Several pattern recognition and classification techniques have been applied to the biometrics domain. Among them, an interesting technique is Scale Invariant Feature Transform (SIFT), originally devised for object recognition. Even if SIFT features emerged as a very powerful image descriptors, their employment in face analysis context has never systematically investigated. This paper investigates application of approach authentication. In order determine real potential applicability method,...

In recent years face recognition systems have been applied in various useful applications, such as surveillance, access control, criminal investigations, law enforcement, and others. However biometric can be highly vulnerable to spoofing attacks where an impostor tries bypass the system using a photo or video sequence. this paper novel liveness detection method, based on 3D structure of face, is proposed. Processing curvature acquired data, proposed approach allows distinguish real from...

Dementia represents a global public health concern, with the early detection of Alzheimer's disease, most prevalent form dementia, being paramount importance. Given limited availability suitable biomarkers, research has shown that cognitive impairment can be identified through patients' spoken language. This paper presents multi-modal system for automatic disease using speech. The been trained on recordings healthy individuals and patients describing an image, task requiring linguistic...

In recent years the number of credentials and digital identities needed to access online services is exploding, increasing risk fraud unauthorized personal data.Biometric verification techniques can help build more secure authentication systems; in particular face recognition algorithms have reached near perfect accuracy under ideal conditions. Europe 60% population owns an electronic identification (eID) document which contains biometric data, almost all member states provide tools verify...

Visual surveillance in outdoor environments requires the monitoring of both objects and events. The analysis is generally driven by target application which, turn, determines set relevant events to be analyzed. In this paper we concentrate on scenes, particular for vehicle traffic control. scenario, weather conditions considered signal potentially dangerous situations like presence snow, fog, or heavy rain. developed system uses a statistical framework based mixture Gaussians identify...

This paper develops and demonstrates an original approach to face-image analysis based on identifying distinctive areas of each individual's face by its comparison others in the population. The method differs from most others—that we refer as unary —where salient regions are defined analyzing only images same individual. We extract a set multiscale patches image before projecting them into common feature space. degree “distinctiveness” any patch depends distance space mapped other...

This paper proposes a procedure for facial template synthesis based on features extracted from multiple instances with varying pose. The proposed system extracts the rotation, scale and translation invariant SIFT features, also having high discrimination ability, frontal half left right profiles of an individual face images. These affine obviate need ad-hoc algorithms to register side against feature-set augmentation. An augmented feature set is then formed fusion individual, after removing...

Abstract Traditional local image descriptors such as SIFT and SURF are based on processings similar to those that take place in the early visual cortex. Nowadays, convolutional neural networks still draw inspiration from human vision system, integrating computational elements typical of higher cortical areas. Deep CNN’s architectures intrinsically hard interpret, so much effort has been made dissect them order understand which type features they learn. However, considering resemblance no...

Face recognition with Deep Learning is generally approached as a problem of capacity. The field has seen progressively deeper, more complex models or larger, highly variant datasets. However, the carbon footprint machine learning (ML) concern. A real push developing to reduce energy consumption ML we strive for eco-friendly society. Lower compute budget always desirable, if accuracy not reduced below usable level. We present an approach using state art Vision Transformer and Early Exits...

This research focuses on developing improved diagnostic tools for Alzheimer's Disease (AD), a condition impacting approximately 50 million individuals globally. In the paper, we achieve automatic AD detection by leveraging pre-trained Large Language Models (LLMs) linguistic analysis applied to ADReSS/ADReSSo (Alzheimer's Dementia Recognition through Spontaneous Speech/only) Challenges datasets, following speech-to-text conversion. While recent advancements in LLMs offer robust foundation...

When dealing with face recognition, multimodal algorithms, their potential to capture complementary characteristics from the 2D and 3D data channels, can reach high level of efficiency robustness. In this paper, we explore different combinations iconic descriptors coupled a shape descriptor propose fully automatic, multimodal, recognition paradigm. Two features extractors, Scale Invariant Feature Transform (SIFT) Speeded-Up Robust Features (SURF), are used, in turn, extract salient points...

Abstract Multiscale models are among the cutting-edge technologies used for face detection and recognition. An example is Deformable part-based (DPMs), which encode a as multiplicity of local areas (parts) at different resolution scales their hierarchical spatial relationship. Although these have proven successful incredibly efficient in practical applications, mutual position parts involved arbitrarily defined by human specialist final choice optimal based on heuristics. This work seeks to...

Attention in Machine Learning allows a model to selectively up-weight informative parts of an input relation others. The Vision Transformer (ViT) is entirely based on attention. ViTs have shown state the art performance multiple fields including person re-identification, presentation attack detection and object recognition. Several works that embedding human attention into pipeline can improve or compensate for lack data. However correlation between computer vision models has not yet been...

Following other researchers, we investigated the premise that visual judgment of kinship might be modelled as a signal-detection task, strictly related to similar facial features. We measured subjects' response times face-pair stimuli while they performed judgments kinship, similarity, or dissimilarity, and examined some priming effects involved. Our results show takes longer on average than either similarity dissimilarity judgment—which is compatible with existing models, yet also suggest...

In recent years, CNNs are capturing the attention of a large community researchers, attracted by high performance this approach and surprising results obtained in many recognition/classification activities. Unfortunately, excellent CNN-based systems is accompanied worryingly poor understanding why they work so well. document, basic mechanisms related to extraction points interest (following first convolution phases) considered compared human fixations, with aim better analogies differences...

Hybrid face recognition methods combine holistic and feature based approaches with the aim of reaching a high level efficiency robustness. In this paper we propose fully automatic algorithm for multimodal data consisting 2D images their corresponding 3D scans. The is on extraction simple image features using Scale Invariant Feature Transform validation key-points scans by means Joint Differential Invariants local global shape information. process goes through an optimisation procedure: first...

Early hierarchical computational visual models as well recent deep neural networks have been inspired by the functioning of primate cortex system. Although much effort has made to dissect visualize features they learn at individual units, scope visualizations limited a categorization in terms their semantic level. Considering ability humans select high level regions scene, question whether can match this ability, and if similarity with attention is correlated performance naturally arise. To...