SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2008.05105 · 2020

Local Temperature Scaling for Probability Calibration

arXiv · PDF · Open in the Atlas

Code that ran

3 of 3 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 0 failures of our instrument; 0 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2008.05105.json · badge: /api/ran/2008.05105.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
uncbiag/LTS community (classifier label)
classifier basis not recorded on the link
3 ran · 0 no recorded run
FunctionStatusWhere it lives
Calculate_ECE Ran
ran on a synthesized input
uncbiag/LTS/code/probability_measure_CamVid.py
code served (permissive licence) · get_code("9cf240d39679a80a")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

np.random.seed(0)

# Create small segmentation maps (2x3 spatial dimensions)
height, width = 2, 3

# confidence: probability map, values in [0, 1]
confidence = np.random.rand(height, width).astype(np.float32)

# prediction: predicted class labels
prediction = np.random.randint(0, 2, size=(height, width)).astype(np.int32)

# gt: ground-truth labels (values != 100 are valid regions, 100 is unlabeled)
gt = np.array([[0, 1, 0],
               [1, 0, 100]], dtype=np.int32)

# boundary: boundary mask (0 for non-boundary, 1 for boundary)
boundary = np.array([[0, 1, 0],
                     [1, 0, 0]], dtype=np.int32)

ARGS = {
    "confidence": confidence,
    "prediction": prediction,
    "gt": gt,
    "boundary": boundary,
    "boundary_on": False,
    "n_bins": 10
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
Calculate_MCE Ran
ran on a synthesized input
uncbiag/LTS/code/probability_measure_CamVid.py
code served (permissive licence) · get_code("ebb35b9431b0b649")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create small numpy arrays that match the function's usage patterns
height, width = 4, 3

# confidence: probability map, should be float between 0 and 1
confidence = np.random.rand(height, width).astype(np.float32)

# prediction: prediction map of segmentation (class indices)
prediction = np.random.randint(0, 5, size=(height, width)).astype(np.int32)

# gt: ground-truth segmentation map
# Use values 0-4 for valid regions and 100 for invalid/unlabeled regions
gt = np.random.randint(0, 5, size=(height, width)).astype(np.int32)
gt[0, 0] = 100  # Mark at least one pixel as invalid (gt == 100)

# boundary: boundary areas of segmentation map
# Non-zero values indicate boundary pixels
boundary = np.random.randint(0, 2, size=(height, width)).astype(np.int32)

# boundary_on: whether to evaluate on boundary or all region
boundary_on = False

# n_bins: number of bins for calibration error
n_bins = 10

ARGS = {
    "confidence": confidence,
    "prediction": prediction,
    "gt": gt,
    "boundary": boundary,
    "boundary_on": boundary_on,
    "n_bins": n_bins,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
Calculate_SCE Ran
ran on a synthesized input
uncbiag/LTS/code/probability_measure_CamVid.py
code served (permissive licence) · get_code("d817a87ee7f89f5d")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Set random seed for reproducibility
np.random.seed(0)

# Create small dummy segmentation maps
# confidence: probability map, values between 0 and 1
confidence = np.random.rand(4, 3).astype(np.float32)

# prediction: predicted segmentation labels (0-11 for 12 classes)
prediction = np.random.randint(0, 12, size=(4, 3)).astype(np.int32)

# gt: ground-truth segmentation (0-11 for valid labels, 100 for ignored regions)
gt = np.random.randint(0, 12, size=(4, 3)).astype(np.int32)
# Ensure some valid regions (gt != 100)
gt[0, 0] = 5

# boundary: boundary areas (0 for non-boundary, non-zero for boundary)
boundary = np.random.randint(0, 2, size=(4, 3)).astype(np.int32)

# boundary_on: whether to evaluate on boundary or all region
boundary_on = False

# n_bins: number of bins for calibration evaluation
n_bins = 10

ARGS = {
    'confidence': confidence,
    'prediction': prediction,
    'gt': gt,
    'boundary': boundary,
    'boundary_on': boundary_on,
    'n_bins': n_bins
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block

Also present in other papers' code

Of 3 samples lifted from this paper's linked repositories, 3 carry a function identity and 0 do not. Of the 3 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 3 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

For semantic segmentation, label probabilities are often uncalibrated as they are typically only the by-product of a segmentation task. Intersection over Union (IoU) and Dice score are often used as criteria for segmentation success, while metrics related to label probabilities are not often explored. However, probability calibration approaches have been studied, which match probability outputs with experimentally observed errors. These approaches mainly focus on classification tasks, but not on semantic segmentation. Thus, we propose a learning-based calibration method that focuses on multi-label semantic segmentation. Specifically, we adopt a convolutional neural network to predict local temperature values for probability calibration. One advantage of our approach is that it does not change prediction accuracy, hence allowing for calibration as a post-processing step. Experiments on the COCO, CamVid, and LPBA40 datasets demonstrate improved calibration performance for a range of different metrics. We also demonstrate the good performance of our method for multi-atlas brain segmentation from magnetic resonance images.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2008.05105")
get_code_for_paper("2008.05105")
have("2008.05105")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2008.05105.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2008.05105.svg)](https://syntology.ai/paper/2008.05105)

Connect an agent — have() is free.