SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2106.11958 · NeurIPS · 2021

Prototypical Cross-Attention Networks for Multiple Object Tracking and Segmentation

Martin Danelljan, Hkust, Yu-Wing Tai, Chi-Keung Tang, Lei Ke, Xia Li, Fisher Yu, E Operation

arXiv · PDF · Open in the Atlas

Code that ran

5 of 5 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 5 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/2106.11958.json · badge: /api/ran/2106.11958.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
SysCV/pcan canonical (classifier label)
classifier basis: census
5 ran · 0 no recorded run
FunctionStatusWhere it lives
cal_similarity Ran
ran on a synthesized input
SysCV/pcan/pcan/core/track/similarity.py
code served (permissive licence) · get_code("ecbbdf2760634c6f")
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 Tensor, ndim 2, dtype family float, shape (2, 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 torch
import torch.nn.functional as F

torch.manual_seed(0)

# Create small embedding tensors
# key_embeds: (batch_size=2, embedding_dim=4)
# ref_embeds: (batch_size=3, embedding_dim=4)
key_embeds = torch.randn(2, 4)
ref_embeds = torch.randn(3, 4)

ARGS = {
    'key_embeds': key_embeds,
    'ref_embeds': ref_embeds,
    'method': 'dot_product',
    'temperature': -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 Tensor, ndim 2, dtype family float, shape (2, 3) — shape (2, 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
intersection_over_area Ran
ran on a synthesized input
SysCV/pcan/pcan/core/evaluation/mot.py
code served (permissive licence) · get_code("4cd14952230cdf84")
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 2, dtype family float, shape (2, 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

# Analyze the function:
# - preds: iterable of boxes, each box has indices [0], [1], [2], [3]
#   - [0] = x coordinate, [1] = y coordinate, [2] = width, [3] = height
# - gts: iterable of ground truth boxes, same format
# - Function computes intersection over predicted box area
# - Returns matrix of shape (len(preds), len(gts))

# Create sample bounding boxes in format [x, y, width, height]
preds = np.array([
    [10.0, 20.0, 30.0, 40.0],  # box 1: x=10, y=20, w=30, h=40
    [50.0, 60.0, 25.0, 35.0]   # box 2: x=50, y=60, w=25, h=35
])

gts = np.array([
    [15.0, 25.0, 40.0, 50.0],  # gt box 1: x=15, y=25, w=40, h=50
    [55.0, 65.0, 20.0, 30.0],  # gt box 2: x=55, y=65, w=20, h=30
    [100.0, 100.0, 10.0, 10.0] # gt box 3: x=100, y=100, w=10, h=10
])

ARGS = {
    "preds": preds,
    "gts": gts
}
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 2, dtype family float, shape (2, 3) — shape (2, 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
segtrack2result Ran
ran on a synthesized input
SysCV/pcan/pcan/core/track/transforms.py
code served (permissive licence) · get_code("d58e5f0465880838")
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 defaultdict. 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 numpy as np

torch.manual_seed(0)

# Create small tensors for bboxes (batch_size=2, bbox_dim=4)
bboxes = torch.tensor([
    [10.0, 20.0, 30.0, 40.0],
    [50.0, 60.0, 70.0, 80.0]
], dtype=torch.float32)

# Create small tensors for labels (batch_size=2)
labels = torch.tensor([1, 2], dtype=torch.long)

# Create segmentation masks (list of 2 numpy arrays)
segms = [
    np.array([[1, 0, 1], [0, 1, 0]]),
    np.array([[1, 1], [1, 1]])
]

# Create ids tensor (batch_size=2) with valid ids (> -1)
ids = torch.tensor([5, 10], dtype=torch.long)

ARGS = {
    'bboxes': bboxes,
    'labels': labels,
    'segms': segms,
    'ids': ids
}
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 defaultdict — a defaultdict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a defaultdict 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
track2result Ran
ran on a synthesized input
SysCV/pcan/pcan/core/track/transforms.py
code served (permissive licence) · get_code("8803773e1a105ac1")
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 defaultdict. 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 numpy as np
from collections import defaultdict

torch.manual_seed(0)

# Create small test tensors
# bboxes: should be indexable and have .cpu().numpy() method
bboxes = torch.tensor([
    [10.0, 20.0, 30.0, 40.0],
    [50.0, 60.0, 70.0, 80.0],
    [15.0, 25.0, 35.0, 45.0]
], dtype=torch.float32)

# labels: should be indexable and have .cpu().numpy() method
labels = torch.tensor([1, 2, 0], dtype=torch.int64)

# segms: list of segmentation masks (can be anything, just needs to be indexable)
segms = [
    np.array([[1, 0, 1], [0, 1, 0]]),
    np.array([[1, 1], [1, 1]]),
    np.array([[0, 0, 0, 0]])
]

# ids: tensor with values > -1 for valid entries, and some -1 for invalid
ids = torch.tensor([100, -1, 101], dtype=torch.int64)

ARGS = {
    "bboxes": bboxes,
    "labels": labels,
    "segms": segms,
    "ids": ids
}
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 defaultdict — a defaultdict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a defaultdict 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
xyxy2xywh Ran
ran on a synthesized input
SysCV/pcan/pcan/core/evaluation/mot.py
code served (permissive licence) · get_code("08df0b7800bb9cc2")
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 list, ndim 1, dtype family int; 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

torch.manual_seed(0)

# The function xyxy2xywh takes a bbox parameter and expects it to support
# indexing with [0], [1], [2], [3] and subtraction operations.
# A simple list or tuple works perfectly for this.

# bbox format: [x1, y1, x2, y2] where (x1,y1) is top-left and (x2,y2) is bottom-right
# The function converts to [x, y, width, height] format

bbox = [10, 20, 50, 80]

ARGS = {
    "bbox": bbox
}
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 list, ndim 1, dtype family int — a list 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

Also present in other papers' code

Of 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 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 5 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

Multiple object tracking and segmentation requires detecting, tracking, and segmenting objects belonging to a set of given classes. Most approaches only exploit the temporal dimension to address the association problem, while relying on single frame predictions for the segmentation mask itself. We propose Prototypical Cross-Attention Network (PCAN), capable of leveraging rich spatio-temporal information for online multiple object tracking and segmentation. PCAN first distills a space-time memory into a set of prototypes and then employs cross-attention to retrieve rich information from the past frames. To segment each object, PCAN adopts a prototypical appearance module to learn a set of contrastive foreground and background prototypes, which are then propagated over time. Extensive experiments demonstrate that PCAN outperforms current video instance tracking and segmentation competition winners on both Youtube-VIS and BDD100K datasets, and shows efficacy to both one-stage and two-stage segmentation frameworks.

For agents

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

get_harvested_code_for_paper("2106.11958")
get_code_for_paper("2106.11958")
have("2106.11958")

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

curl https://syntology.ai/api/ran/2106.11958.json

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

[![Syntology run record](https://syntology.ai/api/ran/2106.11958.svg)](https://syntology.ai/paper/2106.11958)

Connect an agent — have() is free.