Search

RAS is a signaling protein associated with the cell membrane that is mutated in up to 30% of human cancers. RAS signaling has been proposed to be regulated by dynamic heterogeneity of the cell membrane. Investigating such a mechanism requires near-atomistic detail at macroscopic temporal and spatial scales, which is not possible with conventional computational or experimental techniques. We demonstrate here a multiscale simulation infrastructure that uses machine learning to create a scale-bridging ensemble of over 100,000 simulations of active wild-type KRAS on a complex, asymmetric membrane. Initialized and validated with experimental data (including a new structure of active wild-type KRAS), these simulations represent a substantial advance in the ability to characterize RAS-membrane biology. We report distinctive patterns of local lipid composition that correlate with interfacially promiscuous RAS multimerization. These lipid fingerprints are coupled to RAS dynamics, predicted to influence effector binding, and therefore may be a mechanism for regulating cell signaling cascades.

RAS is a signaling protein associated with the cell membrane that is mutated in up to 30% of human cancers. RAS signaling has been proposed to be regulated by dynamic heterogeneity of the cell membrane. Investigating such a mechanism requires near-atomistic detail at macroscopic temporal and spatial scales, which is not possible with conventional computational or experimental techniques. We demonstrate here a multiscale simulation infrastructure that uses machine learning to create a scale-bridging ensemble of over 100,000 simulations of active wild-type KRAS on a complex, asymmetric membrane. Initialized and validated with experimental data (including a new structure of active wild-type KRAS), these simulations represent a substantial advance in the ability to characterize RAS-membrane biology. We report distinctive patterns of local lipid composition that correlate with interfacially promiscuous RAS multimerization. These lipid fingerprints are coupled to RAS dynamics, predicted to influence effector binding, and therefore may be a mechanism for regulating cell signaling cascades.

In recent years, the proliferation of unmanned aerial vehicles (UAVs) has increased dramatically. UAVs can accomplish complex or dangerous tasks in a reliable and cost-effective way but are still limited by power consumption problems, which pose serious constraints on the flight duration and completion of energy-demanding tasks. The possibility of providing UAVs with advanced decision-making capabilities in an energy-effective way would be extremely beneficial. In this paper, we propose a practical solution to this problem that exploits deep learning on the edge. The developed system integrates an OpenMV microcontroller into a DJI Tello Micro Aerial Vehicle (MAV). The microcontroller hosts a set of machine learning-enabled inference tools that cooperate to control the navigation of the drone and complete a given mission objective. The goal of this approach is to leverage the new opportunistic features of TinyML through OpenMV including offline inference, low latency, energy efficiency, and data security. The approach is successfully validated on a practical application consisting of the onboard detection of people wearing protection masks in a crowded environment.

The amount of multimedia content shared everyday, combined with the level of realism reached by recent fake-generating technologies, threatens to impair the trustworthiness of online information sources. The process of uploading and sharing data tends to hinder standard media forensic analyses, since multiple re-sharing steps progressively hide the traces of past manipulations. At the same time though, new traces are introduced by the platforms themselves, enabling the reconstruction of the sharing history of digital objects, with possible applications in information flow monitoring and source identification. In this work, we propose a supervised framework for the reconstruction of image sharing chains on social media platforms. The system is structured as a cascade of backtracking blocks, each of them tracing back one step of the sharing chain at a time. Blocks are designed as ensembles of classifiers trained to analyse the input image independently from one another by leveraging different feature representations that describe both content and container of the media object. Individual decisions are then properly combined by a late fusion strategy. Results highlight the advantages of employing multiple clues, which allow accurately tracing back up to three steps along the sharing chain.

