[Config Support]:

[loneswimmer]

Describe the problem you are having

I have enabled object detection and facial recognition, but have questions/issues:

1. I am getting recognition in the "facial recognition" tab...

2. ... but no recognition or labelling is showing in review tab. I just see a video of the motion detected under "alerts" (nothing at at under "detections" or "motion"). From what I understand this is ordered alerts > detections > motion)?

3. In explore tab it shows boxes around objects, and type (e.g. cat or person), but not who they are. Is that correct?

4. What is the "label" option? Can I get it to learn to detect between different cats?

5. Also, whilst mostly correct, sometimes it gets muddled between "person" and "cat". How can I improve on that?

Version

0.17.0-f0d69f7

Frigate config file

#####  Modified date:  17/04/2026



face_recognition:
  # Optional: Enable face recognition (default: shown below)
  enabled: true
  # Optional: Minimum face distance score required to mark as a potential match (default: shown below)
  unknown_score: 0.8
  # Optional: Minimum face detection score required to detect a face (default: shown below)
  # NOTE: This only applies when not running a Frigate+ model
  detection_threshold: 0.7
  # Optional: Minimum face distance score required to be considered a match (default: shown below)
  recognition_threshold: 0.9
  # Optional: Min area of detected face box to consider running face recognition (default: shown below)
  min_area: 500
  # Optional: Min face recognitions for the sub label to be applied to the person object (default: shown below)
  min_faces: 1
  # Optional: Number of images of recognized faces to save for training (default: shown below)
  save_attempts: 200
  # Optional: Apply a blur quality filter to adjust confidence based on the blur level of the image (default: shown below)
  blur_confidence_filter: true
  # Optional: Set the model size used face recognition. (default: shown below)
  model_size: small
  # Optional: Target a specific device to run the model (default: shown below)
  # NOTE: See https://onnxruntime.ai/docs/execution-providers/ for more information
  device: None



objects:
  # Optional: list of objects to track from labelmap.txt (default: shown below)
  track:
    - person
    - 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: 0.000,0.000,0.781,0.000,0.781,0.278,0.000,0.278
  # Optional: filters to reduce false positives for specific object types
  filters:
    person:
      # Optional: minimum size of the bounding box for the detected object (default: 0).
      # Can be specified as an integer for width*height in pixels or as a decimal representing the percentage of the frame (0.000001 to 0.99).
      min_area: 5000
      # Optional: maximum size of the bounding box for the detected object (default: 24000000).
      # Can be specified as an integer for width*height in pixels or as a decimal representing the percentage of the frame (0.000001 to 0.99).
      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: 0.000,0.000,0.781,0.000,0.781,0.278,0.000,0.278




review:
  # Optional: alerts configuration
  alerts:
    # Optional: enables alerts for the camera (default: shown below)
    enabled: true
    # Optional: labels that qualify as an alert (default: shown below)
    labels:
      - car
      - person
      # Time to cutoff alerts after no alert-causing activity has occurred (default: shown below)
    cutoff_time: 40
    # Optional: required zones for an object to be marked as an alert (default: none)
    # NOTE: when settings required zones globally, this zone must exist on all cameras
    #       or the config will be considered invalid. In that case the required_zones
    #       should be configured at the camera level.

    # Optional: detections configuration
  detections:
    # Optional: enables detections for the camera (default: shown below)
    enabled: true
    # Optional: labels that qualify as a detection (default: all labels that are tracked / listened to)
    labels:
      - cat
    # Time to cutoff detections after no detection-causing activity has occurred (default: shown below)
    cutoff_time: 30
    # Optional: required zones for an object to be marked as a detection (default: none)
    # NOTE: when settings required zones globally, this zone must exist on all cameras
    #       or the config will be considered invalid. In that case the required_zones
    #       should be configured at the camera level.





mqtt:
  enabled: false

detectors:
  coral:
    type: edgetpu
    device: pci

audio:
  enabled: true

detect:
  enabled: true

record:
  enabled: true
  continuous:
    days: 7
  alerts:
    retain:
      days: 30
      mode: motion
  detections:
    retain:
      days: 30
      mode: motion
  motion:
    days: 7

snapshots:
  enabled: true
  retain:
    default: 30

ffmpeg:
  hwaccel_args: preset-vaapi
  output_args:
    record: preset-record-generic-audio-aac

go2rtc:
  streams:
    Cam_01_Kitchen:
      - rtsp://redacted:redacted@192.168.123.81:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_01_Kitchen#audio=opus
    Cam_02_Workshop:
      - rtsp://redacted:redacted@192.168.123.82:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_02_Workshop#audio=opus
    Cam_03_Lounge:
      - rtsp://redacted:redacted@192.168.123.83:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_03_Lounge#audio=opus
    Cam_04_Landing:
      - rtsp://redacted:redacted@192.168.123.84:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_04_Landing#audio=opus
    Cam_05_Cabin:
      - rtsp://redacted:redacted@192.168.123.85:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_05_Cabin#audio=opus
    Cam_06_B3:
      - rtsp://redacted:redacted@192.168.123.86:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_06_B3#audio=opus
    Cam_07_B4:
      - rtsp://redacted:redacted@192.168.123.87:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_07_B4#audio=opus
    Cam_08_B2:
      - rtsp://redacted:redacted@192.168.123.88:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_08_B2#audio=opus
    Cam_09_B1:
      - rtsp://redacted:redacted@192.168.123.89:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_09_B1)#audio=opus
    Cam_10_Garage:
      - rtsp://redacted:redacted@192.168.123.90:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_10_Garage#audio=opus
    Cam_11_Cons:
      - rtsp://redacted:redacted@192.168.123.91:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_11_Cons#audio=opus
    Cam_12_Porch:
      - rtsp://redacted:redacted@192.168.123.92:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_12_Porch#audio=opus
    Cam_13_Drive:
      - rtsp://redacted:redacted@192.168.123.93:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_13_Drive#audio=opus
    Cam_14_Garden:
      - rtsp://redacted:redacted@192.168.123.94:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_14_Garden#audio=opus
    Cam_15_Kitchen2:
      - rtsp://redacted:redacted@192.168.123.95:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_15_Kitchen2#audio=opus
    Cam_16:
      - rtsp://redacted:redacted@192.168.123.96:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_15_Kitchen2#audio=opus
    Cam_17:
      - rtsp://redacted:redacted@192.168.123.97:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_15_Kitchen2#audio=opus
    Cam_18:
      - rtsp://redacted:redacted@192.168.123.98:554/H265/ch1/main/av_stream
      - ffmpeg:Cam_15_Kitchen2#audio=opus


#################################################   C A M E R A S   #################################################

cameras:


##################################   Cam01   ##################################

  Cam_01_Kitchen:
    enabled: true
    ffmpeg:
      hwaccel_args: preset-vaapi
      output_args:
        record: preset-record-generic-audio-aac
      inputs:
        - path: 
            rtsp://redacted:redacted@192.168.123.81:554/H265/ch2/sub/av_stream
          roles:
            - detect
        - path: 
            rtsp://redacted:redacted@192.168.123.81:554/H265/ch1/main/av_stream
          roles:
            - record
        - path: rtsp://127.0.0.1:8554/Cam_01_Kitchen
          input_args: preset-rtsp-restream
          roles:
            - audio
    detect:
      enabled: true
      width: 1280
      height: 720
      fps: 6
    motion:
      mask:
        - 0.452,0.085,0.45,0.178,0.477,0.183,0.479,0.092
        - 0.022,0.011,0.017,0.1,0.323,0.11,0.317,0.005
    record:
      enabled: true



etc. for other cams...

