A5054837437- Face recognition and analysis
- Human Pose and Action Recognition
- Video Surveillance and Tracking Methods
- Hand Gesture Recognition Systems
- Anomaly Detection Techniques and Applications
- Generative Adversarial Networks and Image Synthesis
- Face and Expression Recognition
- Biometric Identification and Security
- Advanced Neural Network Applications
- Advanced Memory and Neural Computing
- Advanced Vision and Imaging
- CCD and CMOS Imaging Sensors
- Advanced Image and Video Retrieval Techniques
- Gait Recognition and Analysis
- Industrial Vision Systems and Defect Detection
- Advanced Image Processing Techniques
- Linguistic Studies and Language Acquisition
- Domain Adaptation and Few-Shot Learning
- Ancient Mediterranean Archaeology and History
- Digital Media Forensic Detection
- Linguistics and language evolution
- Classical Antiquity Studies
- Gaze Tracking and Assistive Technology
- Neural dynamics and brain function
- Data Stream Mining Techniques
University of Bologna
2006-2024
University of Modena and Reggio Emilia
2016-2024
Ferrari (Italy)
2017-2024
This article presents FRCSyn-onGoing, an ongoing challenge for face recognition where researchers can easily benchmark their systems against the state of art in open common platform using large-scale public databases and standard experimental protocols. FRCSyn-onGoing is based on Face Recognition Challenge Era Synthetic Data (FRCSyn) organized at WACV 2024. first international aiming to explore use real synthetic data independently, also fusion, order address existing limitations technology....
Fast and accurate upper-body head pose estimation is a key task for automatic monitoring of driver attention, challenging context characterized by severe illumination changes, occlusions extreme poses. In this work, we present new deep learning framework localization on depth images. The core the proposal regressive neural network, called POSEidon, which composed three independent convolutional nets followed fusion layer, specially conceived understanding depth. addition, to recover...
Depth cameras allow to set up reliable solutions for people monitoring and behavior understanding, especially when unstable or poor illumination conditions make unusable common RGB sensors. Therefore, we propose a complete framework the estimation of head shoulder pose based on depth images only. A detection localization module is also included, in order develop end-to-end system. The core element Convolutional Neural Network, called POSEidon <sup...
Despite the widespread adoption of face recognition technology around world, and its remarkable performance on current benchmarks, there are still several challenges that must be covered in more detail. This paper offers an overview Face Recognition Challenge Era Synthetic Data (FRCSyn) organized at WACV 2024. is first international challenge aiming to explore use synthetic data address existing limitations technology. Specifically, FRCSyn targets concerns related privacy issues, demographic...
Designing efficient neural networks for embedded devices is a critical challenge, particularly in applications requiring real-time performance, such as aerial imaging with drones and UAVs emergency responses. In this work, we introduce TakuNet, novel light-weight architecture which employs techniques depth-wise convolutions an early downsampling stem to reduce computational complexity while maintaining high accuracy. It leverages dense connections fast convergence during training uses 16-bit...
Transformer-based neural networks represent a successful self-attention mechanism that achieves state-of-the-art results in language understanding and sequence modeling. However, their application to visual data and, particular, the dynamic hand gesture recognition task has not yet been deeply investigated. In this paper, we propose transformer-based architecture for task. We show employment of single active depth sensor, specifically usage maps surface normals estimated from them, results,...
Background and objective: The aim of this prospective, randomized, double-blind investigation was to assess the dose–effect characteristics postoperative nausea vomiting after intrathecal administration small doses morphine (from 0.015 0.25 mg) in opioid-naïve, non-surgical patients. Methods: With Ethic Committee approval written informed consent 144 opioid-naïve patients suffering from non-cancerous chronic back-pain, receiving as diagnostic test for their pain, were randomly allocated...
The correct estimation of the head pose is a problem great importance for many applications.For instance, it an enabling technology in automotive driver attention monitoring.In this paper, we tackle through deep learning network working regression manner.Traditional methods usually rely on visual facial features, such as landmarks or nose tip position.In contrast, exploit Convolutional Neural Network (CNN) to perform directly from depth data.We Siamese architecture and propose novel loss...
Face morphing and related attacks have emerged as a serious security threat for automatic face recognition systems challenging research field. Therefore, the availability of effective reliable attack detectors is strongly needed. In this paper, we proposed framework based on double Siamese architecture to tackle detection task in differential scenario, which two images, trusted live acquired image probe (morphed or bona fide) are given input system. particular, presented aimed merge...
applied to morphed images, aiming reduce or even remove any visible artifact and then produce high-quality images.However, this procedure is subjective, tedious, time-consuming: these issues prevent the creation of large datasets with retouched images that can be used train test Morphing Attack Detection (MAD) algorithms [4], i.e. automated tools explicitly devised detect presence morphing in input images.More recently, based on GANs [5] have been introduced as alternatives traditional...
An accurate and fast driver's head pose estimation is a rich source of information, in particular the automotive context. Head key element for behavior investigation, analysis, attention monitoring also useful component to improve efficacy Human-Car Interaction systems. In this paper, Recurrent Neural Network exploited tackle problem driver estimation, directly only working on depth images be more reliable presence varying or insufficient illumination. Experimental results, obtained from two...
The ability to detect, localize and classify objects that are anomalies is a challenging task in the computer vision community. In this paper, we tackle these tasks developing framework automatically inspect railway during night. Specifically, it able predict presence, image coordinates class of obstacles. To deal with low-light environment, based on thermal images consists three different modules address problem detecting anomalies, predicting their classifying them. Moreover, due absolute...
The ability to detect, localize and track the hands is crucial in many applications requiring understanding of person behavior, attitude interactions. In particular, this true for automotive context, which hand analysis allows predict preparatory movements maneuvers or investigate driver's attention level. Moreover, due recent diffusion cameras inside new car cockpits, it feasible use gestures develop Human-Car Interaction systems, more user-friendly safe. paper, we propose a dataset, called...
The recent spread of low-cost and high-quality RGB-D infrared sensors has supported the development Natural User Interfaces (NUIs) in which interaction is carried without use physical devices such as keyboards mouse. In this paper, we propose a NUI based on dynamic hand gestures, acquired with RGB, depth sensors. system developed for challenging automotive context, aiming at reducing driver’s distraction during driving activity. Specifically, proposed framework multimodal combination...
Event cameras are biologically-inspired sensors that gather the temporal evolution of scene.They capture pixel-wise brightness variations and output a corresponding stream asynchronous events.Despite having multiple advantages with respect to traditional cameras, their use is partially prevented by limited applicability data processing vision algorithms.To this aim, we present framework which exploits event synthesize RGB frames, relying on an initial or periodic set color key-frames...
Nowadays, we are witnessing the wide diffusion of active depth sensors. However, generalization capabilities and performance deep face recognition approaches that based on data hindered by different sensor technologies currently available depth-based datasets, which limited in size acquired through same device. In this paper, present an analysis use maps, as obtained sensors neural architectures for task. We compare representations (depth normal images, voxels, point clouds), models...
Existing Continual Learning benchmarks only partially address the complexity of real-life applications, limiting realism learning agents. In this letter, we propose and focus on characterized by common key elements scenarios, including temporally ordered streams as input data, strong correlation samples in short time ranges, high data distribution drift over long frame, heavy class unbalancing. Moreover, enforce online training constraints such need for frequent model updates without...