Face morphing versus face averaging: Vulnerability and detection.

DISCLAIMER

Distribution of FRGC-Morph has been suspended.

Database Description

FERET-Morphs is a dataset of morphed faces selected from the publicly available FERET dataset (https://www.nist.gov/itl/products-and-services/color-feret-database). We created the database by selecting similar looking pairs of people, and made 3 types of morphs for each pair using the following morphing tools:

OpenCV (https://www.learnopencv.com/face-morph-using-opencv-cpp-python/) FaceMorpher (https://github.com/yaopang/FaceMorpher/tree/master/facemorpher) StyleGAN 2 (https://github.com/NVlabs/stylegan2)

Instructions

This dataset is planned for vulnerability analysis experiments in the context of face recognition. Therefore, it is intended to be used in conjunction with the original FERET dataset. The copy_original_feret.py file included with this dataset helps with preparing the file structure so this folder may easily be used for such experiments.

$ python copy_original_feret.py /path/to/original/feret/folder

Once completed the directory's structure should be as given below:

+-- feret | +-- morph_facemorpher | +-- morph_opencv | +-- morph_stylegan | +-- raw | +-- protocols | +-- copy_original_feret.py | +-- feret_selection.csv | +-- README.txt

Protocols

The vulnerability analysis can be conducted in two ways, using:

morphed images as references (reverse-protocol)

morphed images as probes (scores-protocol)

The protocols for both types of experiments are provided in the protocols folder, each of which contains the file lists of detailing the exact images used as references (for_models.lst) and as probes (for_probes.lst) for each morphing tool.

The data is not split into subsets, rather a single set is provided both as the development (dev) and evaluation set (eval) in order to be easily used by a toolkit such as bob (https://www.idiap.ch/software/bob).

References

Any publication (eg. conference paper, journal article, technical report, book chapter, etc) resulting from the usage of FERET_Morphs must cite the following papers:

@INPROCEEDINGS{9746477, author = {Sarkar, Eklavya and Korshunov, Pavel and Colbois, Laurent and Marcel, Sébastien}, booktitle = {ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)}, title = {Are GAN-based morphs threatening face recognition?}, year={2022}, pages={2959-2963}, url={https://doi.org/10.1109/ICASSP43922.2022.9746477} doi={10.1109/ICASSP43922.2022.9746477} }

@article{Sarkar2020, title={Vulnerability Analysis of Face Morphing Attacks from Landmarks and Generative Adversarial Networks}, author={Eklavya Sarkar and Pavel Korshunov and Laurent Colbois and S\'{e}bastien Marcel}, year={2020}, month=oct, journal={arXiv preprint}, url={https://arxiv.org/abs/2012.05344} }

Any publication (eg. conference paper, journal article, technical report, book chapter, etc) resulting from the usage of FERET and subsequently FERET_Morphs must also cite the following paper:

P. Jonathon Phillips, Harry Wechsler, Jeffery Huang, and Patrick J. Rauss, The feret database and evaluation procedure for face-recognition algorithms. Image and Vision Computing, Vol. 16, no.5, pp. 295–306, 1998.

DISCLAIMER

The `preprocess.py` script included in this dataset is no longer necessary, and thus should NOT be run.

Database Description

FRLL-Morphs is a dataset of morphed faces based on images selected from the publicly available Face Research London Lab dataset.
We created 4 types of morphs for each pre-selected pair of images using the following morphing tools:

• OpenCV
• FaceMorpher
• StyleGAN 2
• WebMorpher

Instructions

This dataset is planned for vulnerability analysis experiments in the context of face recognition.
Therefore, it is intended to be used in conjunction with the original Face Research London Lab dataset.

To prepare this folder's file structure so it may easily be used for such experiments:

1. Download and extract only the `neutral_front` and `smiling_front` datasets from the Face London Research Dataset.
2. Place them in a new facelab_london/raw folder.
3. Rename them simply as `neutral` and `smiling` respectively.
4. Remove the `.tem` files from the `neutral` folder if not specifically required for any experiments as these could clash with other operations.

Once completed the directory's structure should be as given below:

+-- facelab_london
| +-- morph_amsl
| +-- morph_facemorpher
| +-- morph_opencv
| +-- morph_stylegan
| +-- morph_webmorph
| +-- raw
| +-- protocols
| +-- preprocess.py
| +-- README.txt

Protocols

The vulnerability analysis can be conducted in two ways, using:

• morphed images as references (`reverse-protocol`)
• morphed images as probes (`scores-protocol`)

The protocols for both types of experiments are provided in the `protocols` folder, each of which contains the file lists of detailing the exact images used as references (`for_models.lst`) and as probes (`for_probes.lst`) for each morphing tool.

The data is split into two sets, development (`dev`) and evaluation (`eval`), in order to be easily used by a toolkit such as bob (https://www.idiap.ch/software/bob). The split is made such as that no original identities used to make the morphed images overlap with one another, make the two sets completely independent of one another.

References

Any publication (eg. conference paper, journal article, technical report, book chapter, etc) resulting from the usage of FRLL_Morphs must cite the following papers:

@INPROCEEDINGS{9746477,
author = {Sarkar, Eklavya and Korshunov, Pavel and Colbois, Laurent and Marcel, Sébastien},
booktitle = {ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
title = {Are GAN-based morphs threatening face recognition?},
year={2022},
pages={2959-2963},
url={https://doi.org/10.1109/ICASSP43922.2022.9746477}
doi={10.1109/ICASSP43922.2022.9746477}
}

@article{Sarkar2020,
title={Vulnerability Analysis of Face Morphing Attacks from Landmarks and Generative Adversarial Networks},
author={Eklavya Sarkar and Pavel Korshunov and Laurent Colbois and S\'{e}bastien Marcel},
year={2020},
month=oct,
journal={arXiv preprint},
url={https://arxiv.org/abs/2012.05344}
}

Any publication (eg. conference paper, journal article, technical report, book chapter, etc) resulting from the usage of Face Research London Lab must cite the following source:

@misc{debruine_jones_2017,
title={Face Research Lab London Set},
url={https://figshare.com/articles/dataset/Face_Research_Lab_London_Set/5047666/3},
DOI={10.6084/m9.figshare.5047666.v3},
publisher={figshare},
author={DeBruine, Lisa and Jones, Benedict},
year={2017},
month={May}
}

Any publication (eg. conference paper, journal article, technical report, book chapter, etc) resulting from the usage of Advanced Multimedia Security Lab’s (AMSL) Face Morph Image dataset must cite the following source:

@article{https://doi.org/10.1049/iet-bmt.2017.0147,
author={Neubert, Tom and Makrushin, Andrey and Hildebrandt, Mario and Kraetzer, Christian and Dittmann, Jana},
title={Extended StirTrace benchmarking of biometric and forensic qualities of morphed face images},
journal={IET Biometrics},
volume={7},
number={4},
pages={325-332},
doi={https://doi.org/10.1049/iet-bmt.2017.0147},
url={https://ietresearch.onlinelibrary.wiley.com/doi/abs/10.1049/iet-bmt.2017.0147},
eprint={https://ietresearch.onlinelibrary.wiley.com/doi/pdf/10.1049/iet-bmt.2017.0147},
year={2018}
}

Face Morphing Attacks pose a threat to the security of identity documents, especially with respect to a subsequent access control process, because they allow both involved individuals to use the same document. Several algorithms are currently being developed to detect Morphing Attacks, often requiring large data sets of morphed face images for training. In the present study, face embeddings are used for two different purposes: first, to pre-select images for the subsequent large-scale generation of Morphing Attacks, and second, to detect potential Morphing Attacks. Previous studies have demonstrated the power of embeddings in both use cases. However, we aim to build on these studies by adding the more powerful MagFace model to both use cases, and by performing comprehensive analyses of the role of embeddings in pre-selection and attack detection in terms of the vulnerability of face recognition systems and attack detection algorithms. In particular, we use recent developments to assess the attack potential, but also investigate the influence of morphing algorithms. For the first objective, an algorithm is developed that pairs individuals based on the similarity of their face embeddings. Different state-of-the-art face recognition systems are used to extract embeddings in order to pre-select the face images and different morphing algorithms are used to fuse the face images. The attack potential of the differently generated morphed face images will be quantified to compare the usability of the embeddings for automatically generating a large number of successful Morphing Attacks. For the second objective, we compare the performance of the embeddings of two state-of-the-art face recognition systems with respect to their ability to detect morphed face images. Our results demonstrate that ArcFace and MagFace provide valuable face embeddings for image pre-selection. Various open-source and commercial-off-the-shelf face recognition systems are vulnerable to the generated Morphing Attacks, and their vulnerability increases when image pre-selection is based on embeddings compared to random pairing. In particular, landmark-based closed-source morphing algorithms generate attacks that pose a high risk to any tested face recognition system. Remarkably, more accurate face recognition systems show a higher vulnerability to Morphing Attacks. Among the systems tested, commercial-off-the-shelf systems were the most vulnerable to Morphing Attacks. In addition, MagFace embeddings stand out as a robust alternative for detecting morphed face images compared to the previously used ArcFace embeddings. The results endorse the benefits of face embeddings for more effective image pre-selection for face morphing and for more accurate detection of morphed face images, as demonstrated by extensive analysis of various designed attacks. The MagFace model is a powerful alternative to the often-used ArcFace model in detecting attacks and can increase performance depending on the use case. It also highlights the usability of embeddings to generate large-scale morphed face databases for various purposes, such as training Morphing Attack Detection algorithms as a countermeasure against attacks.

MORPH Consists of:- - Images divided into Train, Validation, Test Sets(80-10-10) - Labels of Age and Gender are provided in csv files in Index**(The age is scaled according to the starting and ending ages in dataset)**. - Images are Cropped and Aligned using DLIB and cv2 library in python

This data set contains the Satellite CPC morphing technique (CMORPH). For more information please see CPC Morphing Technique (CMORPH).

Matching unfamiliar faces is known to be difficult, and this can give an opportunity to those engaged in identity fraud. Here we examine a relatively new form of fraud, the use of photo-ID containing a graphical morph between two faces. Such a document may look sufficiently like two people to serve as ID for both. We present two experiments with human viewers, and a third with a smartphone face recognition system. In Experiment 1, viewers were asked to match pairs of faces, without being warned that one of the pair could be a morph. They very commonly accepted a morphed face as a match. However, in Experiment 2, following very short training on morph detection, their acceptance rate fell considerably. Nevertheless, there remained large individual differences in people’s ability to detect a morph. In Experiment 3 we show that a smartphone makes errors at a similar rate to ‘trained’ human viewers—i.e. accepting a small number of morphs as genuine ID. We discuss these results in reference to the use of face photos for security.

Reconfigurable morphing surfaces provide new opportunities for advanced human-machine interfaces and bio-inspired robotics. Morphing into arbitrary surfaces on demand requires a device with a sufficiently large number of actuators and an inverse control strategy that can calculate the actuator stimulation necessary to achieve a target surface. The programmability of a morphing surface can be improved by increasing the number of independent actuators, but this increases the complexity of the control system. Thus, developing compact and efficient control interfaces and control algorithms is a crucial knowledge gap for the adoption of morphing surfaces in broad applications. In this work, we describe a passively addressed robotic morphing surface (PARMS) composed of matrix-arranged ionic actuators. To reduce the complexity of the physical control interface, we introduce passive matrix addressing. Matrix addressing allows the control of N2 independent actuators using only 2N control inputs,...

Included here are data to support the publication: Ciqun Xu, Saba Firouznia, Charl F. J. Faul, Majid Taghavi, and Jonathan Rossiter, Charge-induced Morphing Gels for Bioinspired Actuation, Advanced Functional Materials (2025): e17447 https://doi.org/10.1002/adfm.202517447 Complete download (zip, 2.5 MiB)

All of the final data files required to run the Python and R code associated with: "Stability shifts in gliding flight: Hawks morph from an unstable to stable state when navigating a gap".

Data files necessary to reproduce figures.

The GPM Ground Validation NOAA CPC Morphing Technique (CMORPH) IFloodS dataset consists of global precipitation analyses data produced by the NOAA Climate Prediction Center (CPC). The Iowa Flood Studies (IFloodS) campaign was a ground measurement campaign that took place in eastern Iowa from May 1 to June 15, 2013. The goals of the campaign were to collect detailed measurements of precipitation at the Earth'ssurface using ground instruments and advanced weather radars and, simultaneously, collect data from satellites passing overhead. The CPC morphing technique uses precipitation estimates from low orbiter satellite microwave observations to produce global precipitation analyses at a high temporal and spatial resolution. Data has been selected for the Iowa Flood Studies (IFloodS) field campaign which took place from April 1, 2013 to June 30, 2013. The dataset includes both the near real-time raw data and bias corrected data from NOAA in binary and netCDF format.

This dataset contains the bias-corrected CPC MORPHing technique (CMORPH) global precipitation analyses, version 1, and is obtained from the NOAA Climate Data Record.

The following description is from the NOAA Climate Data Record CMORPH dataset page:

This data set is for the bias-corrected, reprocessed CPC Morphing technique (CMORPH) high-resolution global satellite precipitation estimates. The CMORPH satellite precipitation estimates are created in two steps. First, the purely satellite-based global fields of precipitation are constructed through integrating Level 2 retrievals of instantaneous precipitation rates from all available passive microwave measurements aboard low earth orbiting platforms. Bias in these integrated satellite precipitation estimates is then removed through comparison against CPC daily gauge analysis over land and adjustment against the Global Precipitation Climatology Project (GPCP) merged analysis of pentad precipitation over ocean. The bias corrected CMORPH satellite precipitation estimates are created on an 8 km by 8 km grid over the global domain from 60 degrees S to 60 degrees N and in a 30-minute interval from January 1, 1998. Due to the delay of some input data sets, this formal version (Version 1) bias corrected CMORPH is produced manually once a month at a latency of 3-4 months.

For the CDR production, the bias corrected CMORPH generated at its native resolution of 8 km by 8 km / 30-minute is upscaled to form THREE sets of data files of different time/space resolution for improved user experience:

a) the full-resolution CMORPH data
Output variable: precipitation rate in mm/hour
spatial resolution: 8 km by 8km (at equator)
spatial coverage: global (60S-60N)
temporal resolution: 30min
data period: January 1, 1998 to the present

b) Hourly CMORPH
Output variable:...

This dataset contains the predicted prices of the asset Mecha Morphing over the next 16 years. This data is calculated initially using a default 5 percent annual growth rate, and after page load, it features a sliding scale component where the user can then further adjust the growth rate to their own positive or negative projections. The maximum positive adjustable growth rate is 100 percent, and the minimum adjustable growth rate is -100 percent.

The GPM Ground Validation NOAA CPC Morphing Technique (CMORPH) IPHEx dataset consists of global precipitation analyses data produced by the NOAA Climate Prediction Center (CPC) during the Global Precipitation Mission (GPM) Integrated Precipitation and Hydrology Experiment (IPHEx) field campaign in North Carolina. The goal of IPHEx was to evaluate the accuracy of satellite precipitation measurements and use the collected data for hydrology models in the region. The CPC morphing technique uses precipitation estimates from low orbiter satellite microwave observations to produce global precipitation analyses at a high temporal and spatial resolution. CMORPH data has been selected from May 1, 2014 through June 14, 2014, during the IPHEx field campaign. These data files are available in raw binary and netCDF-4 file format.

This data set contains all of the input files that are required to use the "run_script.R" to obtain the final output results. This includes the body and feather morphological measurements as well as the aligned optitrack outputs. This includes "Assumption_Tracking.xlsx" that details every measurement assumption that was made within the study and "VerificationData.xlsx" that details the torso densities and estimates of the head only CG position as discussed in the supplemental methods.

Bird-inspired reflexive morphing enables rudderless flight

https://doi.org/10.5061/dryad.zkh1893h5

Description of the data and file structure

PigeonBotII.STEP

PigeonBotII.STEP is a STEP AP214 3D CAD file of PigeonBot II.

Software: Prepared in Solidworks 2023. Compatible with most CAD software. A free web-based alternative to view this file is Autodesk Viewer.

PigeonBot_II_datasupplement.xfl

PigeonBot_II_datasupplement.xfl is a xflr5 project file for simulating open-loop dynamic stability of PigeonBot II's three-by-three wing and tail morph combinations.

Software: Prepared in xflr5 v6.61. xflr5 is free.

WindTunnelTrials.mat

WindTunnelTrials.mat is a MATLAB data file containing PigeonBot II virtual flight data in the wind tunnel.

Software: Prepared in MATLAB R2018b. Free alternatives are HDF5 for Python or [GNU Octave](https://octa...

This dataset contains the predicted prices of the asset Morph AI over the next 16 years. This data is calculated initially using a default 5 percent annual growth rate, and after page load, it features a sliding scale component where the user can then further adjust the growth rate to their own positive or negative projections. The maximum positive adjustable growth rate is 100 percent, and the minimum adjustable growth rate is -100 percent.

Data and models for journal publication The role of geometry in the rapid snap-through of hummingbird beak-inspired morphing structures

This dataset contains the predicted prices of the asset MORPH over the next 16 years. This data is calculated initially using a default 5 percent annual growth rate, and after page load, it features a sliding scale component where the user can then further adjust the growth rate to their own positive or negative projections. The maximum positive adjustable growth rate is 100 percent, and the minimum adjustable growth rate is -100 percent.