Skip to content

Releases: SCIInstitute/ShapeWorks

Development Build for Windows

Pre-release

Choose a tag to compare

@github-actions github-actions released this 19 Jul 20:13
b0f3bbb
Fix particle/time estimates when using initial landmarks (#2275) (#2602)

The optimization progress bar and time-remaining estimate are computed
from ComputeTotalIterations(), which assumed every domain starts from a
single seed particle and doubles at each split. When initial landmarks
(or initial points) are used, each domain instead starts with many
particles, so far fewer splits occur. The old estimate therefore greatly
overcounted the total iterations, causing the progress bar to stall at a
low percentage and the time-remaining estimate to be far too high.

Account for the initial seed count per domain: derive it from the
sampler's initial points / point files via GetInitialParticleCount(), and
drive the split count from the least-seeded domain across all shapes so
fixed-domain cases (which still seed some shapes with a single particle)
keep the full split count. The non-landmark case is unchanged.

Development Build for Mac Arm64

Pre-release

Choose a tag to compare

@github-actions github-actions released this 19 Jul 19:40
b0f3bbb
Fix particle/time estimates when using initial landmarks (#2275) (#2602)

The optimization progress bar and time-remaining estimate are computed
from ComputeTotalIterations(), which assumed every domain starts from a
single seed particle and doubles at each split. When initial landmarks
(or initial points) are used, each domain instead starts with many
particles, so far fewer splits occur. The old estimate therefore greatly
overcounted the total iterations, causing the progress bar to stall at a
low percentage and the time-remaining estimate to be far too high.

Account for the initial seed count per domain: derive it from the
sampler's initial points / point files via GetInitialParticleCount(), and
drive the split count from the least-seeded domain across all shapes so
fixed-domain cases (which still seed some shapes with a single particle)
keep the full split count. The non-landmark case is unchanged.

Development Build for Mac

Pre-release

Choose a tag to compare

@github-actions github-actions released this 19 Jul 20:22
b0f3bbb
Fix particle/time estimates when using initial landmarks (#2275) (#2602)

The optimization progress bar and time-remaining estimate are computed
from ComputeTotalIterations(), which assumed every domain starts from a
single seed particle and doubles at each split. When initial landmarks
(or initial points) are used, each domain instead starts with many
particles, so far fewer splits occur. The old estimate therefore greatly
overcounted the total iterations, causing the progress bar to stall at a
low percentage and the time-remaining estimate to be far too high.

Account for the initial seed count per domain: derive it from the
sampler's initial points / point files via GetInitialParticleCount(), and
drive the split count from the least-seeded domain across all shapes so
fixed-domain cases (which still seed some shapes with a single particle)
keep the full split count. The non-landmark case is unchanged.

Development Build for Linux

Pre-release

Choose a tag to compare

@github-actions github-actions released this 19 Jul 19:32
b0f3bbb
Fix particle/time estimates when using initial landmarks (#2275) (#2602)

The optimization progress bar and time-remaining estimate are computed
from ComputeTotalIterations(), which assumed every domain starts from a
single seed particle and doubles at each split. When initial landmarks
(or initial points) are used, each domain instead starts with many
particles, so far fewer splits occur. The old estimate therefore greatly
overcounted the total iterations, causing the progress bar to stall at a
low percentage and the time-remaining estimate to be far too high.

Account for the initial seed count per domain: derive it from the
sampler's initial points / point files via GetInitialParticleCount(), and
drive the split count from the least-seeded domain across all shapes so
fixed-domain cases (which still seed some shapes with a single particle)
keep the full split count. The non-landmark case is unchanged.

ShapeWorks v6.7.0

Choose a tag to compare

@akenmorris akenmorris released this 24 Apr 21:52

ShapeWorks v6.6.1

Choose a tag to compare

@akenmorris akenmorris released this 19 May 20:09

ShapeWorks 6.6.1 is a patch release to the recent 6.6.0 release. The release notes are here:

https://sciinstitute.github.io/ShapeWorks/6.6/about/release-notes.html

ShapeWorks v6.6.0

Choose a tag to compare

@akenmorris akenmorris released this 06 May 06:03

ShapeWorks v6.5.1

Choose a tag to compare

@akenmorris akenmorris released this 20 Aug 17:30
  • Fix VTK Python version for windows (#2308)
  • Fix some alignment issues in grooming
  • Fix crash during ICP alignment with multiple domains

ShapeWorks v6.5.0

Choose a tag to compare

@akenmorris akenmorris released this 16 Mar 20:20

ShapeWorks v6.4.2

Choose a tag to compare

@akenmorris akenmorris released this 10 Jul 06:49