Digital images are becoming the most commonly used multimedia data nowadays thanks to the massive manufacturing of cheap acquisition devices coupled with the unprecedented popularity of Online Social Networks. As two sides of a coin, massive use of digital images triggers the development of user-friendly editing tools and intelligent techniques that violate image authenticity. By this respect, digital images are less and less trustable as they are easily modified not only by experts or researchers, but also by unexperienced users. It has been also witnessed that malicious use of images has tremendous impact on human perception as well as system reliability. Those concerns highlight the importance to verify image authenticity. In practice, digital images are created, manipulated, and diffused world-wide via many channels. Simply answering to the question "Is an image authentic?'' appears insufficient. Further steps aiming at understanding the provenance of images with respect to acquisition device or distributed platforms as well as the processing history have to be considered significant. This doctoral study contributes solutions to recover digital image provenance under multiple aspects: i) image acquisition device, ii) social network origin, and iii) source-target disambiguation in image copy-move forgery.

Digital images are becoming the most commonly used multimedia data nowadays thanks to the massive manufacturing of cheap acquisition devices coupled with the unprecedented popularity of Online Social Networks. As two sides of a coin, massive use of digital images triggers the development of user-friendly editing tools and intelligent techniques that violate image authenticity. By this respect, digital images are less and less trustable as they are easily modified not only by experts or researchers, but also by unexperienced users. It has been also witnessed that malicious use of images has tremendous impact on human perception as well as system reliability. Those concerns highlight the importance to verify image authenticity. In practice, digital images are created, manipulated, and diffused world-wide via many channels. Simply answering to the question "Is an image authentic?'' appears insufficient. Further steps aiming at understanding the provenance of images with respect to acquisition device or distributed platforms as well as the processing history have to be considered significant. This doctoral study contributes solutions to recover digital image provenance under multiple aspects: i) image acquisition device, ii) social network origin, and iii) source-target disambiguation in image copy-move forgery.

Digital images are becoming the most commonly used multimedia data nowadays thanks to the massive manufacturing of cheap acquisition devices coupled with the unprecedented popularity of Online Social Networks. As two sides of a coin, massive use of digital images triggers the development of user-friendly editing tools and intelligent techniques that violate image authenticity. By this respect, digital images are less and less trustable as they are easily modified not only by experts or researchers, but also by unexperienced users. It has been also witnessed that malicious use of images has tremendous impact on human perception as well as system reliability. Those concerns highlight the importance to verify image authenticity. In practice, digital images are created, manipulated, and diffused world-wide via many channels. Simply answering to the question "Is an image authentic?'' appears insufficient. Further steps aiming at understanding the provenance of images with respect to acquisition device or distributed platforms as well as the processing history have to be considered significant. This doctoral study contributes solutions to recover digital image provenance under multiple aspects: i) image acquisition device, ii) social network origin, and iii) source-target disambiguation in image copy-move forgery.

Digital images are becoming the most commonly used multimedia data nowadays thanks to the massive manufacturing of cheap acquisition devices coupled with the unprecedented popularity of Online Social Networks. As two sides of a coin, massive use of digital images triggers the development of user-friendly editing tools and intelligent techniques that violate image authenticity. By this respect, digital images are less and less trustable as they are easily modified not only by experts or researchers, but also by unexperienced users. It has been also witnessed that malicious use of images has tremendous impact on human perception as well as system reliability. Those concerns highlight the importance to verify image authenticity. In practice, digital images are created, manipulated, and diffused world-wide via many channels. Simply answering to the question "Is an image authentic?'' appears insufficient. Further steps aiming at understanding the provenance of images with respect to acquisition device or distributed platforms as well as the processing history have to be considered significant. This doctoral study contributes solutions to recover digital image provenance under multiple aspects: i) image acquisition device, ii) social network origin, and iii) source-target disambiguation in image copy-move forgery.

