SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2002.11088 · AAAI · 2020

Model Watermarking for Image Processing Networks

Jie Zhang, Dongdong Chen, Weiming Zhang, Nenghai Yu, Jing Liao, Wenbo Zhou, Han Fang, Hao Cui

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 (2 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; 0 failures of our instrument; 0 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/2002.11088.json · badge: /api/ran/2002.11088.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
ZJZAC/Deep-Model-Watermarking community
Papers with Code archive
2 ran · 0 no recorded run
FunctionStatusWhere it lives
HidingRes Ran
ran on a synthesized input · constructed an object; no result was computed
ZJZAC/Deep-Model-Watermarking/Initial stage/models/HidingRes.py
code served (permissive licence) · get_code("e1c61c8e39b44dce")
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 an instance of its own class, HidingRes: the run constructed the object and did not call it, so no result was computed. 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 torch.nn as nn
import torch.nn.functional as F

torch.manual_seed(0)

class ResidualBlock(nn.Module):

    def __init__(self, channel_num, dilation=1, group=1):
        super(ResidualBlock, self).__init__()
        self.conv1 = nn.Conv2d(channel_num, channel_num, 3, 1, padding=dilation, dilation=dilation, groups=group, bias=False)
        self.norm1 = nn.InstanceNorm2d(channel_num, affine=True)
        self.conv2 = nn.Conv2d(channel_num, channel_num, 3, 1, padding=dilation, dilation=dilation, groups=group, bias=False)
        self.norm2 = nn.InstanceNorm2d(channel_num, affine=True)

    def forward(self, x):
        y = F.relu(self.norm1(self.conv1(x)))
        y = self.norm2(self.conv2(y))
        return F.relu(x + y)

class HidingRes(nn.Module):

    def __init__(self, in_c=4, out_c=3, only_residual=False, requires_grad=True):
        super(HidingRes, self).__init__()
        self.conv1 = nn.Conv2d(in_c, 128, 3, 1, 1, bias=False)
        self.norm1 = nn.InstanceNorm2d(128, affine=True)
        self.conv2 = nn.Conv2d(128, 128, 3, 1, 1, bias=False)
        self.norm2 = nn.InstanceNorm2d(128, affine=True)
        self.conv3 = nn.Conv2d(128, 128, 3, 2, 1, bias=False)
        self.norm3 = nn.InstanceNorm2d(128, affine=True)
        self.res1 = ResidualBlock(128, dilation=2)
        self.res2 = ResidualBlock(128, dilation=2)
        self.res3 = ResidualBlock(128, dilation=2)
        self.res4 = ResidualBlock(128, dilation=2)
        self.res5 = ResidualBlock(128, dilation=4)
        self.res6 = ResidualBlock(128, dilation=4)
        self.res7 = ResidualBlock(128, dilation=4)
        self.res8 = ResidualBlock(128, dilation=4)
        self.res9 = ResidualBlock(128, dilation=1)
        self.deconv3 = nn.ConvTranspose2d(128, 128, 4, 2, 1)
        self.norm4 = nn.InstanceNorm2d(128, affine=True)
        self.deconv2 = nn.Conv2d(128, 128, 3, 1, 1)
        self.norm5 = nn.InstanceNorm2d(128, affine=True)
        self.deconv1 = nn.Conv2d(128, out_c, 1)
        self.only_residual = only_residual
        if not requires_grad:
            for param in self.parameters():
                param.requires_grad = False

    def forward(self, x):
        y = F.relu(self.norm1(self.conv1(x)))
        y = F.relu(self.norm2(self.conv2(y)))
        y = F.relu(self.norm3(self.conv3(y)))
        y = self.res1(y)
        y = self.res2(y)
        y = self.res3(y)
        y = self.res4(y)
        y = self.res5(y)
        y = self.res6(y)
        y = self.res7(y)
        y = self.res8(y)
        y = self.res9(y)
        y = F.relu(self.norm4(self.deconv3(y)))
        y = F.relu(self.norm5(self.deconv2(y)))
        if self.only_residual:
            y = self.deconv1(y)
        else:
            y = F.relu(self.deconv1(y))
        return y

ARGS = {
    "in_c": 4,
    "out_c": 3,
    "only_residual": False,
    "requires_grad": True
}
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 HidingRes — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a HidingRes 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
ResidualBlock Ran
ran on a synthesized input · constructed an object; no result was computed
ZJZAC/Deep-Model-Watermarking/Initial stage/models/HidingRes.py
code served (permissive licence) · get_code("4c34226f3cf4c833")
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 an instance of its own class, ResidualBlock: the run constructed the object and did not call it, so no result was computed. 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 torch.nn as nn

torch.manual_seed(0)

# ResidualBlock.__init__ takes: self, channel_num, dilation, group
# ResidualBlock.forward takes: self, x
# The forward method expects x to be a tensor that:
#   - can be passed through Conv2d with channel_num input channels
#   - can be added to the output of conv2 (which has channel_num output channels)
# So x should have shape (batch, channel_num, height, width)

channel_num = 4
dilation = 1
group = 1

# Create input tensor: (batch=2, channels=4, height=8, width=8)
x = torch.randn(2, channel_num, 8, 8)

ARGS = {
    "channel_num": channel_num,
    "dilation": dilation,
    "group": group,
}
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 ResidualBlock — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a ResidualBlock 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

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 2026-11-21.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Deep learning has achieved tremendous success in numerous industrial applications. As training a good model often needs massive high-quality data and computation resources, the learned models often have significant business values. However, these valuable deep models are exposed to a huge risk of infringements. For example, if the attacker has the full information of one target model including the network structure and weights, the model can be easily finetuned on new datasets. Even if the attacker can only access the output of the target model, he/she can still train another similar surrogate model by generating a large scale of input-output training pairs. How to protect the intellectual property of deep models is a very important but seriously under-researched problem. There are a few recent attempts at classification network protection only. In this paper, we propose the first model watermarking framework for protecting image processing models. To achieve this goal, we leverage the spatial invisible watermarking mechanism. Specifically, given a black-box target model, a unified and invisible watermark is hidden into its outputs, which can be regarded as a special task-agnostic barrier. In this way, when the attacker trains one surrogate model by using the input-output pairs of the target model, the hidden watermark will be learned and extracted afterward. To enable watermarks from binary bits to high-resolution images, both traditional and deep spatial invisible watermarking mechanism are considered. Experiments demonstrate the robustness of the proposed watermarking mechanism, which can resist surrogate models learned with different network structures and objective functions. Besides deep models, the proposed method is also easy to be extended to protect data and traditional image processing algorithms.

For agents

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

get_harvested_code_for_paper("2002.11088")
get_code_for_paper("2002.11088")
have("2002.11088")

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

curl https://syntology.ai/api/ran/2002.11088.json

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

[![Syntology run record](https://syntology.ai/api/ran/2002.11088.svg)](https://syntology.ai/paper/2002.11088)

Connect an agent — have() is free.