, but despite it the detection is poor.
For example, my cat is never seen.

Configuration

Frigate 0.13.2
Home Assistant
Core2024.7.1
Supervisor2024.06.2
Operating System12.4
Frontend20240705.0

Frigtate configuration

## Connection to xxx ##
mqtt:
  host: MonIP
  port: 1883
  client_id: frigate
  topic_prefix: frigate
  user: xxx
  password: xxx


## Camera selection ##
go2rtc:
  streams:
    r4252_smart_outdoor_camera:
    - echo:bash /config/custom_components/expose_camera_stream_source/get_stream.sh
      camera.r4252_smart_outdoor_camera

## List of cameras with their connection addresses ##
cameras:
  r4252_smart_outdoor_camera:
    ffmpeg:
      inputs:
      - path: rtsp://127.0.0.1:8554/r4252_smart_outdoor_camera?video
        input_args: preset-rtsp-restream-low-latency
        roles:
        - detect

# motion: ## Create ignore zones ##
# mask:
# - 0,461,3,0,1919,0,1919,843,1699,492,1344,458,1346,336,973,317,869,375,866,432
    motion:
      mask:
      - 262,0,259,36,32,38,31,0

## Select detection and recording options ##
record:
  enabled: true
  retain:
    days: 7
    mode: all
  events:
# Recording seconds before and after event
    pre_capture: 60
    post_capture: 60
    retain:
#Duration

Coral is well detected on Frigate
image
But only one object was detected since yesterday, my dog, camera never see my cat.
image