normalised_rect_to_region()

MAYAFLUX_API Region MayaFlux::Kakshya::normalised_rect_to_region	(	float	nx,
		float	ny,
		float	nw,
		float	nh,
		uint32_t	pixel_w,
		uint32_t	pixel_h
	)

Convert a normalised image rectangle to a pixel-space Region.

Input coordinates are in image space: origin top-left, x and y in [0, 1], +Y downward. Output Region uses the IMAGE_COLOR coordinate convention: start/end as { y0, x0 } / { y1, x1 } (inclusive, clamped to [0, w-1] x [0, h-1]).

This is the inverse of the normalisation performed by ConnectedComponents and Contours, and is the correct input form for get_region_data() on any pixel-bearing SignalSourceContainer.

Parameters

nx	Normalised left edge in [0, 1].
ny	Normalised top edge in [0, 1].
nw	Normalised width in [0, 1].
nh	Normalised height in [0, 1].
pixel_w	Source image width in pixels.
pixel_h	Source image height in pixels.

Returns

Region with 2D spatial coordinates in pixel integers.

Definition at line 377 of file CoordUtils.cpp.

{
    const auto fw = static_cast<float>(pixel_w);
    const auto fh = static_cast<float>(pixel_h);
 
    const auto x0 = static_cast<uint64_t>(std::clamp(nx * fw, 0.0F, fw - 1.0F));
    const auto y0 = static_cast<uint64_t>(std::clamp(ny * fh, 0.0F, fh - 1.0F));
    const auto x1 = static_cast<uint64_t>(std::clamp((nx + nw) * fw - 1.0F, 0.0F, fw - 1.0F));
    const auto y1 = static_cast<uint64_t>(std::clamp((ny + nh) * fh - 1.0F, 0.0F, fh - 1.0F));
 
    return Region(std::vector { y0, x0 }, std::vector { y1, x1 });
}

Referenced by MayaFlux::Yantra::VisionExtractor< InputType, OutputType >::extract_implementation(), and normalised_points_tight_region().

Here is the caller graph for this function: