SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2001.08499 · 2020

A Large Scale Event-based Detection Dataset for Automotive

arXiv · PDF · Open in the Atlas

Code that ran

6 of 7 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 6 have no contract verdict; 0 failures of our instrument; 1 no recorded run), measured 2026-11-21; 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/2001.08499.json · badge: /api/ran/2001.08499.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
prophesee-ai/prophesee-automotive-dataset-toolbox community (classifier label)
classifier basis not recorded on the link
6 ran · 1 no recorded run
FunctionStatusWhere it lives
count_events Ran
ran on a synthesized input
prophesee-ai/prophesee-automotive-dataset-toolbox/src/io/dat_events_tools.py
code served (permissive licence) · get_code("6588176504cc10b8")
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 uint8, ndim 0, dtype family int, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape 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 os
import sys
import tempfile
import numpy as np

# Create a temporary dat file with a valid format
temp_dir = tempfile.gettempdir()
temp_file = os.path.join(temp_dir, 'test_events.dat')

# Write a minimal valid dat file
with open(temp_file, 'wb') as f:
    # Write header lines
    f.write(b'% Date 2024 01 15\n')
    f.write(b'% Height 480\n')
    f.write(b'% Width 640\n')
    
    # Write event type and size bytes
    f.write(np.uint8(0).tobytes())  # event type
    f.write(np.uint8(8).tobytes())  # event size in bytes
    
    # Write some event data (8 bytes per event)
    num_events = 5
    for i in range(num_events):
        event_data = np.array([i, i*2], dtype=[('t', 'u4'), ('_', 'i4')])
        f.write(event_data.tobytes())

ARGS = {
    'filename': temp_file
}
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 uint8, ndim 0, dtype family int, shape () — shape () as a later run recorded it
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
filter_boxes Ran
ran on a synthesized input
prophesee-ai/prophesee-automotive-dataset-toolbox/src/io/box_filtering.py
code served (permissive licence) · get_code("03ea6bf76fae4681")
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 ndarray, ndim 1, dtype family float, shape (2,) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape 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 a structured numpy array with the required dtype for boxes
BBOX_DTYPE = np.dtype([
    ('t', np.int64),
    ('x', np.float32),
    ('y', np.float32),
    ('w', np.float32),
    ('h', np.float32),
    ('class_id', np.int32),
    ('track_id', np.int32),
    ('class_confidence', np.float32)
])

# Create sample box data with 5 boxes
boxes = np.array([
    (100000, 10.0, 20.0, 100.0, 80.0, 1, 1, 0.95),
    (600000, 50.0, 60.0, 150.0, 120.0, 2, 2, 0.87),
    (700000, 15.0, 25.0, 30.0, 25.0, 1, 3, 0.92),
    (800000, 100.0, 150.0, 200.0, 180.0, 3, 4, 0.88),
    (900000, 5.0, 10.0, 10.0, 15.0, 2, 5, 0.79),
], dtype=BBOX_DTYPE)

ARGS = {
    'boxes': boxes,
    'skip_ts': int(500000.0),
    'min_box_diag': 60,
    'min_box_side': 20
}
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 ndarray, ndim 1, dtype family float, shape (2,) — shape (2,) as a later run recorded it
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
load_td_data Ran
ran on a synthesized input
prophesee-ai/prophesee-automotive-dataset-toolbox/src/io/dat_events_tools.py
code served (permissive licence) · get_code("fe7ffc1e33554f5d")
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 ndarray, ndim 1, dtype family float, shape (5,) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape 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 tempfile
import os
import numpy as np

# Create a temporary dat file with proper header and event data
temp_file = tempfile.NamedTemporaryFile(mode='wb', delete=False, suffix='.dat')
temp_filename = temp_file.name

# Write header
header = b'% Date 2024 01 01\n% Height 480\n% Width 640\n'
temp_file.write(header)

# Write event type and size
temp_file.write(np.uint8(0).tobytes())  # ev_type
temp_file.write(np.uint8(8).tobytes())  # ev_size (4 bytes for 't' + 4 bytes for '_')

# Write some event data
# EV_TYPE = [('t', 'u4'), ('_', 'i4')]
# Create 10 events with timestamp and packed x,y,p data
num_events = 10
for i in range(num_events):
    timestamp = np.uint32(i * 1000)
    temp_file.write(timestamp.tobytes())
    
    # Pack x, y, p into a 32-bit integer
    # x: bits 0-13 (14 bits)
    # y: bits 14-27 (14 bits)
    # p: bits 28-31 (4 bits)
    x = np.uint32(i % 100)  # x coordinate
    y = np.uint32((i * 2) % 480)  # y coordinate
    p = np.uint32(i % 2)  # polarity
    packed = x | (y << 14) | (p << 28)
    temp_file.write(np.int32(packed).tobytes())

temp_file.close()

ARGS = {
    'filename': temp_filename,
    'ev_count': 5,
    'ev_start': 0
}
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 ndarray, ndim 1, dtype family float, shape (5,) — shape (5,) as a later run recorded it
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
make_binary_histo Ran
ran on a synthesized input
prophesee-ai/prophesee-automotive-dataset-toolbox/src/visualize/vis_utils.py
code served (permissive licence) · get_code("e75fa5288180d870")
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 ndarray, ndim 3, dtype family int, shape (240, 304, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape 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 a structured numpy array with the required fields: x, y, p
# x and y are coordinates, p is polarity (0 or 1)
events = np.array(
    [(10, 20, 1), (50, 100, 0), (150, 80, 1), (200, 150, 0)],
    dtype=[('x', np.int32), ('y', np.int32), ('p', np.float32)]
)

# Optional: create a pre-existing image
img = None

width = 304
height = 240

ARGS = {
    "events": events,
    "img": img,
    "width": width,
    "height": height,
}
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 ndarray, ndim 3, dtype family int, shape (240, 304, 3) — shape (240, 304, 3) as a later run recorded it
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
parse_header Ran
ran on a synthesized input
prophesee-ai/prophesee-automotive-dataset-toolbox/src/io/dat_events_tools.py
code served (permissive licence) · get_code("fbdd5fe13d8d02d7")
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 os
import sys
import io
import numpy as np

# Create a mock binary file with header content
def create_mock_dat_file():
    # Create an in-memory binary file
    mock_file = io.BytesIO()
    
    # Write header lines (comments starting with "% ")
    header_lines = [
        b"% Date 2024-01-15 10:30:00\n",
        b"% Height 480\n",
        b"% Width 640\n",
    ]
    
    for line in header_lines:
        mock_file.write(line)
    
    # Write event type (1 byte)
    mock_file.write(np.uint8(1).tobytes())
    
    # Write event size (1 byte)
    mock_file.write(np.uint8(8).tobytes())
    
    # Write some dummy event data
    mock_file.write(b"\x00" * 100)
    
    # Reset to beginning
    mock_file.seek(0)
    
    return mock_file

# Create the mock file
mock_dat_file = create_mock_dat_file()

ARGS = {
    'f': mock_dat_file
}
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 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
reformat_boxes Ran
ran on a synthesized input
prophesee-ai/prophesee-automotive-dataset-toolbox/src/io/box_loading.py
code served (permissive licence) · get_code("9260f2e88f5825de")
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 ndarray, ndim 1, dtype family float, shape (2,) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape 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

# Define the old format (with 'ts' and 'confidence' instead of 't' and 'class_confidence')
OLD_BBOX_DTYPE = np.dtype({
    'names': ['ts', 'x', 'y', 'w', 'h', 'class_id', 'track_id', 'confidence'],
    'formats': ['<i8', '<f4', '<f4', '<f4', '<f4', '<u4', '<u4', '<f4'],
    'offsets': [0, 8, 12, 16, 20, 24, 28, 32],
    'itemsize': 40
})

# Create sample boxes with old format
boxes_old = np.zeros((2,), dtype=OLD_BBOX_DTYPE)
boxes_old['ts'] = [100, 101]
boxes_old['x'] = [10.0, 20.0]
boxes_old['y'] = [15.0, 25.0]
boxes_old['w'] = [5.0, 6.0]
boxes_old['h'] = [8.0, 9.0]
boxes_old['class_id'] = [1, 2]
boxes_old['track_id'] = [10, 11]
boxes_old['confidence'] = [0.9, 0.85]

ARGS = {
    'boxes': boxes_old
}
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 ndarray, ndim 1, dtype family float, shape (2,) — shape (2,) as a later run recorded it
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
parse_header Not yet run
unverified (harvested, no recorded run)
prophesee-ai/prophesee-automotive-dataset-toolbox/src/io/npy_events_tools.py
code served (permissive licence) · get_code("dc3efec1267082d4")
How it ran

no invocation recorded.

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 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 7 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 2026-11-21.

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 introduce the first very large detection dataset for event cameras. The dataset is composed of more than 39 hours of automotive recordings acquired with a 304x240 ATIS sensor. It contains open roads and very diverse driving scenarios, ranging from urban, highway, suburbs and countryside scenes, as well as different weather and illumination conditions. Manual bounding box annotations of cars and pedestrians contained in the recordings are also provided at a frequency between 1 and 4Hz, yielding more than 255,000 labels in total. We believe that the availability of a labeled dataset of this size will contribute to major advances in event-based vision tasks such as object detection and classification. We also expect benefits in other tasks such as optical flow, structure from motion and tracking, where for example, the large amount of data can be leveraged by self-supervised learning methods.

For agents

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

get_harvested_code_for_paper("2001.08499")
get_code_for_paper("2001.08499")
have("2001.08499")

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

curl https://syntology.ai/api/ran/2001.08499.json

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

[![Syntology run record](https://syntology.ai/api/ran/2001.08499.svg)](https://syntology.ai/paper/2001.08499)

Connect an agent — have() is free.