Backends¶

The framework supports pluggable detection backends via the PipelineBackend interface. This allows comparing different detection approaches (classical CV, YOLO, external repositories) using the same evaluation pipeline and metrics.

Interface¶

from abc import ABC, abstractmethod

class PipelineBackend(ABC):
    @abstractmethod
    def run(self, image: np.ndarray, params: dict) -> PipelineResult: ...

Available Backends¶

Contour (Default)¶

The built-in 9-stage CV pipeline. This is the default and primary backend. See Pipeline Overview.

from wire_detection.pipeline.backends.contour import ContourBackend

backend = ContourBackend()
result = backend.run(image, params)

SINA¶

Wraps the CCL-based net discovery approach from Aldowaish et al. (arXiv:2601.22114) for benchmarking comparison.

The SINA approach masks components, runs CCL on remaining wires, and treats each CCL region as an electrical net. It produces blobs rather than individual wire segments.

from wire_detection.pipeline.backends.sina import SINABackend

backend = SINABackend()
result = backend.run(image, params)

Custom Backend¶

To add a new backend, implement the interface and register it:

from wire_detection.pipeline.registry import register_backend

class MyBackend(PipelineBackend):
    def run(self, image, params):
        lines = my_detection_function(image, params)
        return PipelineResult(
            lines=lines,
            raw_lines=lines,
            blob_count=0,
            stage_outputs={},
            params_used=params,
            elapsed_ms=0.0,
        )

register_backend("my_backend", MyBackend)

Once registered, the backend is available in the experiment engine and CLI:

wire-eval --backend my_backend --dataset hand_drawn