=Paper=
{{Paper
|id=Vol-3910/aics2024_p23
|storemode=property
|title=An Investigation into the Application of Anomaly Detection and the Meijering Filter in the eKYC Process to Detect Recaptured Identity Documents
|pdfUrl=https://ceur-ws.org/Vol-3910/aics2024_p23.pdf
|volume=Vol-3910
|authors=John Magee,Stephen Sheridan,Christina Thorpe
|dblpUrl=https://dblp.org/rec/conf/aics/MageeST24
}}
==An Investigation into the Application of Anomaly Detection and the Meijering Filter in the eKYC Process to Detect Recaptured Identity Documents==
https://ceur-ws.org/Vol-3910/aics2024_p23.pdf
An Investigation Into the Application of Anomaly
Detection and the Meijering Filter in the eKYC Process to
Detect Recaptured Identity Documents
John Magee1,† , Stephen Sheridan1 and Christina Thorpe1
1
School of Informatics and Cybersecurity, Technological University Dublin, Dublin, Ireland
Abstract
As financial institutions move to offer more digital based services, the need for secure and accurate eKYC services
increases. Identity documents submitted during the eKYC process are fundamental to establish the identity of
customers. The ability of bad actors to modify identity documents using digital imaging software opens these
eKYC services to new threats, resulting in identity theft and reputational damage. In this exploratory research we
apply auto-encoder/decoder anomaly detection to the domain of recaptured identity document detection, using
the Meijering filter as the feature extraction mechanism.
Keywords
image processing, anomaly detection, auto-encoder/decoder, Meijering filter, document fraud, recaptured identity
documents,
1. Introduction
Traditional Know Your Customer (KYC) solutions require banking customers to present themselves at a
physical building, a slow and costly process [1]. eKYC solutions allow customers to sign up for new
services remotely, greatly benefiting banking institutions. It also opens up new attack vectors from bad
actors looking to commit fraud or identity theft. Regulations 1 , such as Anti-Money Laundering (AML,
including Politically Exposed Persons (PEPs) screening 2 ) and Countering the Financing of Terrorism
(CFT) regulations impose strict identity proofing standards. The ability to prove the authenticity of an
identity document during a remote account opening is fundamental in establishing the true identity of
a new customer and to ensure compliance with such regulations.
The advanced nature of computer graphic software means it is easy to create/modify images of
identity documents. Recent research by Magee et al. [2, 3, 4] demonstrated that biomedical imaging
filters (the Meijering filter, the Sato filter and Frangi filter) can be used to help identify recaptured
identity documents. Their best results were obtained using the Meijering filter and a Random Forest
classifier (APCER=7.5%, BPCER=6.6%)[4]. However, it is not possible to know what new attack vectors
bad actors will use when attempting to by-pass security mechanisms designed to detect fraudulent
documents, thus exposing a limitation on the use of a classification based systems in this domain.
Machine learning techniques are divided into two categories, classification and regression. Classifica-
tion techniques are designed around the premise that a dataset is divided into discrete instances, referred
to as ’classes’. Training techniques ensure the model learns features to distinguish between classes.
Regression techniques predict a value across a continuous range [5]. These techniques require datasets
that are labeled. It is not always possible to have such datasets available when researching sensitive
domains, such as identity document fraud. A paradigm known as anomaly detection [6, 7] can be
useful in such circumstances. This technique takes advantage of an understanding of expected data and
AICS’24: 32nd Irish Conference on Artificial Intelligence and Cognitive Science, December 09–10, 2024, Dublin, Ireland
∗
Corresponding author.
†
Primary author.
$ B00149241@mytudublin.ie (J. Magee); stephen.sheridan@mytudublin.ie (S. Sheridan); christina.thorpe@mytudublin.ie
(C. Thorpe)
© 2024 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
1
https://www.centralbank.ie/regulation/anti-money-laundering-and-countering-the-financing-of-terrorism
2
https://www.dataprotection.ie/en/faqs/banking-insurance-sector/what-politically-exposed-persons-pep-screening
CEUR
ceur-ws.org
Workshop ISSN 1613-0073
Proceedings
attempts to identify anomalous patterns [7]. Hilton & Salakhutdinov [8] showed that auto-encoders can
reduce the dimensions of the input data into a more efficient latent space. The decoder regenerates the
latent space into the input format. The difference between the input and reconstruction is leveraged to
detect anomalous events [7]. The auto-encoder/decoder architecture is also fundamental to the modern
development of generative AI [9]. Based on our previous research, we pose the following research
question: "Can anomaly detection using neural network based auto-encoder/decoder architecture be
used to detect recaptured identity documents processed by the Meijering filter?".
