A5042746008- Advanced Image and Video Retrieval Techniques
- Digital Media Forensic Detection
- Advanced Vision and Imaging
- Image Retrieval and Classification Techniques
- Video Surveillance and Tracking Methods
- Advanced Steganography and Watermarking Techniques
- Image Processing Techniques and Applications
- Generative Adversarial Networks and Image Synthesis
- Image and Signal Denoising Methods
- Advanced Image Processing Techniques
- Video Analysis and Summarization
- Image Enhancement Techniques
- Image Processing and 3D Reconstruction
- Image and Video Stabilization
- Advanced Data Compression Techniques
- Robotics and Sensor-Based Localization
- Anomaly Detection Techniques and Applications
- Human Pose and Action Recognition
- Digital and Cyber Forensics
- Medical Image Segmentation Techniques
- 3D Surveying and Cultural Heritage
- Industrial Vision Systems and Defect Detection
- Visual Attention and Saliency Detection
- Computer Graphics and Visualization Techniques
- Autonomous Vehicle Technology and Safety
University of Catania
2016-2025
University of Brighton
2023
Edinburgh Napier University
2022
STMicroelectronics (Italy)
2002-2021
Laboratoire d'Informatique de Paris-Nord
2021
University of Messina
2020
International Society for Optics and Photonics
2015
Society for Imaging Science and Technology
2015
STMicroelectronics (Switzerland)
2002-2003
STMicroelectronics (United States)
2003
The Deepfake phenomenon has become very popular nowadays thanks to the possibility create incredibly realistic images using deep learning tools, based mainly on ad-hoc Generative Adversarial Networks (GAN). In this work we focus analysis of Deepfakes human faces with objective creating a new detection method able detect forensics trace hidden in images: sort fingerprint left image generation process. proposed technique, by means an Expectation Maximization (EM) algorithm, extracts set local...
The analysis of financial data represents a challenge that researchers had to deal with. rethinking the basis markets has led an urgent demand for developing innovative models understand assets. In past few decades, have proposed several systems based on traditional approaches, such as autoregressive integrated moving average (ARIMA) and exponential smoothing model, in order devise accurate representation. Despite their efficacy, existing works face some drawbacks due poor performance when...
Detecting and recognizing deepfakes is a pressing issue in the digital age. In this study, we first collected dataset of pristine images fake ones properly generated by nine different Generative Adversarial Network (GAN) architectures four Diffusion Models (DM). The contained total 83,000 images, with equal distribution between real deepfake data. Then, to address detection recognition tasks, proposed hierarchical multi-level approach. At level, classified from AI-generated ones. second...
This paper presents a video stabilization algorithm based on the extraction and tracking of scale invariant feature transform features through frames. Implementation SIFT operator is analyzed adapted to be used in feature-based motion estimation algorithm. are extracted from frames then their trajectory evaluated estimate interframe motion. A modified version iterative least squares method adopted avoid errors tracked as they appear nearby improve stability. Intentional camera eventually...
This paper presents a video stabilization algorithm based on the extraction and tracking of scale invariant feature transform features through frames. Implementation SIFT operator is analyzed adapted to be used in feature-based motion estimation algorithm. are extracted from frames then their trajectory evaluated estimate interframe motion. A modified version iterative least squares method adopted avoid errors tracked as they appear nearby improve stability. Intentional camera eventually...
Physiological signals are widely used to perform medical assessment for monitoring an extensive range of pathologies, usually related cardio-vascular diseases. Among these, both PhotoPlethysmoGraphy (PPG) and Electrocardiography (ECG) those more employed. PPG emerging non-invasive measurement technique study blood volume pulsations through the detection analysis back-scattered optical radiation coming from skin. ECG is process recording electrical activity heart over a period time using...
Advances in Artificial Intelligence and Image Processing are changing the way people interacts with digital images video. Widespread mobile apps like FACEAPP make use of most advanced Generative Adversarial Networks (GAN) to produce extreme transformations on human face photos such gender swap, aging, etc. The results utterly realistic extremely easy be exploited even for non-experienced users. This kind media object took name Deepfake raised a new challenge multimedia forensics field:...
Magnetic resonance imaging is a fundamental tool to reach diagnosis of multiple sclerosis and monitoring its progression. Although several attempts have been made segment lesions using artificial intelligence, fully automated analysis not yet available. State-of-the-art methods rely on slight variations in segmentation architectures (e.g. U-Net, etc.). However, recent research has demonstrated how exploiting temporal-aware features attention mechanisms can provide significant boost...
This paper presents a novel and fully automatic technique to estimate depth information from single input image. The proposed method is based on new image classification able classify digital images (also in Bayer pattern format) as indoor, outdoor with geometric elements or without elements. Using the collected step suitable map estimated. unsupervised generate view of scene, requiring low computational resources.
Today, many people in the world without any (or with little) knowledge about video recording, thanks to widespread use of mobile devices (personal digital assistants, phones, etc.), take videos. However, unwanted movements their hands typically blur and introduce disturbing jerkiness recorded sequences. Many stabilization techniques have been hence developed different performances but only fast strategies can be implemented on embedded devices. A fundamental issue is overall robustness...
The classification of food images is an interesting and challenging problem since the high variability image content which makes task difficult for current state-of-the-art methods. representation to be employed in engine plays important role. We believe that texture features have been not properly considered this application domain. This paper points out, through a set experiments, textures are fundamental recognize different items. For purpose bag visual words model (BoW) employed. Images...
In the midst of novel coronavirus (SARS-CoV-2) epidemic, examining reported case data could lead to biased speculations and conclusions. Indeed, estimation unreported infections is crucial for a better understanding current emergency in China other countries. this study, we aimed estimate number prior 23 January 2020 restrictions. To do this, developed Susceptible–Exposed–Infectious–Recovered–Dead (SEIRD) model that estimated from deaths. Our approach relied on fact observed deaths were less...
Multimedia data manipulation and forgery has never been easier than today, thanks to the power of Artificial Intelligence (AI). AI-generated fake content, commonly called Deepfakes, have raising new issues concerns, but also challenges for research community. The Deepfake detection task become widely addressed, unfortunately, approaches in literature suffer from generalization issues. In this paper, Face Detection Reconstruction Challenge is described. Two different tasks were proposed...
The <i>Journal of Electronic Imaging</i> (JEI), copublished bimonthly with the Society for Imaging Science and Technology, publishes peer-reviewed papers that cover research applications in all areas electronic imaging science technology.