pub fn resize_and_pad(
img: &RgbImage,
vcfg: &VisionConfig,
operator_max: Option<usize>,
) -> RgbImageExpand description
2026-09-25: Fit img inside the smart_resize canvas preserving aspect,
then zero-pad right/bottom to the canvas exactly. Returns the padded canvas.
- The pad is black pixels applied before normalisation, so padded cells
normalise to
(0 - mean)/std, not0.0. - The fit never upscales: image c of the reference fixtures (640x360) sits unresampled in a 644x364 canvas.
tests/glm_vision_preprocess_pin.rs compares the output with the
reference’s pixel_values for the fixture images.