With the growing complexity of computational and experimental facilities, many scientific researchers are turning to machine learning (ML) techniques to analyze large scale ensemble data. With complexities such as multi-component workflows, heterogeneous machine architectures, parallel file systems, and batch scheduling, care must be taken to facilitate this analysis in a high performance computing (HPC) environment. In this paper, we present Merlin, a workflow framework to enable large ML-friendly ensembles of scientific HPC simulations. By augmenting traditional HPC with distributed compute technologies, Merlin aims to lower the barrier for scientific subject matter experts to incorporate ML into their analysis. In addition to its design, we describe some example applications that Merlin has enabled on leadership-class HPC resources, such as the ML-augmented optimization of nuclear fusion experiments and the calibration of infectious disease models to study the progression of and possible mitigation strategies for COVID-19., Comment: 28 pages, 9 figures; Submitted to FGCS

Knowing how to exploit unlabeled data is a fundamental problem in machine learning. This dissertation provides contributions in different contexts, including semi-supervised learning, positive unlabeled learning and representation learning. In particular, we ask (i) whether is possible to learn a classifier in the context of limited data, (ii) whether is possible to scale existing models for positive unlabeled learning, and (iii) whether is possible to train a deep generative model with a single minimization problem.

Knowing how to exploit unlabeled data is a fundamental problem in machine learning. This dissertation provides contributions in different contexts, including semi-supervised learning, positive unlabeled learning and representation learning. In particular, we ask (i) whether is possible to learn a classifier in the context of limited data, (ii) whether is possible to scale existing models for positive unlabeled learning, and (iii) whether is possible to train a deep generative model with a single minimization problem.

Knowing how to exploit unlabeled data is a fundamental problem in machine learning. This dissertation provides contributions in different contexts, including semi-supervised learning, positive unlabeled learning and representation learning. In particular, we ask (i) whether is possible to learn a classifier in the context of limited data, (ii) whether is possible to scale existing models for positive unlabeled learning, and (iii) whether is possible to train a deep generative model with a single minimization problem.

Knowing how to exploit unlabeled data is a fundamental problem in machine learning. This dissertation provides contributions in different contexts, including semi-supervised learning, positive unlabeled learning and representation learning. In particular, we ask (i) whether is possible to learn a classifier in the context of limited data, (ii) whether is possible to scale existing models for positive unlabeled learning, and (iii) whether is possible to train a deep generative model with a single minimization problem.

Events have always been fundamental building blocks of individual lives as well as of the whole world. Nowadays, thanks to the several technological advances achieved within the digital age, the processes of capturing, describing and spreading events have never been so simple and intuitive. This results in an ubiquitous presence of event-related information, which is digitally embedded in any form of media. Both the pervasiveness of such information as well as the benefits of its exploitation for many purposes have fostered decades of research effort to detect and summarize it. However, several issues emerge at subsequent stages and shall be addressed to support the proper exploitation and consumption of event-related information. The work presented within this thesis is indeed committed to this goal. The aforementioned ubiquity of events makes them exhibit different characteristics and appear in a diverse range of scenarios. Therefore, we categorize events according to three main aspects that come into play when considering the management and usage of event-related information over time, once it has been created. These are the degree of privacy, as events can be of public domain or rather pertain to a more personal sphere, the type of description, which is the form (e.g. textual or visual) in which events are described, and the time of usage, namely the temporal horizon over which event-related information is expected to be accessed and used. The problems addressed in this thesis regard different combinations of such aspects, each one subject to specific issues to be dealt with. Concerning the private sphere, we aim at properly managing large amounts of photographs taken during personal events, so that they can be easily revisited and enjoyed in the future. The common habit of dumping every single picture, encouraged by the availability of cheap storage devices, poses serious threats to their future revisiting and calls for more selective strategies to identify the

Knowing how to exploit unlabeled data is a fundamental problem in machine learning. This dissertation provides contributions in different contexts, including semi-supervised learning, positive unlabeled learning and representation learning. In particular, we ask (i) whether is possible to learn a classifier in the context of limited data, (ii) whether is possible to scale existing models for positive unlabeled learning, and (iii) whether is possible to train a deep generative model with a single minimization problem.

Knowing how to exploit unlabeled data is a fundamental problem in machine learning. This dissertation provides contributions in different contexts, including semi-supervised learning, positive unlabeled learning and representation learning. In particular, we ask (i) whether is possible to learn a classifier in the context of limited data, (ii) whether is possible to scale existing models for positive unlabeled learning, and (iii) whether is possible to train a deep generative model with a single minimization problem.

Knowing how to exploit unlabeled data is a fundamental problem in machine learning. This dissertation provides contributions in different contexts, including semi-supervised learning, positive unlabeled learning and representation learning. In particular, we ask (i) whether is possible to learn a classifier in the context of limited data, (ii) whether is possible to scale existing models for positive unlabeled learning, and (iii) whether is possible to train a deep generative model with a single minimization problem.

Knowing how to exploit unlabeled data is a fundamental problem in machine learning. This dissertation provides contributions in different contexts, including semi-supervised learning, positive unlabeled learning and representation learning. In particular, we ask (i) whether is possible to learn a classifier in the context of limited data, (ii) whether is possible to scale existing models for positive unlabeled learning, and (iii) whether is possible to train a deep generative model with a single minimization problem.

Clustering images according to their acquisition devices is a well-known problem in multimedia forensics, which is typically faced by means of camera Sensor Pattern Noise (SPN). Such an issue is challenging since SPN is a noise-like signal, hard to be estimated and easy to be attenuated or destroyed by many factors. Moreover, the high dimensionality of SPN hinders large-scale applications. Existing approaches are typically based on the correlation among SPNs in the pixel domain, which might not be able to capture intrinsic data structure in union of vector subspaces. In this paper, we propose an accurate clustering framework, which exploits linear dependencies among SPNs in their intrinsic vector subspaces. Such dependencies are encoded under sparse representations which are obtained by solving a LASSO problem with non-negativity constraint. The proposed framework is highly accurate in number of clusters estimation and image association. Moreover, our framework is scalable to the number of images and robust against double JPEG compression as well as the presence of outliers, owning big potential for real-world applications. Experimental results on Dresden and Vision database show that our proposed framework can adapt well to both medium-scale and large-scale contexts, and outperforms state-of-the-art methods.

In this paper, we present a novel framework that can produce a visual description of a tourist attraction by choosing the most diverse pictures from community-contributed datasets, that describe different details of the queried location. The main strength of the proposed approach is its flexibility that permits to filter out non-relevant images, and to obtain a reliable set of diverse and relevant images by first clustering similar images according to their textual descriptions and their visual content, and then extracting images from different clusters according to a measure of user's credibility. Clustering is based on a two-step process where textual descriptions are used first, and the clusters are then refined according to the visual features. The degree of diversification can be further increased by exploiting users' judgments on the results produced by the proposed algorithm through a novel approach, where users not only provide a relevance feedback, but also a diversity feedback. Experimental results performed on the MediaEval 2015 ``Retrieving Diverse Social Images" dataset show that the proposed framework can achieve very good performance both in the case of automatic retrieval of diverse images, and in the case of the exploitation of the users' feedback. The effectiveness of the proposed approach has been also confirmed by a small case study involving a number of real users.

In this paper, we present a novel framework that can produce a visual description of a tourist attraction by choosing the most diverse pictures from community-contributed datasets, that describe different details of the queried location. The main strength of the proposed approach is its flexibility that permits to filter out non-relevant images, and to obtain a reliable set of diverse and relevant images by first clustering similar images according to their textual descriptions and their visual content, and then extracting images from different clusters according to a measure of user's credibility. Clustering is based on a two-step process where textual descriptions are used first, and the clusters are then refined according to the visual features. The degree of diversification can be further increased by exploiting users' judgments on the results produced by the proposed algorithm through a novel approach, where users not only provide a relevance feedback, but also a diversity feedback. Experimental results performed on the MediaEval 2015 ``Retrieving Diverse Social Images" dataset show that the proposed framework can achieve very good performance both in the case of automatic retrieval of diverse images, and in the case of the exploitation of the users' feedback. The effectiveness of the proposed approach has been also confirmed by a small case study involving a number of real users.