The structure of this paper is as follows: Section 2 presents a brief overview of the related work;
Section 3 introduces the experiment design; Section 4 presents our results and discussion; Section 5
presents our conclusion and future work.
2. Related Research
This section introduces a review of the identity document datasets, techniques used to date for document
recapture/forgery detection and the usage of anomaly detection across industries.
2.1. Identity Document Data Sets
The Mobile Identity Document Video (MIDV) is a data set released by Smart Engines 3 . It consists of
captured synthetic identity documents, the purpose is to further advancement in Optical Character
Recognition (OCR) and Machine Readable Zone (MRZ) processing. Three versions of MIDV exist,
MIDV-500[10], MIDV-2019[11] and MIDV-2020[12]. MIDV-500 consists of 500 video clips for 50 identity
document types as well as annotated images in TIFF format. MIDV-2019 extended the original dataset
by including high resolution images captured under different lighting conditions. MIDV-2020 further
expanded the data set, including JPEG images of the documents. As this dataset does not include
recaptured identity documents, it is not applicable for use in this experiment.
Kumar et al. [13] describe an image generation system for synthetic passports, driver’s licenses and
Visa stickers. They have published a data set consisting of 15,000 images, available on Kaggle 4 . Special
character sets and fonts exist in official documents. The goal of this dataset is to further the study in
this field as one of the leading indicators of forged documents is typically font related i.e. the forger
used the wrong fonts.
More recently, Benalcazar et al. [14] published another mechanism for synthetic data generation using
three different approaches, including computer vision algorithms and Generative Adversarial Networks
(GANs). Their dataset includes print/scan recapture and screen recapture images of documents. Using
the Chilean ID Card as a template, they used a range of random facial images (from the FERET database),
signatures, names and dates to generate random instances of ID Cards. They developed the ability to
transfer different types of noise into the images to simulate actual capture from a device e.g. screen
recapture. Their data set is not publicly available as access is controlled, otherwise this would considered
for use in this experiment.
Soares et al. [15] released the Brazilian Identity Document (BID) data set, a synthetic identity
document dataset generated using the Brazilian ID Card as a template. The purpose of this data set is
to advance the domain of document segmentation and Optical Character Recognition (OCR) research.
This is the source dataset used to build the dataset for our experiments as it contains both portrait and
landscape captures and when printed are approximately the same size as the information page on a
passport book. This is published on Kaggle 5 .
Chen et al. [16] published a data augmentation method to assist with the training of algorithms for
forged document detection. Their method was shown to increase the accuracy of deeply trained models
by 6.8%. While this work doesn’t specifically focus on identity documents, some of the techniques are
3
https://smartengines.com/
4
https://www.kaggle.com/turabbajeer
5
https://www.kaggle.com/datasets/johnmageetud/recaptured-identity-documents
applicable to this domain as special documents e.g. certificates, contain complex features specifically
designed to ensure authenticity, and by design to make the documents harder to forge. It is not clear if
this approach can be used in the context of anomaly detection.
2.2. Classification for Document Recapture/Forgery Detection
Berenguel et al. [17] developed a document classification based on texture analysis. Their system
uses a Naïve Bayes classifier based on multivariate Bernoulli distribution to detect if a document was
genuine or counterfeit. The Naïve Bayes classifier inputs are constructed by a sequence of components
that include Principal Component Analysis (PCA) and Linear Support Vector Machines (SVM). They
construct a dataset of crowd sourced recaptures of the Spanish ID Card, capturing both sides of the ID
Card. This is a binary classification network and they report a classification F-score of approximately
98% across each side of the ID Card. Further work by Berenguel et al. [18] was the development of the
Counterfeit Recurrent Comparator (CRC) to identify counterfeit documents. This design takes into
consideration existing research of the human perception system [19]. This network is trained using an
existing dataset of counterfeit bank notes from their own research [20] and they report a AUC of 0.984.
Yang et al. [21] developed a Convolutional Neural Network (CNN) to detect recaptured images. Their
network is a binary classification network, classifying if an image is genuine or if it is recaptured. This
research is not specific to identity documents, but it can be applied to identity document images if
a sufficient dataset was available. Their network is trained using lower quality images, only 512 by
512 pixels in size. Their research reports a classification accuracy of 99.74%. Testing with even lower
sized images reduced this accuracy slightly. However, this approach shows that processing images in
patches is possible alternative to processing whole images which is something we use in the experiment
documented in this paper.
Chen et al. [22] developed a Siamese neural network to detect recaptured documents. A Siamese
network is a complicated neural network design containing of two identical components that find
similarities between inputs. Despite its complexity, in this case it was ultimately configured to be a
binary classification network. They train and test their network using synthetic document data and
achieve 6.92% Attack Presentation Classification Error Rate (APCER) and 8.51% Bonafide Presentation
Classification Error Rate (BPCER).
Research from Magee et al. [2, 3, 4]. has also focused on a classification approach to recaptured
identity document detection. This research utilises the Meijering filter as a feature extraction process
and the Random Forest classification algorithm. They achieved APCER 7.53% and BPCER 6.59% using
an optimised Meijering filter configuration discovered using a grid search approach. These accuracy
results are comprable to the results obtained by Chen et al. using their complex Siamese neural network
architecture.
The papers presented above [18, 20, 21, 22] show a clear trend toward the use of neural network
architectures for classification models. Our own work [2, 3, 4] used traditional machine learning models
for the purposes of classification. All these algorithms depend on the existence of structured data
for training, which is a limitation we have identified. There are no examples of the use of anomaly
detection.
2.3. Anomaly Detection Algorithms
A comparative analysis by Kharitonov et al.[23] shows the most common anomaly detection algorithms
include neural network based auto-encoders, Isolation Forest, K-Nearest Neighbour(KNN). The success
of neural networks in recent years has propelled the use of auto-encoders in this domain.
Autonomous vehicles are a recent literature source for anomaly detection. Rezaei et al. [24] used a
GAN based auto-encoder/decoder architecture for sensor fault detection as well as to protect vehicle
systems from cyber attacks. Their network follows the auto-encoder/decoder architecture where the
encoder is under-complete [25] (the number of neurons in each layer decreases, forcing the network to
learn a smaller dimension latent space representation of the input). The decoder network then uses the
over-complete network architecture to reconstruct the reduced dimensional representation back to that
of the original input. They report their results in terms of accuracy, sensitivity and specificity derived
from the range of anomaly scores. Han et al. [26] developed an anomaly detection system based on
the transformer architecture for GPS spoofing threats. Their system combines image and positional
data to detect anomalies, the images broken down into patches. The image and positional data are
merged and encoded by the transformer. This encoded feature set is the input to a one-class classifier
Multi-Layer Perceptron (MLP), computing the probability that the input is anomalous. Han et al. report
results using the F1 and Area Under the Curve (AUC) metrics but no threshold is defined. Di Biase et
al. [27] present a computer vision based pixel-wise anomaly detection framework. This is specifically
designed to detect anomalous objects in the path of a vehicle (e.g. a person, a dog or another vehicle)
using an image re-synthesis approach. Their encoder is an under-complete architecture based on the
VGG-16 [28] network architecture. Their decoder component is an over-complete architecture but adds
the novelty by concatenating a generated semantic image into each decoder network layer, augmenting
the role of the decoder. Their metrics are based on the false positive rate as a fixed threshold of 95%
true positive rate.
An overview of the application of anomaly detection in industry is presented by Kharitonov et al. [23].
They show traditional machine learning methods, e.g. KNN and Isolation Forest, have a wide application.
They investigate the robustness of each algorithm to predict machine breakdowns based on historical
data. Their experiments show that traditional machine learning algorithms can still outperform neural
network based auto-encoders. Gruber & Heselmann [29] developed a Frangi filter [30] based detection
system to detect defects in transparent materials. The Frangi filter is a biomedical image filter, designed
to enhance vessels in images. Their system uses feature extraction from filtered images and the KNN
algorithm for fault detection. Chen et al. [31] present a system for fibre inspection in an industrial
textile setting. They use an under-complete LeNet-5 [32] CNN based encoder component and a Support
Vector Machine to predict fibre classification. Input images are split into patches for processing. They
measure their solution effectiveness using the precision metric when varying the amount of data used
in the training process. When their system is trained with 80% of the available training data it is able to
outperform existing state of the art.
Medical applications of anomaly detection is using autoencoder/decoder architectures is common in
electrocardiogram (ECG) event monitoring. Verardo et al. [33], Lomoio et al. [34] and Shan et al. [35]
all follow this approach, using the MSE or MAE loss functions.
2.4. Gap Analysis
It is clear that anomaly detection is a widely accepted method in industry and we take inspiration from
the diverse examples of its application across different domains. We note that there are no examples of
anomaly detection being used in the domain of recaptured identity document detection, classification
models[17, 18, 20, 21, 22, 2, 3, 4] are used exclusively. Therefore, we will apply anomaly detection to
the domain of recaptured identity document detection in this experiment.
3. Methodology
3.1. Data set
This research reuses our dataset used in [4] that consists of recaptured images from the BID dataset
[15]. The scope of this research extends the base dataset to include captures from an Android device, a
Motorola G62.
The recaptured dataset consists of two types of recaptured identity documents, 306 screen recaptured
images and 918 printed recaptured documents, captured using two iPhone models (8 & 12) and a Motorola
G62 Android device. Printed recaptured documents are subdivided into printed paper recaptures and
plastic covered printed paper recaptures. The final breakdown of the dataset used in this research is
represented in Table 1. A sample capture of the raw documents and the resultant images after being
processed by the Meijering filter are shown in Figure 1.
Table 1
Type and count of recaptured documents per device.
Source Printer iPhone 8 iPhone 12 M G62
Paper Inkjet 102 102 102
Laser 102 102 102
Plastic covered Inkjet 102 102 102
Laser 102 102 102
Screen Recapture N/A 102 102 102
Figure 1: Shows (A) the raw captured image of an inkjet printout from the Motorola G62 and (B) the resultant
image after being processed using the Meijering filter. (C) the raw captured image of an inkjet printout from the
iPhone8 and (D) the resultant image after being processed using the Meijering filter.
3.2. Network Architecture
The network architecture used in this research is described in Table 2. The encoder component
represents a traditional under-complete network, where the network is forced to learn a lower dimension
representation of the input [25]. The decoder component is the traditional over-complete network that
learns to reconstruct the original input. The network is not symmetrical as the decoder component
contains one more layer than the encoder component. It is well-known that auto-encoder/decoders
always produce a lower resolution reconstruction of the input data, therefore this is included to provide
the decoder with additional weights (degrees of freedom) to better reconstruct the input. As this research
is a proof of concept, a minimal network architecture is used consisting only of dense layers.
Table 2
Describes the individual layers in the encoder-decoder architecture.
Component Layer Input Output Activation
Encoder Input 9176 9176
1 9176 1248 ReLU
2 1248 512 ReLU
Decoder Input 512 ReLU
1 512 768 ReLU
2 768 1248 ReLU
Output 1248 9176 ReLU
3.3. Loss Functions
The most common loss functions used by neural network based auto encoders are the Mean Absolute
Error (MAE), Mean Square Error (MSE) and Cosine Similarity (CS). The MSE [36] is defined in Equation
1 and computes the squared error between the value, y, and the predicted value f (x), and computes
the mean of these values across the number of samples n.
n
1X
M SE = (y i − f (xi )) 2 (1)
n
i=1
The MAE [37] loss function is defined in Equation 2. A variation on the MSE, the absolute error value
is computed, and the mean is taken across the number of samples n.
n
1X
M AE = |y i − f (xi ) | (2)
n
i=1
The CS [38] loss function is defined in Equation 3. This measures the angular distance between the
given and predicted values. This has been shown to be better than metrics like MSE and MAE that
attempt to measure a linear distance. This metric has resulted in exceptional results in the domain of
natural language processing and is integral to the functioning of the transformer architecture [39].
Pn
(y i × f (xi ))
CS = pPn i=1 pPn (3)
i=1 (y i ) × i=1 f (xi )
2 2
3.4. Data Processing
The image processing pipeline used in this research consists of the following steps:
1. Image filtering. Apply the Meijering filter to each of the input images. Input image dimensions
are 1344x848 pixels. The filtered output are gray scale images of dimensions 1244x748 pixels,
saved using the viridis colour space.
2. Image cropping. The images are cropped to a final dimensions of 1240x740 by removing the outer
most pixels.
3. Generating image patches. Each image is divided evenly into 100 patches of dimensions 124x74
pixels.
The final step is to aggregate the individual files containing the intensity pixel values so they can be
used in the training process. This step simply copies individual patch data into CSV files, generating
the training and validation dataset files. The output of this process provides us with:
1. Training dataset. Consists of 27,540 patches of filtered screen recaptured documents.
2. Test dataset. Consists of 3,060 patches, reserved from the training dataset, that is used to test all
models. This test dataset remains static across all models and all training files.
3. Validation dataset. This consists of 122,400 patches of the filtered printed recaptured documents.
This is considered anomalous data for the purposes of this experiment.
3.5. Training the network
The NVIDIA GeForce GTX 1650 Ti GPU used in this research can only accommodate approximately
2GB of data, therefore the data is distributed across 10 separate files. The training procedure uses each
file sequentially, enabling the ability to train in batch mode. The Keras 6 framework, version 2.10.0, is
used to train the models, with the ADAM optimizer. The three popular loss functions in the field of
anomaly detection are used, MSE, MAE and CS. The following safe guards are employed in order to
avoid over fitting the training data: a) we randomise the input data so the algorithm cannot learn from
a predictable input sequence, b) we enable shuffling of training data after each epoch. The models are
trained using four different epoch levels, 50, 100, 150 and 200. We then compare the performance of
models trained for the same number of epochs. No significant increase in performance was observed
past 200 epoches.
6
https://keras.io/
3.6. Testing the Model
The performance of machine learning classifiers typically report results using the confusion matrix and
its constituent metrics i.e., True Positive Rate (TPR), True Negative Rate (TNR), False Positive Rate (FPR)
and the False Negative Rate (FNR) [37]. When measuring the performance of an anomaly detection
model, these metrics cease to be as useful because the model is not aware of classes. The results in
this research are presented using the range based Equal Error Rate (EER) metric [40]. This approach
determines a threshold value at which both error rates (FPR and FNR) are equal. Measuring error rates
is industry standard for biometric and document classification systems as error rates are a measure of
security of the system. This is formalized in the ISO Presentation Attack Detection standard (30107) 7 .
3.6.1. Algorithm to compute EER
The output of the testing procedure is a range of MAE scores, one for each patch, from the test and
validation datasets. These are processed by the algorithm described in Algorithm 1.
Algorithm 1 Algorithm to compute logistics curves to determine EER
Require: len(test_scores) > 0
Require: len(validation_scores) > 0
1: procedure ComputeEER(test_scores, validation_scores)
2: min_threshold ← min(test_scores)
3: max_threshold ← max(validation_scores)
4: threshold_range ← seq(min_threshold, max_threshold, 0.1)
5: index ← 0
6: index ←next_ value
7: while index ≤ len(threshold_range) do
8: threshold_value ←threshold_range[index]
9: f a ← count(test_scores > threshold_value)
10: f n ← count(validation_scores <= threshold_value)
11: save( threshold_value, f a, f n)
12: index ←index + 1
13: end while
14: end procedure
The algorithm is based on a computation of the false classifications at a specific threshold value.
The algorithm is initialised on lines 2 and 3 by obtaining the minimum score from the test dataset
and the maximum score from the validation dataset. These values are used in line 4 to compute a
sequence of threshold values, in increments of 0.1, between the minimum and maximum values using
the seq function in R. The procedure iterates over the threshold values and computes the number of
misclassifications at each threshold (lines 8 & 9). The threshold and the misclassifcation rates are then
saved to file (line 11). A chart plotting the misclassification rates is shown in Figure 2, where the reader
can see the intersection point of the two logistic curves, representing the threshold to provides the
equal error rate.
4. Results and Discussion
The results are shown in Table 3. The model trained using Cosine Similarity produced lower EER
values than the MAE and MSE models. The lowest EER obtained using MAE=9.03% compared to Cosine
Similarity=8.08%, while MSE=9.12%. The lowest mean EER is achieved by the models trained using
Cosine Similarity=39.58% compared to MAE=44.94% and MSE=45.45%. The highest EERs are produced
by dark patches where no real distinguishing information is available.
7
https://www.iso.org/standard/83828.html
Figure 2: Chart showing the EER plotted to show the intercept point of the two curves.
Table 3
EER statistics for models trained using loss functions MAE, MSE and Cosine Similarity
MAE MSE Cosine
Epochs Min Mean Max Min Mean Max Min Mean Max
50 11.39% 47.10% 93.97% 9.16% 46.36% 93.86% 8.57% 41.33% 81.86%
100 9.13% 46.06% 93.35% 9.57% 45.78% 93.49% 8.08% 40.73% 82.48%
150 9.03% 45.55% 93.37% 9.55% 45.64% 93.25% 8.14% 39.71% 83.33%
200 9.13% 44.94% 92.95% 9.12% 45.45% 93.07% 8.08% 39.58% 82.76%
All models show a large EER range, due to the relevant performance of different patches. We see
that each loss function responds differently to the input patch. Models trained using the MAE and
MSE loss function produce models that result in significantly lower EER for vertical document features
while models trained using the Cosine Similarity produce lower EER on horizontal document features.
Figures 3 and 4 show patches with vertical and horizontal features, overlaid with the mean patch EER,
highlighting for the reader the difference in performance. Further research is required to understand
if this is a true response to features intrinsic to the documents or if this is just co-incidence. Visual
inspection shows that patches tend to be reconstructed to a higher degree of accuracy for test samples
(see Figure 5) compared to validation samples (see Figure 6). Validation patches reconstructed by the
models also shows signs of hallucination, as shown in Figures 7 and 8.
Figure 3: Comparing the performance of MSE and Cosine Similarity models based on vertical features. See the
patch EER for highlighting the difference in performance.
The loss function used to train a model is critical. While MSE and MAE are extensively used, as
shown in our literature review in Section 2, they do not have the sensitivity to accurately represent the
error in image reconstruction scenarios like the one we undertook in this research. Similarly, the CS
loss function was capable of training a model that produced lower mean EER, but still far too high for
practical use.
Figure 4: Comparing the performance of MSE and Cosine Similarity models based on horizontal features. See
the patch EER for highlighting the difference in performance.
Figure 5: Manual reconstruction of patches from a screen recaptured document, image recaptured using an
iPhone8.
Figure 6: Manual reconstruction of patches from a paper recaptured document from an inkjet printed, image
recaptured using an iPhone8.
5. Conclusion
We posed the following research question: "Can anomaly detection using neural network based
autoencoder-decoder architecture be used to detect recaptured identity documents processed by the
Meijering filter?". Based on our results, the answer is a qualified "Yes". We see the MSE and CS loss
learn different patterns, CS performed better with horizontal features while MSE performed better with
vertical features. These features correspond to parts of the document that produce visually different
impulse responses from the Meijering filter, thus validating this approach.
Care is needed to ensure we don’t infer too much about how real identity document images will
behave under the same experimental conditions, but we have demonstrated the potential of this
approach. Another aspect that requires caution is that this research evaluated the EER at a patch
level, not considering the document as a whole entity with a single metric. Despite its limitations,
this proof-of-concept shows the potential for the use of anomaly detection when detecting recaptured
identity documents. While this paper specifically targeted the eKYC process for financial institutions,
the practical use of this technology can extent to any sector where establishment of identity is important,
Figure 7: Visualising the reconstruction of sample patches for models trained using MSE and Cosine Similarity
loss functions.
Figure 8: Examples of generated patches that demonstrate hallucinations by the network. Samples from the
test dataset recreate the actual patch whereas the same patches from the validation set appear to make up
features that are not present in the actual input.
such as the government and health sectors.
5.1. Future Work
Based on the limitations of the loss functions, investigations into different loss functions is a priority.
Research by Yan et al. made use of a loss function to combine MSE and Cosine Similarity. Other
loss functions typically associated with generative models, MinMax, Wasserstein and Diffusion loss
functions, are obvious next steps. Experimenting with deep model architectures like ResNet-50 and
Diffusion models are on our list of next steps. In this research we trained our own model from the
beginning as this was simply a proof of concept, however, it is well established that transfer learning is
a valid strategy to improve the training of neural networks by leveraging the weights from models that
are already trained. Expanding our dataset is critical to ensure the results of this research generalise
across other mobile devices, and to capture realistic input data to represent genuine identity document
captured images. Investigating the use of an error metric for the entire document rather than individual
patches is also in the scope for future work.
References
[1] R. Soltani, U. Trang Nguyen, A. An, A new approach to client onboarding using self-sovereign
identity and distributed ledger, in: 2018 IEEE International Conference on Internet of Things
(iThings), IEEE, 2018, pp. 1129–1136. doi:10.1109/Cybermatics_2018.2018.00205.
[2] J. Magee, S. Sheridan, C. Thorpe, An investigation into the application of the meijering filter for
document recapture detection, in: International Conference on Cloud and Big Data Computing
(ICCBDC2023), 2023. doi:10.12720/jait.15.1.132-137.
[3] J. Magee, S. Sheridan, C. Thorpe, Classifying recaptured identity documents using the biomedical
meijering and sato algorithms, in: APWG.EU Technical Summit and Researchers Sync-Up, APWG,
2023. URL: https://ceur-ws.org/Vol-3631/paper2.pdf.
[4] J. Magee, S. Sheridan, C. Thorpe, Optimization of biomedical imaging filters for use in recaptured
identity document classification, in: Cybersecurity Ireland (ICCBDC2023), 2023. doi:10.1109/
Cyber-RCI59474.2023.10671521.
