5 July 2024

Changes DIPlib 3.5.1

Changes to DIPlib

dip::HoughTransformCircleCenters() now draws lines in the parameter space with an intensity proportional to the magnitude of the input gradient. Low gradients are likely less precise in orientation, so down-weighing them improves precision.

dip::saturated_mul() for dip::sint64 inputs was broken on systems that don’t have 128-bit arithmetic.
Improved dip::PhysicalQuantity::Normalize() to avoid floating-point rounding errors.
dip::DrawEllipsoid() and dip::DrawDiamond() could produce wonky shapes for even diameters, due to a floating-point rounding error.
JPEG encoding and decoding had a small memory leak.
dip::GetOptimalDFTSize() could loop indefinitely for inputs close to the maximum dip::uint value, because of an integer overflow in arithmetic.
The classes dip::StatisticsAccumulator and dip::VarianceAccumulator divided by 0 when adding two empty accumulators together, producing NaN values as output. This caused the function dip::SampleStatistics() to sometimes produce NaN statistics when called with a mask image.

None, but see bugfixes to DIPlib.

dip.ChainCode.Length() had a parameter boundaryPixels added in DIPlib 3.3.0, but the Python bindings did not know about its default value.

(See also bugfixes to DIPlib.)

The dip.viewer.SliceViewer object now also maps the complex and projection properties (matching the third and fourth columns of the “control panel”).

None