Relevant Frigate log output

2026-04-11 21:21:38.225435600  [INFO] Preparing Frigate...
2026-04-11 21:21:38.669688799  [INFO] Starting Frigate...
2026-04-11 21:21:43.106671742  [2026-04-11 21:21:43] frigate.util.config            INFO    : Checking if frigate config needs migration...
2026-04-11 21:21:43.186799051  [2026-04-11 21:21:43] frigate.util.config            INFO    : frigate config does not need migration...
2026-04-11 21:21:43.433073924  [2026-04-11 21:21:43] frigate.app                    INFO    : Starting Frigate (0.17.0-f0d69f7)
2026-04-11 21:21:43.623563312  [2026-04-11 21:21:43] peewee_migrate.logs            INFO    : Starting migrations
2026-04-11 21:21:43.624612170  [2026-04-11 21:21:43] peewee_migrate.logs            INFO    : There is nothing to migrate
2026-04-11 21:21:44.075631106  [2026-04-11 21:21:44] frigate.app                    INFO    : Recording process started: 1338
2026-04-11 21:21:44.498995828  [2026-04-11 21:21:44] frigate.app                    INFO    : Review process started: 1351
2026-04-11 21:21:44.500526709  [2026-04-11 21:21:44] frigate.app                    INFO    : go2rtc process pid: 128
2026-04-11 21:21:45.151892591  [2026-04-11 21:21:45] frigate.app                    INFO    : Embedding process started: 1361
2026-04-11 21:21:45.596292446  [2026-04-11 21:21:45] frigate.detectors.plugins.edgetpu_tfl INFO    : Attempting to load TPU as pci
2026-04-11 21:21:45.600749854  INFO: Created TensorFlow Lite XNNPACK delegate for CPU.
2026-04-11 21:21:45.604411864  [2026-04-11 21:21:45] frigate.detectors.plugins.edgetpu_tfl INFO    : TPU found
2026-04-11 21:21:46.066940165  [2026-04-11 21:21:46] frigate.app                    INFO    : Output process started: 1444
2026-04-11 21:21:46.728091751  [2026-04-11 21:21:46] frigate.api.fastapi_app        INFO    : Starting FastAPI app
2026-04-11 21:21:46.926604675  [2026-04-11 21:21:46] frigate.audio_manager          INFO    : Audio processor started (pid: 1492)
2026-04-11 21:21:47.046149936  [2026-04-11 21:21:47] frigate.api.fastapi_app        INFO    : FastAPI started
2026-04-11 21:21:47.070824534  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Camera processor started for Cam_01_Kitchen: 1801
2026-04-11 21:21:47.115466236  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Capture process started for Cam_01_Kitchen: 1862
2026-04-11 21:21:47.157092789  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Camera processor started for Cam_02_Workshop: 1928
2026-04-11 21:21:47.196140302  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Capture process started for Cam_02_Workshop: 1985
2026-04-11 21:21:47.234870284  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Camera processor started for Cam_03_Lounge: 2051
2026-04-11 21:21:47.348894210  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Capture process started for Cam_03_Lounge: 2108
2026-04-11 21:21:47.400932945  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Camera processor started for Cam_04_Landing: 2175
2026-04-11 21:21:47.438059420  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Capture process started for Cam_04_Landing: 2232
2026-04-11 21:21:47.480724937  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Camera processor started for Cam_05_Cabin: 2298
2026-04-11 21:21:47.518748093  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Capture process started for Cam_05_Cabin: 2355
2026-04-11 21:21:47.562235476  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Camera processor started for Cam_06_B3: 2421
2026-04-11 21:21:47.600020140  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Capture process started for Cam_06_B3: 2478
2026-04-11 21:21:47.639038390  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Camera processor started for Cam_07_B4: 2544
2026-04-11 21:21:47.676963919  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Capture process started for Cam_07_B4: 2601
2026-04-11 21:21:47.718824356  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Camera processor started for Cam_08_B2: 2667
2026-04-11 21:21:47.757071951  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Capture process started for Cam_08_B2: 2724
2026-04-11 21:21:47.800946020  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Camera processor started for Cam_09_B1: 2797
2026-04-11 21:21:47.841863985  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Capture process started for Cam_09_B1: 2909
2026-04-11 21:21:47.888969285  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Camera processor started for Cam_10_Garage: 3029
2026-04-11 21:21:47.937400660  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Capture process started for Cam_10_Garage: 3158
2026-04-11 21:21:47.979196564  [2026-04-11 21:21:47] frigate.camera.maintainer      INFO    : Camera processor started for Cam_11_Cons: 3260
2026-04-11 21:21:48.061050866  [2026-04-11 21:21:48] frigate.camera.maintainer      INFO    : Capture process started for Cam_11_Cons: 3317
2026-04-11 21:21:48.102725336  [2026-04-11 21:21:48] frigate.camera.maintainer      INFO    : Camera processor started for Cam_12_Porch: 3384
2026-04-11 21:21:48.141257120  [2026-04-11 21:21:48] frigate.camera.maintainer      INFO    : Capture process started for Cam_12_Porch: 3441
2026-04-11 21:21:48.181732419  [2026-04-11 21:21:48] frigate.camera.maintainer      INFO    : Camera processor started for Cam_13_Drive: 3507
2026-04-11 21:21:48.219902594  [2026-04-11 21:21:48] frigate.camera.maintainer      INFO    : Capture process started for Cam_13_Drive: 3564
2026-04-11 21:21:48.258933326  [2026-04-11 21:21:48] frigate.camera.maintainer      INFO    : Camera processor started for Cam_14_Garden: 3630
2026-04-11 21:21:48.297183043  [2026-04-11 21:21:48] frigate.camera.maintainer      INFO    : Capture process started for Cam_14_Garden: 3687
2026-04-11 21:21:48.337977940  [2026-04-11 21:21:48] frigate.camera.maintainer      INFO    : Camera processor started for Cam_15_Kitchen2: 3753
2026-04-11 21:21:48.392914942  [2026-04-11 21:21:48] frigate.camera.maintainer      INFO    : Capture process started for Cam_15_Kitchen2: 3864
2026-04-11 21:21:48.433745746  [2026-04-11 21:21:48] frigate.camera.maintainer      INFO    : Camera processor started for Cam_16: 3930
2026-04-11 21:21:48.472868055  [2026-04-11 21:21:48] frigate.camera.maintainer      INFO    : Capture process started for Cam_16: 3987
2026-04-11 21:21:48.514724074  [2026-04-11 21:21:48] frigate.camera.maintainer      INFO    : Camera processor started for Cam_17: 4053
2026-04-11 21:21:48.555829646  [2026-04-11 21:21:48] frigate.camera.maintainer      INFO    : Capture process started for Cam_17: 4110
2026-04-11 21:21:48.594865527  [2026-04-11 21:21:48] frigate.camera.maintainer      INFO    : Camera processor started for Cam_18: 4176
2026-04-11 21:21:48.637066111  [2026-04-11 21:21:48] frigate.camera.maintainer      INFO    : Capture process started for Cam_18: 4237