[5] C. M. Bishop, Pattern Recognition and Machine Learning, Springer, New York, 2009.
[6] C. C. Aggarwal, Neural Networks and Deep Learning, Springer, New York, 2018.
[7] S. K. Adari, S. Alla, Beginning Anomaly Detection Using Python-Based Deep Learning, Springer,
2024.
[8] G. E. Hinton, R. R. Salakhutdinov, Reducing the dimensionality of data with neural networks
(2006) 504–507. doi:10.1126/science.1127647.
[9] J. Ho, A. Jain, P. Abbeel, Denoising diffusion probabilistic models (2020). doi:10.48550/ARXIV.
2006.11239.
[10] V. V. Arlazarov, K. Bulatov, T. Chernov, V. L. Arlazarov, MIDV-500: A dataset for identity
documents analysis and recognition on mobile devices in video stream (2019). doi:10.18287/
2412-6179-2019-43-5-818-824.
[11] K. Bulatov, D. Matalov, V. V. Arlazarov, MIDV-2019: Challenges of the modern mobile-based
document OCR (2020) 64. doi:10.1117/12.2558438. arXiv:1910.04009.
[12] K. Bulatov, E. Emelianova, D. Tropin, N. Skoryukina, Y. Chernyshova, A. Sheshkus, S. Usilin,
Z. Ming, J.-C. Burie, M. M. Luqman, V. V. Arlazarov, MIDV-2020: A comprehensive benchmark
dataset for identity document analysis (2022). doi:10.18287/2412-6179-CO-1006.
[13] T. Kumar, M. Turab, S. Talpur, R. Brennan, M. Bendechache, Forged character detection datasets:
Passports, driving licences and visa stickers (2022). doi:10.5121/ijaia.2022.13202.
[14] D. Benalcazar, J. E. Tapia, S. Gonzalez, C. Busch, Synthetic id card image generation for improving
presentation attack detection, IEEE Transactions on Information Forensics and Security (2023)
1814–1824. doi:10.1109/TIFS.2023.3255585.
[15] D. S. Soares, R. B. Das Neves Junior, B. L. D. Bezerra, BID dataset: a challenge dataset for document
processing tasks, in: Anais Estendidos da Conference on Graphics, Patterns and Images (SIBRAPI
Estendido 2020), Sociedade Brasileira de Computação, 2020.
[16] C. Chen, B. Li, R. Cai, J. Zeng, J. Huang, Distortion model-based spectral augmentation for
generalized recaptured document detection (2024). doi:10.1109/TIFS.2023.3333548.
[17] A. Berenguel, O. R. Terrades, J. Llados, C. Canero, E-counterfeit: A mobile-server platform for
document counterfeit detection, in: 2017 14th IAPR International Conference on Document
Analysis and Recognition (ICDAR), IEEE, 2017, pp. 15–20. doi:10.1109/ICDAR.2017.390.
[18] A. Berenguel, O. Ramos Terrades, J. Llados Canet, C. Canero Morales, Recurrent comparator
with attention models to detect counterfeit documents, in: 2019 International Conference on
Document Analysis and Recognition (ICDAR), IEEE, 2019, pp. 1332–1337. doi:10.1109/ICDAR.
2019.00215.
[19] P. Shyam, S. Gupta, A. Dukkipati, Attentive recurrent comparators, in: D. Precup, Y. W. Teh (Eds.),
Proceedings of the 34th International Conference on Machine Learning, volume 70 of Proceedings
of Machine Learning Research, PMLR, 2017, pp. 3173–3181.
[20] A. B. Centeno, O. R. Terrades, J. L. i. Canet, C. C. Morales, Evaluation of texture descriptors for
validation of counterfeit documents, in: 2017 14th IAPR International Conference on Document
Analysis and Recognition (ICDAR), IEEE, 2017, pp. 1237–1242. doi:10.1109/ICDAR.2017.204.
[21] P. Yang, R. Ni, Y. Zhao, Recapture image forensics based on laplacian convolutional neural networks,
in: Digital Forensics and Watermarking, volume 10082, Springer International Publishing, 2017,
pp. 119–128. doi:10.1007/978-3-319-53465-7_9.
[22] C. Chen, S. Zhang, F. Lan, J. Huang, Domain-agnostic document authentication against practical
recapturing attacks (2022) 2890–2905. doi:10.1109/TIFS.2022.3197054.
