SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1803.06799 · 2018

Revisiting RCNN: On Awakening the Classification Power of Faster RCNN

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 2 have no contract verdict; 3 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/1803.06799.json · badge: /api/ran/1803.06799.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
SHI-Labs/Decoupled-Classification-Refinement canonical (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
bowenc0221/Decoupled-Classification-Refinement community (classifier label)
classifier basis not recorded on the link
1 ran · 0 no recorded run
makefile/DCR community (classifier label)
classifier basis not recorded on the link
1 ran · 0 no recorded run
FunctionStatusWhere it lives
get_rcnn_names Ran
ran on a synthesized input
bowenc0221/Decoupled-Classification-Refinement/faster_rcnn/core/metric.py
pointer only (licence: MIT); the stored copy get_code serves, from msracver/Flow-Guided-Feature-Aggregation, is pointer only (licence: MIT) · get_code("bbf8005ce03c08cf")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from msracver/Flow-Guided-Feature-Aggregation, and MIT, as GitHub detects it on that repository, is in the permissive set, but a licence decision keeps the body a pointer: its licence file names an outside party, so it may cover vendored code rather than the repository owner's files (recorded as pointer:licence_names_third_party_corporation). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_rcnn_names Ran
ran on a synthesized input
makefile/DCR/faster_rcnn_dcr/core/metric.py
code served (permissive licence) · get_code("d556ecc93005b402")
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 2, 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

class Config:
    def __init__(self):
        self.TRAIN = TrainConfig()

class TrainConfig:
    def __init__(self):
        self.ENABLE_OHEM = False
        self.END2END = False

torch.manual_seed(0)

cfg = Config()

ARGS = {
    'cfg': cfg
}
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, ndim 2, 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
sample_rcnn Ran
ran; our contract draft was wrong, not the code
SHI-Labs/Decoupled-Classification-Refinement/faster_rcnn_dcr/core/dcr.py
code served (permissive licence) · get_code("015b5733d3e0b8a6")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 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 numpy.random as npr

# Create a config object with the required TRAIN attributes
class Config:
    class TRAIN:
        FG_THRESH = 0.5
        BG_THRESH_HI = 0.5
        BG_THRESH_LO = 0.1

cfg = Config()

# Create roi_rec dictionary with max_classes and max_overlaps
# Let's create 20 ROIs with varying overlaps
num_rois = 20
roi_rec = {
    'max_classes': np.array([0, 1, 2, 0, 1, 3, 0, 2, 1, 0, 1, 2, 3, 0, 1, 2, 0, 1, 2, 3]),
    'max_overlaps': np.array([0.8, 0.7, 0.3, 0.2, 0.6, 0.9, 0.15, 0.4, 0.55, 0.05, 
                              0.75, 0.25, 0.35, 0.85, 0.45, 0.2, 0.65, 0.3, 0.5, 0.1])
}

# Set sample_per_img to be less than the number of ROIs
sample_per_img = 12

ARGS = {
    'roi_rec': roi_rec,
    'cfg': cfg,
    'sample_per_img': sample_per_img
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
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
sample_rois_hard_fg Ran
ran; our contract draft was wrong, not the code
SHI-Labs/Decoupled-Classification-Refinement/faster_rcnn_dcr/core/dcr.py
code served (permissive licence) · get_code("20ad0fa4e4fec335")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 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 numpy.random as npr

# Create a simple config object
class DCRConfig:
    def __init__(self):
        self.hard_fp_score = 0.7
        self.hard_fg_score = 0.5

class TrainConfig:
    def __init__(self):
        self.FG_THRESH = 0.5

class Config:
    def __init__(self):
        self.DCR = DCRConfig()
        self.train = TrainConfig()

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

# Create roi_rec dictionary with all required fields
num_rois = 20

roi_rec = {
    'pred_classes': npr.randint(0, 5, size=num_rois),
    'scores': npr.rand(num_rois),
    'max_classes': npr.randint(0, 5, size=num_rois),
    'max_overlaps': npr.rand(num_rois)
}

# Make some overlaps exactly 1.0 to simulate ground truth
roi_rec['max_overlaps'][0] = 1.0
roi_rec['max_overlaps'][1] = 1.0
roi_rec['max_overlaps'][2] = 1.0

# Make some overlaps in the FG range
roi_rec['max_overlaps'][5] = 0.6
roi_rec['max_overlaps'][6] = 0.7

# Make some scores high to trigger hard FP conditions
roi_rec['scores'][10] = 0.8
roi_rec['scores'][11] = 0.75

cfg = Config()
sample_per_img = 10

ARGS = {
    'roi_rec': roi_rec,
    'cfg': cfg,
    'sample_per_img': sample_per_img
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
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
sample_rois_random Ran
ran; our contract draft was wrong, not the code
SHI-Labs/Decoupled-Classification-Refinement/faster_rcnn_dcr/core/dcr.py
code served (permissive licence) · get_code("6a2e623b98652915")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 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 numpy.random as npr

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

# Create roi_rec dictionary with required keys
# The function accesses roi_rec['max_classes'] and roi_rec['max_overlaps']
num_rois = 10

roi_rec = {
    'max_classes': np.array([1, 2, 0, 3, 1, 2, 0, 1, 3, 2]),  # class labels for each ROI
    'max_overlaps': np.array([0.8, 0.9, 0.3, 0.7, 0.85, 0.6, 0.4, 0.75, 0.65, 0.55])  # overlap scores
}

# Create a simple config dict (cfg is passed but not used in the function)
cfg = {}

# sample_per_img: number of samples to draw
sample_per_img = 5

ARGS = {
    'roi_rec': roi_rec,
    'cfg': cfg,
    'sample_per_img': sample_per_img
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
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

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, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 1 later) and 4 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.

FunctionAlso present inEarliest-datedOthers named
get_rcnn_names 4 other papers
3 dated earlier, 1 later
1611.07715 (2016-11)
Deep Feature Flow for Video Recognition
1703.06211 (2017-03)
Deformable Convolutional Networks
1703.10025 (2017-03)
Flow-Guided Feature Aggregation for Video Object Detection
and 1 more not named here

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

Recent region-based object detectors are usually built with separate classification and localization branches on top of shared feature extraction networks. In this paper, we analyze failure cases of state-of-the-art detectors and observe that most hard false positives result from classification instead of localization. We conjecture that: (1) Shared feature representation is not optimal due to the mismatched goals of feature learning for classification and localization; (2) multi-task learning helps, yet optimization of the multi-task loss may result in sub-optimal for individual tasks; (3) large receptive field for different scales leads to redundant context information for small objects.We demonstrate the potential of detector classification power by a simple, effective, and widely-applicable Decoupled Classification Refinement (DCR) network. DCR samples hard false positives from the base classifier in Faster RCNN and trains a RCNN-styled strong classifier. Experiments show new state-of-the-art results on PASCAL VOC and COCO without any bells and whistles.

For agents

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

get_harvested_code_for_paper("1803.06799")
get_code_for_paper("1803.06799")
have("1803.06799")

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

curl https://syntology.ai/api/ran/1803.06799.json

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

[![Syntology run record](https://syntology.ai/api/ran/1803.06799.svg)](https://syntology.ai/paper/1803.06799)

Connect an agent — have() is free.