SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2410.08397 · 2024

VoxelPrompt: A Vision Agent for End-to-End Medical Image Analysis

arXiv · PDF · Open in the Atlas

Code that ran

3 of 6 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 0 have no contract verdict; 3 failures of our instrument; 3 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/2410.08397.json · badge: /api/ran/2410.08397.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
dalcalab/voxel canonical (classifier label)
classifier basis: census
3 ran · 3 no recorded run
FunctionStatusWhere it lives
cached_transferable Ran
ran; our contract draft was wrong, not the code
dalcalab/voxel/voxel/caching.py
code served (permissive licence) · get_code("bcc1c38f67fb9d91")
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 property. The image it ran in was not recorded. If your copy returns the same type 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 torch
import functools

torch.manual_seed(0)

class TestClass:
    """A minimal class with the _transferable_property_cache attribute."""
    
    def __init__(self):
        self._transferable_property_cache = {}
    
    @functools.wraps(lambda self: None)
    def cached_transferable(func: callable) -> property:
        """
        Decorator that converts a method into a read-only cached property stored
        in `self._transferable_property_cache`. A transferable property is one that
        can be carried over to a new instance of the same class.
        """

        @functools.wraps(func)
        def wrapper(self):
            if func.__name__ not in self._transferable_property_cache:
                self._transferable_property_cache[func.__name__] = func(self)
            return self._transferable_property_cache[func.__name__]
        wrapper = property(wrapper)
        return wrapper


def example_method(self):
    """A simple method to decorate."""
    return torch.tensor([1.0, 2.0, 3.0, 4.0])


ARGS = {
    "func": example_method
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type property — a property output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a property output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
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
expand_slicing Ran
ran; our contract draft was wrong, not the code
dalcalab/voxel/voxel/slicing.py
code served (permissive licence) · get_code("848288d538f87c21")
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 tuple, ndim 1, 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 torch
import numpy as np

torch.manual_seed(0)

# Analyze the function signature and body:
# - slicing: tuple - can contain slice objects, integers, Ellipsis, etc.
# - length: int - target slicing length (positive integer)

# Test case 1: Simple slicing without ellipsis, needs expansion
slicing_1 = (slice(None), 0)
length_1 = 4

# Test case 2: Slicing with ellipsis
slicing_2 = (0, Ellipsis, slice(None))
length_2 = 4

# Test case 3: Exact length match, no expansion needed
slicing_3 = (slice(None), slice(None), slice(None))
length_3 = 3

# Test case 4: Simple ellipsis at the end
slicing_4 = (slice(None), Ellipsis)
length_4 = 3

# Choose a valid test case
ARGS = {
    "slicing": (slice(None), 0),
    "length": 4
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
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
slicing_to_coordinates Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
dalcalab/voxel/voxel/slicing.py
code served (permissive licence) · get_code("6c74ad7a5798845e")
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 tuple. The image it ran in was not recorded. If your copy returns the same type 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 torch

torch.manual_seed(0)

# Test case 1: Simple slicing with slice objects
slicing_1 = (slice(0, 5), slice(1, 10))
shape_1 = (10, 20)

# Test case 2: Slicing with integers
slicing_2 = (2, slice(None))
shape_2 = (5, 8)

# Test case 3: Slicing with ellipsis
slicing_3 = (slice(0, 3), Ellipsis, slice(2, 5))
shape_3 = (4, 6, 8, 12)

# Test case 4: Slicing with steps
slicing_4 = (slice(0, 10, 2), slice(1, 15, 3))
shape_4 = (15, 20)

# Test case 5: Mixed slicing with negative indices
slicing_5 = (slice(-5, None), slice(None, -1))
shape_5 = (12, 16)

# Use the first test case for the ARGS dict
ARGS = {
    "slicing": slicing_1,
    "shape": shape_1
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
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
cached Not yet run
unverified (harvested, no recorded run)
dalcalab/voxel/voxel/caching.py
code served (permissive licence) · get_code("09b012888a989fc7")
How it ran

no invocation recorded.

coordinates_to_slicing Not yet run
unverified (harvested, no recorded run)
dalcalab/voxel/voxel/slicing.py
code served (permissive licence) · get_code("c8fba5cb8d7c05a8")
How it ran

no invocation recorded.

gaussian_kernel_1d Not yet run
unverified (harvested, no recorded run)
dalcalab/voxel/voxel/filters.py
code served (permissive licence) · get_code("c88f6db2a98446ed")
How it ran

no invocation recorded.

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 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 6 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

We present VoxelPrompt, an end-to-end image analysis agent that tackles free-form radiological tasks. Given any number of volumetric medical images and a natural language prompt, VoxelPrompt integrates a language model that generates executable code to invoke a jointly-trained, adaptable vision network. This code further carries out analytical steps to address practical quantitative aims, such as measuring the growth of a tumor across visits. The pipelines generated by VoxelPrompt automate analyses that currently require practitioners to painstakingly combine multiple specialized vision and statistical tools. We evaluate VoxelPrompt using diverse neuroimaging tasks and show that it can delineate hundreds of anatomical and pathological features, measure complex morphological properties, and perform open-language analysis of lesion characteristics. VoxelPrompt performs these objectives with an accuracy similar to that of specialist single-task models for image analysis, while facilitating a broad range of compositional biomedical workflows.

For agents

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

get_harvested_code_for_paper("2410.08397")
get_code_for_paper("2410.08397")
have("2410.08397")

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

curl https://syntology.ai/api/ran/2410.08397.json

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

[![Syntology run record](https://syntology.ai/api/ran/2410.08397.svg)](https://syntology.ai/paper/2410.08397)

Connect an agent — have() is free.