Face authentication has been shown to be vulnerable against three main kinds of attacks: print, replay, and 3D mask. Among those, video replay attacks appear more challenging to be detected. There exist in the literature many countermeasures to face spoofing attacks, but a sophisticated detector is still needed to deal with particularly high-quality video based attacks. In this work, we perform analysis on the noise residual in frequency domain, and extract discriminative features by using a dynamic texture descriptor to characterize video based spoofing attacks. We propose a promising detector, which produces competitive results on the most challenging dataset of video based spoofing.

Face authentication has been shown to be vulnerable against three main kinds of attacks: print, replay, and 3D mask. Among those, video replay attacks appear more challenging to be detected. There exist in the literature many countermeasures to face spoofing attacks, but a sophisticated detector is still needed to deal with particularly high-quality video based attacks. In this work, we perform analysis on the noise residual in frequency domain, and extract discriminative features by using a dynamic texture descriptor to characterize video based spoofing attacks. We propose a promising detector, which produces competitive results on the most challenging dataset of video based spoofing.

Training feedforward neural networks with standard logistic activations is considered difficult because of the intrinsic properties of these sigmoidal functions. This work aims at showing that these networks can be trained to achieve generalization performance comparable to those based on hyperbolic tangent activations. The solution consists on applying a set of conditions in parameter initialization, which have been derived from the study of the properties of a single neuron from an information-theoretic perspective. The proposed initialization is validated through an extensive experimental analysis.

Face authentication has been shown to be vulnerable against three main kinds of attacks: print, replay, and 3D mask. Among those, video replay attacks appear more challenging to be detected. There exist in the literature many countermeasures to face spoofing attacks, but a sophisticated detector is still needed to deal with particularly high-quality video based attacks. In this work, we perform analysis on the noise residual in frequency domain, and extract discriminative features by using a dynamic texture descriptor to characterize video based spoofing attacks. We propose a promising detector, which produces competitive results on the most challenging dataset of video based spoofing.

Positive unlabeled (PU) learning is useful in various practical situations, where there is a need to learn a classifier for a class of interest from an unlabeled data set, which may contain anomalies as well as samples from unknown classes. The learning task can be formulated as an optimization problem under the framework of statistical learning theory. Recent studies have theoretically analyzed its properties and generalization performance, nevertheless, little effort has been made to consider the problem of scalability, especially when large sets of unlabeled data are available. In this work we propose a novel scalable PU learning algorithm that is theoretically proven to provide the optimal solution, while showing superior computational and memory performance. Experimental evaluation confirms the theoretical evidence and shows that the proposed method can be successfully applied to a large variety of real-world problems involving PU learning., Comment: Submitted to IEEE TPAMI

The wide diffusion of multimedia contents of different type and format led to the need of effective methods to efficiently handle such huge amount of information, opening interesting research challenges in the media community. In particular, the definition of suitable content understanding methodologies is attracting the effort of a large number of researchers worldwide, who proposed various tools for automatic content organization, retrieval, search, annotation and summarization. In this thesis, we will focus on an important concept, that is the inherent link between ''media" and the ''events" that such media are depicting. We will present two different methodologies related to such problem, and in particular to the automatic discovery of event-semantics from media contents. The two methodologies address this general problem at two different levels of abstraction. In the first approach we will be concerned with the detection of activities and behaviors of people from a video sequence (i.e., what a person is doing and how), while in the second we will face the more general problem of understanding a class of events from a set visual media (i.e., the situation and context). Both problems will be addressed trying to avoid making strong a-priori assumptions, i.e., considering the largely unstructured and variable nature of events.As to the first methodology, we will discuss about events related to the behavior of a person living in a home environment. The automatic understanding of human activity is still an open problems in the scientific community, although several solutions have been proposed so far, and may provide important breakthroughs in many application domains such as context-aware computing, area monitoring and surveillance, assistive technologies for the elderly or disabled, and more. An innovative approach is presented in this thesis, providing (i) a compact representation of human activities, and (ii) an effective tool to reliably measure the similarity b

