SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2102.08884 · ICCV · 2021

MOSAICOS: A Simple and Effective Use of Object-Centric Images for Long-Tailed Object Detection

Boqing Gong, Wei-Lun Chao, Cheng Zhang, Tai-Yu Pan, Yandong Li, Hexiang Hu, Dong Xuan, Soravit Changpinyo

arXiv · PDF · Open in the Atlas

Code that ran

2 of 2 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; 2 failures of our instrument; 0 no recorded run), measured 2026-11-22; 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/2102.08884.json · badge: /api/ran/2102.08884.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
czhang0528/MosaicOS marked official in Papers with Code
Papers with Code archive
2 ran · 0 no recorded run
FunctionStatusWhere it lives
concat_min Ran
ran; our contract draft was wrong, not the code
czhang0528/MosaicOS/mosaic.py
code served (permissive licence) · get_code("a93d736b86562be2")
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 cv2
import numpy as np

# Create sample images with different sizes
# Image 1: 100x80 (height x width) with 3 channels
img1 = np.random.randint(0, 256, (100, 80, 3), dtype=np.uint8)

# Image 2: 120x60 (height x width) with 3 channels
img2 = np.random.randint(0, 256, (120, 60, 3), dtype=np.uint8)

# Image 3: 90x70 (height x width) with 3 channels
img3 = np.random.randint(0, 256, (90, 70, 3), dtype=np.uint8)

# Create sample bounding boxes for each image
# Each box is [x1, y1, x2, y2]
boxes1 = np.array([[10, 20, 30, 40], [50, 60, 70, 80]], dtype=np.float32)
boxes2 = np.array([[5, 15, 25, 35]], dtype=np.float32)
boxes3 = np.array([], dtype=np.float32).reshape(0, 4)

ARGS = {
    'images': [img1, img2, img3],
    'boxes': [boxes1, boxes2, boxes3],
    'by': 'horizontal',
    'interpolation': cv2.INTER_CUBIC
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, 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
synthesize Ran
ran; our contract draft was wrong, not the code
czhang0528/MosaicOS/mosaic.py
code served (permissive licence) · get_code("24b34df3519b1235")
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) 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 numpy as np
import cv2

# Create small test images with distinct dimensions
# Image 1: 100x80 (height x width)
image1 = np.random.randint(0, 256, (100, 80, 3), dtype=np.uint8)

# Image 2: 120x90 (height x width)
image2 = np.random.randint(0, 256, (120, 90, 3), dtype=np.uint8)

# Image 3: 110x85 (height x width)
image3 = np.random.randint(0, 256, (110, 85, 3), dtype=np.uint8)

# Image 4: 95x75 (height x width)
image4 = np.random.randint(0, 256, (95, 75, 3), dtype=np.uint8)

# Create bounding boxes for each image
# Format: [x1, y1, x2, y2] for each box
boxes1 = np.array([[10, 10, 30, 40], [50, 20, 70, 60]], dtype=np.float32)
boxes2 = np.array([[15, 15, 45, 50]], dtype=np.float32)
boxes3 = np.array([], dtype=np.float32).reshape(0, 4)  # No boxes for image 3
boxes4 = np.array([[5, 5, 20, 25], [40, 30, 65, 55]], dtype=np.float32)

# Organize into 2x2 grid
images = [image1, image2, image3, image4]
boxes = [boxes1, boxes2, boxes3, boxes4]

ARGS = {
    'nrow': 2,
    'ncol': 2,
    'images': images,
    'boxes': boxes,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
cv2, 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

Also present in other papers' code

Of 2 samples lifted from this paper's linked repositories, 2 carry a function identity and 0 do not. Of the 2 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 2 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

Many objects do not appear frequently enough in complex scenes (e.g., certain handbags in living rooms) for training an accurate object detector, but are often found frequently by themselves (e.g., in product images). Yet, these object-centric images are not effectively leveraged for improving object detection in scene-centric images. In this paper, we propose Mosaic of Object-centric images as Scene-centric images (MOSAICOS), a simple and novel framework that is surprisingly effective at tackling the challenges of long-tailed object detection. Keys to our approach are three-fold: (i) pseudo scene-centric image construction from object-centric images for mitigating domain differences, (ii) high-quality bounding box imputation using the object-centric images' class labels, and (iii) a multi-stage training procedure. On LVIS object detection (and instance segmentation), MOSAICOS leads to a massive 60% (and 23%) relative improvement in average precision for rare object categories. We also show that our framework can be compatibly used with other existing approaches to achieve even further gains. Our pre-trained models are publicly available at https:// github.com/czhang0528/MosaicOS/.

For agents

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

get_harvested_code_for_paper("2102.08884")
get_code_for_paper("2102.08884")
have("2102.08884")

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

curl https://syntology.ai/api/ran/2102.08884.json

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

[![Syntology run record](https://syntology.ai/api/ran/2102.08884.svg)](https://syntology.ai/paper/2102.08884)

Connect an agent — have() is free.