pith:YZ5ALXTV
U-SEG: Uncertainty in SEGmentation -- A systematic multi-variable exploration
A broad test of uncertainty estimation in segmentation finds that harder panoptic tasks reduce performance and that results vary sharply across datasets and backbones.
arxiv:2605.15421 v1 · 2026-05-14 · cs.CV
Add to your LaTeX paper
\usepackage{pith}
\pithnumber{YZ5ALXTVSLLOUYYSLIDULAK7PG}
Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge
Record completeness
Claims
a) the more challenging task of panoptic segmentation usually results in worse performance while high performance variance between datasets and backbones indicates that generalization is not guaranteed, b) time series samples can be useful for specific configurations, but in many cases are not worth the cost, c) sample diversity shows the most promise in the downstream task of calibration, but otherwise fails to beat simpler alternatives, d) a deterministic approach is adequate for some downstream tasks, but ensembles allow for significant improvements if the right conditions can be achieved in deployment.
The chosen collection of datasets, backbones, downstream tasks, and uncertainty methods is representative enough of real-world variability that the observed performance patterns can be treated as general guidance rather than artifacts of the specific experimental slice.
Systematic multi-variable experiments show panoptic segmentation yields poorer uncertainty quality than semantic, with high variance across datasets and backbones, limited value from time-series samples, calibration gains from sample diversity, and conditional benefits from ensembles over single det
References
Formal links
Receipt and verification
| First computed | 2026-05-20T00:00:57.695989Z |
|---|---|
| Builder | pith-number-builder-2026-05-17-v1 |
| Signature | Pith Ed25519
(pith-v1-2026-05) · public key |
| Schema | pith-number/v1.0 |
Canonical hash
c67a05de7592d6ea63125a0745815f798923b503cc49daaf38efc42675c612a1
Aliases
· · · · ·Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/YZ5ALXTVSLLOUYYSLIDULAK7PG \
| jq -c '.canonical_record' \
| python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: c67a05de7592d6ea63125a0745815f798923b503cc49daaf38efc42675c612a1
Canonical record JSON
{
"metadata": {
"abstract_canon_sha256": "7750e4fd6ba51733621dd5eca8da598fe49face7593e4090ecd49ef5c3f3f88a",
"cross_cats_sorted": [],
"license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
"primary_cat": "cs.CV",
"submitted_at": "2026-05-14T21:08:04Z",
"title_canon_sha256": "ed071394cbbc7205257ce391a4854594ca7350b78245c014533481619b002697"
},
"schema_version": "1.0",
"source": {
"id": "2605.15421",
"kind": "arxiv",
"version": 1
}
}