The wide diffusion of multimedia contents of different type and format led to the need of effective methods to efficiently handle such huge amount of information, opening interesting research challenges in the media community. In particular, the definition of suitable content understanding methodologies is attracting the effort of a large number of researchers worldwide, who proposed various tools for automatic content organization, retrieval, search, annotation and summarization. In this thesis, we will focus on an important concept, that is the inherent link between ''media" and the ''events" that such media are depicting. We will present two different methodologies related to such problem, and in particular to the automatic discovery of event-semantics from media contents. The two methodologies address this general problem at two different levels of abstraction. In the first approach we will be concerned with the detection of activities and behaviors of people from a video sequence (i.e., what a person is doing and how), while in the second we will face the more general problem of understanding a class of events from a set visual media (i.e., the situation and context). Both problems will be addressed trying to avoid making strong a-priori assumptions, i.e., considering the largely unstructured and variable nature of events.As to the first methodology, we will discuss about events related to the behavior of a person living in a home environment. The automatic understanding of human activity is still an open problems in the scientific community, although several solutions have been proposed so far, and may provide important breakthroughs in many application domains such as context-aware computing, area monitoring and surveillance, assistive technologies for the elderly or disabled, and more. An innovative approach is presented in this thesis, providing (i) a compact representation of human activities, and (ii) an effective tool to reliably measure the similarity b

The wide diffusion of multimedia contents of different type and format led to the need of effective methods to efficiently handle such huge amount of information, opening interesting research challenges in the media community. In particular, the definition of suitable content understanding methodologies is attracting the effort of a large number of researchers worldwide, who proposed various tools for automatic content organization, retrieval, search, annotation and summarization. In this thesis, we will focus on an important concept, that is the inherent link between ''media" and the ''events" that such media are depicting. We will present two different methodologies related to such problem, and in particular to the automatic discovery of event-semantics from media contents. The two methodologies address this general problem at two different levels of abstraction. In the first approach we will be concerned with the detection of activities and behaviors of people from a video sequence (i.e., what a person is doing and how), while in the second we will face the more general problem of understanding a class of events from a set visual media (i.e., the situation and context). Both problems will be addressed trying to avoid making strong a-priori assumptions, i.e., considering the largely unstructured and variable nature of events.As to the first methodology, we will discuss about events related to the behavior of a person living in a home environment. The automatic understanding of human activity is still an open problems in the scientific community, although several solutions have been proposed so far, and may provide important breakthroughs in many application domains such as context-aware computing, area monitoring and surveillance, assistive technologies for the elderly or disabled, and more. An innovative approach is presented in this thesis, providing (i) a compact representation of human activities, and (ii) an effective tool to reliably measure the similarity b

The wide diffusion of multimedia contents of different type and format led to the need of effective methods to efficiently handle such huge amount of information, opening interesting research challenges in the media community. In particular, the definition of suitable content understanding methodologies is attracting the effort of a large number of researchers worldwide, who proposed various tools for automatic content organization, retrieval, search, annotation and summarization. In this thesis, we will focus on an important concept, that is the inherent link between ''media" and the ''events" that such media are depicting. We will present two different methodologies related to such problem, and in particular to the automatic discovery of event-semantics from media contents. The two methodologies address this general problem at two different levels of abstraction. In the first approach we will be concerned with the detection of activities and behaviors of people from a video sequence (i.e., what a person is doing and how), while in the second we will face the more general problem of understanding a class of events from a set visual media (i.e., the situation and context). Both problems will be addressed trying to avoid making strong a-priori assumptions, i.e., considering the largely unstructured and variable nature of events.As to the first methodology, we will discuss about events related to the behavior of a person living in a home environment. The automatic understanding of human activity is still an open problems in the scientific community, although several solutions have been proposed so far, and may provide important breakthroughs in many application domains such as context-aware computing, area monitoring and surveillance, assistive technologies for the elderly or disabled, and more. An innovative approach is presented in this thesis, providing (i) a compact representation of human activities, and (ii) an effective tool to reliably measure the similarity b

