Repository navigation
configs: ship PALETTE alone without the Voronoi split, the label remap or unblurred speckle - #150
Hendrik-code wants to merge 1 commit into
Conversation
…p or unblurred speckle transform_params_paper-palette-clean.json is the `Palette alone` arm (transform_params_paper-synthseg.json with RandomSynthSegGPU replaced by RandomPaletteGPU at p=1, hash 4b92045c) plus three PALETTE keys: * skip_sub_parc_prob 1.0: no Voronoi sub-parcellation, which cuts straight polygon edges through anatomy. * label_remap_prob 0.0: no per-label affine remap, which steps the intensity exactly on the ground-truth boundary. * blur_sigmas [0.5, 0.8]: a blur floor. The k-means bucketing is hard, so voxels next to a cluster threshold flip between two remaps and come out as salt-and-pepper. With the first two off, the synthesis before the blur is a function of input intensity alone; the test checks that, and that the defaults break it. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Results: Dataset803 and Dataset804, fold 0Worse than PALETTE alone out-of-domain on both datasets. Removing the Voronoi split, the label remap and unblurred speckle costs −0.016 per scan out-of-domain on both 803 and 804. On Dataset804 the Dixon water cohort collapses, from 0.778 to 0.539. PALETTE's value for domain generalization sits in the parts this config removes. Setup.
palette-clean is palette-alone plus Palette-clean minus palette-alone, paired mean and 95% CI:
In-domain is equivalent on both datasets. Out-of-domain is worse on both (TOST p = 1.000 and 0.967). Against the paper config, paired mean:
Cohort means, unweighted across cohorts:
Where it moves, per cohort against palette-alone:
Caveat: GPU model. This arm's Dataset804 run trained on an L40S; every reference trained on an A40. On Dataset804 Dixon water, the three arms of this round that trained on the L40S (#147, #149, #150) scored 0.724, 0.698 and 0.539. That is below all 23 other full-capacity arms on 804 (0.756 to 0.778) and all four runs of the paper config (0.767 to 0.771). An earlier paper-family run on the same L40S scored 0.759, with no scan more than 0.05 below the paper config, so the card alone does not produce the drop. With one run per cell, an interaction between card and arm is not excluded; an A40 rerun on 804 would settle it. In training. About 54 s/epoch on the A40 and 38 s/epoch on the L40S, no non-finite losses in either run. PALETTE fired in 20 of 20 pre-flight batches. Bottom line: do not merge this config. Which of the three keys carries the loss is not identified. Any cleanup of PALETTE's look should be tested one key at a time. |
What
One config,
transform_params_paper-palette-clean.json(hashba46ce7b). It is thePalette alonearm plus threeRandomPaletteGPUkeys.Palette aloneistransform_params_paper-synthseg.jsonwithRandomSynthSegGPUreplaced byRandomPaletteGPUatp: 1.0, hash4b92045c, the same config #142 ships astransform_params_paper-palette.json.skip_sub_parc_prob1.0label_remap_prob0.0blur_sigmas[0.5, 0.8]Every other key is
Palette alone's, so the two differ only in these three. No code change: all three are existingRandomPaletteGPUparameters. Based onmainrather than #142, so it does not wait on that PR; the.gitignoreentry is its own.Why
Seen in renders of the PALETTE arm on four SpineGAN datasets:
In 9 of 40 measured feature-only samples, intensity no longer separated the labels at all. The config asks whether PALETTE's measured gain on Dataset804 survives without those artifacts.
Tests
unit_tests/test_palette_clean_config.py:Full suite:
477 passed, 13 skipped. Pre-commit passes.🤖 Generated with Claude Code