Relevant go2rtc log output

2026-04-11 21:21:38.669789999  [INFO] Preparing new go2rtc config...
2026-04-11 21:21:39.017115807  [INFO] Starting go2rtc...
2026-04-11 21:21:39.337512529  21:21:39.337 INF go2rtc platform=linux/amd64 revision=df95ce3 version=1.9.10
2026-04-11 21:21:39.337515994  21:21:39.337 INF config path=/config/go2rtc_homekit.yml
2026-04-11 21:21:39.337963549  21:21:39.337 INF [api] listen addr=:1984
2026-04-11 21:21:39.338308115  21:21:39.338 INF [rtsp] listen addr=:8554
2026-04-11 21:21:39.338668911  21:21:39.338 INF [webrtc] listen addr=:8555
2026-04-11 21:21:48.075384894  [INFO] Starting go2rtc healthcheck service...
2026-04-11 21:22:06.034589531  21:22:06.034 WRN github.com/AlexxIT/go2rtc/internal/rtsp/rtsp.go:260 > error="read tcp 127.0.0.1:8554->127.0.0.1:43388: read: connection reset by peer"

Frigate stats

No response

Operating system

Debian

Install method

Docker Compose

docker-compose file or Docker CLI command

n/a

Object Detector

Coral

Screenshots of the Frigate UI's System metrics pages

No response

Any other information that may be helpful

No response

[inkeep-ai-assistant[bot]]

Here are answers to your questions:

1 & 2. Face recognition working in Face Library but not showing in Review tab:

When a known person is recognized, their name is added as a sub_label. (1) This sub_label should appear in the Review UI. However, face recognition considers recognition scores across all recognition attempts for each person object, and a sub label will only be assigned if a person is confidently recognized consistently. (1)

Your config looks correct. Check that:

• The recognition scores in the Face Library are meeting your recognition_threshold: 0.9 consistently
• You have trained sufficient faces per person (minimum 5-10 images recommended, 20-30 for more robust results) (1)

3. Explore tab showing object type but not identity:

Yes, that's expected. The Explore view shows tracked objects with their labels (person, cat). The sub_label (recognized name) appears in the Tracked Object Details pane within Explore when you select a specific event. (2)

4. Distinguishing between different cats:

Face recognition only works for person objects—Frigate needs to first detect a person before it can detect and recognize a face. (1)

To distinguish between cats, you can use Object Classification, which allows you to train a custom MobileNetV2 classification model to identify finer categories for tracked objects. (2) You would:

• Define classes like Leo, Charlie, None for your cats
• Train the model with images of each cat
• The classification result can be applied as a sub_label (2)

5. Improving person vs cat detection accuracy:

If you're using the default model and getting confusion between person and cat, consider:

• Adjusting the threshold filter for each object type in your config (3)
• Using Frigate+ models, which are trained on real security camera footage and provide better accuracy for surveillance scenarios (4)
• Submitting both true positives and false positives if using Frigate+ to help the model differentiate between objects (4)

📚 Sources:

1. Face Recognition
2. Object Classification
3. Full Reference Config
4. Models

---

Have another question? Just tag @inkeep.