The wide diffusion of multimedia contents of different type and format led to the need of effective methods to efficiently handle such huge amount of information, opening interesting research challenges in the media community. In particular, the definition of suitable content understanding methodologies is attracting the effort of a large number of researchers worldwide, who proposed various tools for automatic content organization, retrieval, search, annotation and summarization. In this thesis, we will focus on an important concept, that is the inherent link between ''media" and the ''events" that such media are depicting. We will present two different methodologies related to such problem, and in particular to the automatic discovery of event-semantics from media contents. The two methodologies address this general problem at two different levels of abstraction. In the first approach we will be concerned with the detection of activities and behaviors of people from a video sequence (i.e., what a person is doing and how), while in the second we will face the more general problem of understanding a class of events from a set visual media (i.e., the situation and context). Both problems will be addressed trying to avoid making strong a-priori assumptions, i.e., considering the largely unstructured and variable nature of events.As to the first methodology, we will discuss about events related to the behavior of a person living in a home environment. The automatic understanding of human activity is still an open problems in the scientific community, although several solutions have been proposed so far, and may provide important breakthroughs in many application domains such as context-aware computing, area monitoring and surveillance, assistive technologies for the elderly or disabled, and more. An innovative approach is presented in this thesis, providing (i) a compact representation of human activities, and (ii) an effective tool to reliably measure the similarity b

The wide diffusion of multimedia contents of different type and format led to the need of effective methods to efficiently handle such huge amount of information, opening interesting research challenges in the media community. In particular, the definition of suitable content understanding methodologies is attracting the effort of a large number of researchers worldwide, who proposed various tools for automatic content organization, retrieval, search, annotation and summarization. In this thesis, we will focus on an important concept, that is the inherent link between ''media" and the ''events" that such media are depicting. We will present two different methodologies related to such problem, and in particular to the automatic discovery of event-semantics from media contents. The two methodologies address this general problem at two different levels of abstraction. In the first approach we will be concerned with the detection of activities and behaviors of people from a video sequence (i.e., what a person is doing and how), while in the second we will face the more general problem of understanding a class of events from a set visual media (i.e., the situation and context). Both problems will be addressed trying to avoid making strong a-priori assumptions, i.e., considering the largely unstructured and variable nature of events.As to the first methodology, we will discuss about events related to the behavior of a person living in a home environment. The automatic understanding of human activity is still an open problems in the scientific community, although several solutions have been proposed so far, and may provide important breakthroughs in many application domains such as context-aware computing, area monitoring and surveillance, assistive technologies for the elderly or disabled, and more. An innovative approach is presented in this thesis, providing (i) a compact representation of human activities, and (ii) an effective tool to reliably measure the similarity b

People perceive any kind of information with different level of attention and involvement. It is due to the way how our brain functions, redundancy and importance of the perceived data. This work deals with visual information, in particular with images. Image analysis and processing is often requires running computationally expensive algorithms. The knowledge of which part of an image is important over other parts allows for reduction of data to be processed. Besides computational cost a broad variety of applications, including image compression, quality assessment, adaptive content display and rendering, can benefit from this kind of information. The development of an accurate visual importance estimation method may bring a useful tool for image processing domain and that is the main goal for this work. In the following two novel approaches to saliency detection are presented. In comparison to previous works in this field the proposed approaches tackles saliency estimation on the object-wise level. In addition, one of the proposed approach solves saliency detection problem through modelling 3-D spatial relationships between objects in a scene. Moreover, a novel idea of the application of saliency to diversification of image retrieval results is presented.

People perceive any kind of information with different level of attention and involvement. It is due to the way how our brain functions, redundancy and importance of the perceived data. This work deals with visual information, in particular with images. Image analysis and processing is often requires running computationally expensive algorithms. The knowledge of which part of an image is important over other parts allows for reduction of data to be processed. Besides computational cost a broad variety of applications, including image compression, quality assessment, adaptive content display and rendering, can benefit from this kind of information. The development of an accurate visual importance estimation method may bring a useful tool for image processing domain and that is the main goal for this work. In the following two novel approaches to saliency detection are presented. In comparison to previous works in this field the proposed approaches tackles saliency estimation on the object-wise level. In addition, one of the proposed approach solves saliency detection problem through modelling 3-D spatial relationships between objects in a scene. Moreover, a novel idea of the application of saliency to diversification of image retrieval results is presented.

