cornet

CorNET heart-rate regressor from a wrist PPG window.

Rebuilt from Biswas et al., “CorNET: Deep Learning Framework for PPG-Based Heart Rate Estimation and Biometric Identification in Ambulant Environment”, IEEE TBioCAS 13(2), 2019 (Sec. III, Fig. 6, Table III). The reference input is 8 s of band-passed, z-scored PPG at 125 Hz (1000 x 1). No code or weights are published, so constructed models are untrained.

Machine-readable model

• buildfunctionConstruct an untrained CorNET regressor with one linear output.

function

build

Python

Construct an untrained CorNET regressor with one linear output.

helia_edge/models/cornet.py:15

build(
    params: CorNetParams,
    input_shape: tuple[int | None, ...],
    *,
    batch_size: int | None = None,
    name: str | None = None,
) -> keras.Model

Construct an untrained CorNET regressor with one linear output.

Each convolution stage is Conv1D (valid, stride 1), batch normalization, ReLU, max pooling and dropout, following Fig. 6; the paper’s text places batch normalization after ReLU instead. Stride and padding are not stated and are inferred from Table III’s MAC counts. Every LSTM but the last returns its sequence.

Parameters

Parameters of build

Name	Type	Default	Description
params	CorNetParams	Required	Model parameters.
input_shape	tuple[int | None, ...]	Required	``(time, channels)``, both known.
batch_size	int | None	None	Static batch size; None for a dynamic batch.
name	str | None	None	Model name; the family when None.

Returns

Returns of build

Type	Description
keras.Model	keras.Model: The model, named ``cornet`` unless ``name`` is given.