version: '1.0'
module:
  uid: ca43f6b9-8dcc-48fc-9b15-4bc157177788
  name: AdaChrom
  version: '1.0'
  description: 'Model-free rPPG heart rate estimation using adaptive chrominance analysis on skin ROI.'
  keywords:
    - Heart Rate
    - rPPG
  requirements:
    - seetapsych-hertz
    - numpy>=2.2.6
    - opencv-python>=4.13
    - scipy>=1.15.3
packages:
  - uid: 3d98f435-d484-4b91-acf4-f690c28b409f
    name: HeartRate-AdaChrom
    description: Adaptive chrominance rPPG heart rate estimator from adaptive forehead ROI, no neural model required
    version: '1.0.1'
    usage_models: []
    inputs: []
    provides:
      - face/heart_rate
    requires:
      - face/dense_landmarks
    entry:
      method: seetapsych_hertz.ada_chrom.package.load
    parameters:
      - name: window_samples
        text: Window Samples
        description: Sliding window frame count for HR estimation. Larger values reduce noise but increase latency; adjust based on real-time demand.
        type: integer
        value: 300
      - name: roi_regions
        text: ROI Regions
        description: List of regions to use for HR estimation. Defaults to ["skin_b_adaptive_forehead"].
        type: selection[]
        value: ["skin_b_adaptive_forehead"]
        selection: ["skin_all", "skin_legacy", "skin_a_fixed_forehead", "skin_b_adaptive_forehead", "skin_c_connected_components"]
    models: []