[23] A. Kharitonov, A. Nahhas, M. Pohl, K. Turowski, Comparative analysis of machine learning models
for anomaly detection in manufacturing (2022). doi:10.1016/j.procs.2022.01.330.
[24] S. Rezaei, N. Masoud, A. Khojandi, GAAD: GAN-enabled autoencoder for real-time sensor anomaly
detection and recovery in autonomous driving (2024). doi:10.1109/JSEN.2024.3361460.
[25] Y. B. Ian Goodfellow, A. Courville, Deep Learning, MIT Press, 2016.
[26] M. Du, Z. Chen, C. Liu, R. Oak, D. Song, Lifelong anomaly detection through unlearning, in:
Proceedings of the 2019 ACM SIGSAC Conference on Computer and Communications Security,
ACM, 2019, pp. 1283–1297. doi:10.1145/3319535.3363226.
[27] G. Di Biase, H. Blum, R. Siegwart, C. Cadena, Pixel-wise anomaly detection in complex driving
scenes, in: 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE,
2021, pp. 16913–16922. doi:10.1109/CVPR46437.2021.01664.
[28] S. Liu, W. Deng, Very deep convolutional neural network based image classification using small
training sample size, in: 2015 3rd IAPR Asian Conference on Pattern Recognition (ACPR), 2015,
pp. 730–734. doi:10.1109/ACPR.2015.7486599.
[29] D. P. Gruber, M. Haselmann, Inspection of transparent objects with varying light scattering using
a frangi filter (2021) 27. doi:10.3390/jimaging7020027.
[30] A. F. Frangi, W. J. Niessen, K. L. Vincken, M. A. Viergever, Multiscale vessel enhancement filtering,
in: W. M. Wells, A. Colchester, S. Delp (Eds.), Medical Image Computing and Computer-Assisted
Intervention — MICCAI’98, volume 1496, Springer Berlin Heidelberg, 1998, pp. 130–137.
[31] Z. Chen, Y. Xiu, Y. Zheng, X. Wang, Q. Wang, D. Guo, Y. Wan, A weakly supervised learning
pipeline for profiled fibre inspection (2024) 772–784. doi:10.1049/ipr2.12984.
[32] Y. Lecun, L. Bottou, Y. Bengio, P. Haffner, Gradient-based learning applied to document recognition,
Proceedings of the IEEE 86 (1998) 2278–2324. doi:10.1109/5.726791.
[33] G. Verardo, M. Boman, S. Bruchfeld, M. Chiesa, S. Koch, G. Q. Maguire, D. Kostic, FMM-
head: Enhancing autoencoder-based ECG anomaly detection with prior knowledge (2023).
doi:10.48550/ARXIV.2310.05848.
[34] U. Lomoio, P. Vizza, R. Giancotti, G. Tradigo, S. Petrolo, S. Flesca, P. Hiram Guzzi, P. Veltri, AUTAN-
ECG: An AUToencoder bAsed system for anomaly detectioN in ECG signals, 2023. doi:10.36227/
techrxiv.24638856.v1.
[35] L. Shan, Y. Li, H. Jiang, P. Zhou, J. Niu, R. Liu, Y. Wei, J. Peng, H. Yu, X. Sha, S. Chang, Abnormal
ECG detection based on an adversarial autoencoder (2022) 961724. doi:10.3389/fphys.2022.
961724.
[36] G. James, D. Witten, T. Hastie, R. Tibshirani, An Introduction to Statistical Learning: with Applica-
tions in R, Springer Texts in Statistics, Springer US, 2021. doi:10.1007/978-1-0716-1418-1.
[37] J. D. Kelleher, B. MacNamee, A. D’Arcy, Fundamentals of Machine Leanring for Predictive Data
Analytics, MIT Press US, 2015.
[38] S. Yan, H. Shao, Y. Xiao, B. Liu, J. Wan, Hybrid robust convolutional autoencoder for unsupervised
anomaly detection of machine tools under noises (2023) 102441. doi:10.1016/j.rcim.2022.
102441.
[39] A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. Kaiser, I. Polosukhin,
Attention is all you need (2017). arXiv:1706.03762.
[40] L. Zhang, X. Wang, E. Cooper, N. Evans, J. Yamagishi, Range-Based Equal Error Rate for Spoof
Localization, in: Proc. INTERSPEECH 2023, 2023, pp. 3212–3216. doi:10.21437/Interspeech.
2023-1214.