aug_edge_artifacts — Configured pipeline of four edge artifacts

Group: Augmentation · C ABI: n4m_augmentation_edge_artifacts_*

Description

Combined edge artifact augmenter.

Parameters

Name

Type

Default

enabled_flags

int

15

overall_strength

float

1.0

detector_model

int

4

wavelengths

—

None

rng

Optional[PCG64]

None

seed

int

0

API and bindings

C ABI (ABI 2): n4m_augmentation_edge_artifacts_apply · n4m_augmentation_edge_artifacts_create · n4m_augmentation_edge_artifacts_destroy. Use the linked public header for the exact signature, configuration, and result handles.

Python (verified public re-export):

from n4m.augmentation.instrument import EdgeArtifactsAugmenter

Source signature: EdgeArtifactsAugmenter(enabled_flags: int = 15, overall_strength: float = 1.0, detector_model: int = 4, wavelengths = None, rng: Optional[PCG64] = None, seed: int = 0).

R: no current source-verified entry point was found for this catalog method.

MATLAB / Octave: no current source-verified entry point was found for this catalog method.

Explanations

Bibliographic source

No paper defines this composite. It is an implementation-defined pipeline combining truncated peaks, curvature, stray light, and detector roll-off.

Mathematical principle

enabled_flags selects four suboperators. Each receives a deterministic child PCG64 stream and source-defined defaults scaled by overall_strength. Enabled stages run in this exact order: truncated peaks, edge curvature, stray light, detector roll-off, with every stage consuming the previous stage’s output.

Appropriate uses

Combined stress tests of several boundary artifacts with one reproducible operator.

Limits and validation

Composition is noncommutative and overall_strength scales suboperators differently; it is not a universal scalar severity. Suboperator defaults are fixed rather than individually exposed.

Implementation

Python role API n4m.augmentation.instrument.EdgeArtifactsAugmenter; ABI 2 family n4m_augmentation_edge_artifacts_{create,apply,destroy}.

The ABI-2 implementation is the n4m_augmentation_edge_artifacts_* lifecycle in libn4m.

Sources and provenance

https://github.com/GBeurier/nirs4all-methods/blob/main/cpp/src/c_api/c_api_augmenters_edge_splines_random.cpp#L426


See also: methods index · interactive dashboard