Multimedia Forensics
Multimedia Forensics includes a set of scientific techniques recently proposed for the analysis of multimedia signals (audio, videos, images) in order to recover probative evidences from them; in particular, such technologies aim to reveal the history of digital contents, such as identifying the acquisition device that produced the data, validating the integrity of the contents and retrieving information from multimedia signals.
AI-generated images detection
Defense solutions against disinformation attacks based on Diffusion Models
Diffusion models are generative models inspired by non-equilibrium thermodynamics. They work by defining a Markov chain of diffusion steps to slowly add random noise to the data and then learn to reverse the diffusion process to build desired data samples from the noise. Unlike other generative models such as VAE or flow models, diffusion models are learned with a fixed procedure and the latent variable has a high dimensionality. These models have proven to be extraordinarily good at generating highly realistic images. A famous example of these models is OpenAI Dall-E 2 (https://openai.com/dall-e-2/), which is a system that allows generating images based on a text description provided by the user. A popular example is "an avocado-shaped armchair". For those of you who don’t know what they are (give a look at this page).
With our research, we seek to study these models to create defense solutions against disinformation attacks based on diffusion models. Is it possible to check if an image has been generated with a diffusion model?
L. Papa, L. Faiella, L. Corvitto, L. Maiano, I. Amerini, “On the use of Stable Diffusion for creating realistic faces: from generation to detection”, 11th International Workshop on Biometrics and Forensics IWBF 2023, April 2023
DeepFake Detection
Modern AI-based technologies have provided easy-to-use tools to create extremely realistic manipulated videos. Such synthetic videos, named Deep Fakes, may constitute a serious threat to attack the reputation of public subjects or to address the general opinion on a certain event. According to this, being able to individuate this kind of fake information becomes fundamental. In this research project, new forensic techniques able to discern between fake and original video sequences are designed.
Frame-based detector are designed together with methodology that take in account temporal structure of the video, audio and multimodal analysis.
Taiba Majid, Irene Amerini, "Deepfakes Audio Detection leveraging audio spectrogram and Convolutional Neural Networks", ICIAP 2023
L. Maiano, L. Papa, K. Vocaj and I. Amerini, “DepthFake: a depth-based strategy for detecting Deepfake videos”, AI4MFDD workshop ICPR 2022.
I. Amerini, M. Conti, P. Giacomazzi, L. Pajola, “PRaNA: PRNU-based Technique to Tell Real and Deepfake Videos Apart”, IJCNN at IEEE WCCI 2022
I. Amerini, A. Anagnostopoulos, L. Maiano and L. R. Celsi (2021), "Deep Learning for Multimedia Forensics", Foundations and Trends® in Computer Graphics and Vision.
Amerini I., Maiano L., Caldelli R., Galteri L., Del Bimbo A., “DeepFake Cracker: a novel tool for deepfake video detection.” (ICPR 2021). [DEMO]
R. Caldelli, L. Galteri; I. Amerini; A. Del Bimbo, “Optical Flow based CNN for detection of unlearnt deepfake manipulations”, Pattern Recognition Letters 2021 .
I. Amerini, L. Galteri, R.Caldelli, A. Del Bimbo, “Deepfake Video Detection through Optical Flow based CNN”, Human Behaviour and Understanding Workshop, ICCV 2019.
Source identification on Social Media platform
Recognizing information about the origin of a digital image has been individuated as a crucial task to be tackled by the image forensic scientific community. Understanding something on the previous history of an image could be strategic to address any successive assessment to be made on it: knowing the kind of device used for acquisition or, better, the model of the camera could focus investigations in a specific direction.
C. Pasquini, I. Amerini, G. Boato, "Media Forensics on Social Media Platforms: a Survey" EURASIP Journal on Information Security, 2021
Maiano, L.; Amerini, I.; Ricciardi Celsi, L.; Anagnostopoulos, A. Identification of Social-Media Platform of Videos through the Use of Shared Features. J. Imaging 2021, 7, 140
I.Amerini, A. Anagnostopoulos, L. Maiano, L. R. Celsi, Learning double-compression video fingerprints left from social media platforms,ICASSP 2021
R. Caldelli, R. Becarelli and I. Amerini, "Image Origin Classification Based on Social Network Provenance," in IEEE Transactions on Information Forensics and Security, vol. 12, no. 6, pp. 1299-1308, June 2017.doi: 10.1109/TIFS.2017.2656842
Adversarial Forensics
Attacks capable of removing SIFT keypoints from images have been recently devised with the intention of compromising the correct functioning of SIFT-based copy–move forgery detection. To tackle with these attacks, we propose three novel forensic detectors for the identification of images whose SIFT keypoints have been globally or locally removed. The detectors look for inconsistencies like the absence or anomalous distribution of keypoints within textured image regions.
A. Costanzo, I. Amerini, R. Caldelli, M. Barni, "Forensic Analysis of SIFT Keypoint Removal and Injection," Information Forensics and Security, IEEE Transactions on, vol.9, no.9, pp.1450,1464, Sept. 2014 doi: 10.1109/TIFS.2014.2337654.
Smartphone Sensors Fingerprinting
Previous works have shown that microphones can be uniquely identified by audio recordings since physical features of the microphone components leave repeatable and distinguishable traces in the recording. This concept can be used in security applications to perform identification of a mobile phone through the built-in microphone. The problem is to determine an accurate but also efficient representation of the physical characteristics, which is not known a priori. There is also usually a trade-off between identification accuracy and the time requested to perform the classification. Various approaches have been used in literature ranging from the identification and application of handcrafted statistical features to the recent application of Deep Learning techniques. This project evaluates the application of different entropy measures and their suitability for microphone classification.
Gianmarco Baldini and Irene Amerini, “Online Distributed Denial of Service (DDoS) intrusion detection based on adaptive sliding window and morphological fractal dimension”, Computer Networks 2022
G. Baldini, I. Amerini, "Microphone Identification based on Spectral Entropy with Convolutional Neural Network", Workshop on Information Forensics and Security 2022.
Baldini, G.; Amerini, I. “An Evaluation of Entropy Measures for Microphone Identification”. Entropy 2020, 22, 1235.
G. Baldini. I. Amerini, "Smartphones identification through the built-in microphones with Convolutional Neural Network", IEEE Access, 2019 DOI 10.1109/ACCESS.2019.2950859.
I. Amerini, R. Becarelli, R. Caldelli, A. Melani and M. Niccolai, "Smartphone Fingerprinting Combining Features of On-Board Sensors," in IEEE Transactions on Information Forensics and Security, vol. 12, no. 10, pp. 2457-2466, Oct. 2017.doi: 10.1109/TIFS.2017.2708685.