People perceive any kind of information with different level of attention and involvement. It is due to the way how our brain functions, redundancy and importance of the perceived data. This work deals with visual information, in particular with images. Image analysis and processing is often requires running computationally expensive algorithms. The knowledge of which part of an image is important over other parts allows for reduction of data to be processed. Besides computational cost a broad variety of applications, including image compression, quality assessment, adaptive content display and rendering, can benefit from this kind of information. The development of an accurate visual importance estimation method may bring a useful tool for image processing domain and that is the main goal for this work. In the following two novel approaches to saliency detection are presented. In comparison to previous works in this field the proposed approaches tackles saliency estimation on the object-wise level. In addition, one of the proposed approach solves saliency detection problem through modelling 3-D spatial relationships between objects in a scene. Moreover, a novel idea of the application of saliency to diversification of image retrieval results is presented.

People perceive any kind of information with different level of attention and involvement. It is due to the way how our brain functions, redundancy and importance of the perceived data. This work deals with visual information, in particular with images. Image analysis and processing is often requires running computationally expensive algorithms. The knowledge of which part of an image is important over other parts allows for reduction of data to be processed. Besides computational cost a broad variety of applications, including image compression, quality assessment, adaptive content display and rendering, can benefit from this kind of information. The development of an accurate visual importance estimation method may bring a useful tool for image processing domain and that is the main goal for this work. In the following two novel approaches to saliency detection are presented. In comparison to previous works in this field the proposed approaches tackles saliency estimation on the object-wise level. In addition, one of the proposed approach solves saliency detection problem through modelling 3-D spatial relationships between objects in a scene. Moreover, a novel idea of the application of saliency to diversification of image retrieval results is presented.

The biological effects of energetic heavy ions are attracting increasing interest for their applications in cancer therapy and protection against space radiation. The cascade of events leading to cell death or late effects starts from stochastic energy deposition on the nanometer scale and the corresponding lesions in biological molecules, primarily DNA. We have developed experimental techniques to visualize DNA nanolesions induced by heavy ions. Nanolesions appear in cells as “streaks” which can be visualized by using different DNA repair markers. We have studied the kinetics of repair of these “streaks” also with respect to the chromatin conformation. Initial steps in the modeling of the energy deposition patterns at the micrometer and nanometer scale were made with MCHIT and TRAX models, respectively.

The automatic understanding of human activity is probably one of the most challenging problems for the scientific community. Several application domains would benefit of such an analysis, from context-aware computing, to area monitoring and surveillance, to assistive technologies for elderly or disabled, and more. In a broad sense, we can define the activity analysis as the problem of finding an explanation coherent with a set of observations. These observations are typically influenced by several factors from different disciplines, such as sociology or psychology, but also mathematics and physics, making the problem particularly hard. In the last years, also the computer vision community focused its attention on this area, producing the latest advances in the acquisition and understanding of human motion data from image sequences. Despite the increasing effort spent in this field, there still exists a consistent gap between the numerical low-level pixel information that can be observed and measured, and the high abstraction level of the semantic that describes a given activity. In other words, there exist a conceptual ambiguity between the image sequence observations and their possible interpretations. Although several factors are involved, the activity modeling and the comparison strategy play crucial roles. In this proposal, a correlation between activity and corresponding path has been assumed. In light of this, the work carried out tackles two strictly related issues: (i) obtaining a proper representation of human activity; (ii) define an effective tool for reliably measuring the similarity between activity instances. In particular, the object activity is modeled with a signature obtained through a symbolic abstraction of its spatio-temporal trace, allowing the application of particular high-level reasoning for computing the activity similarity. This representation is particularly effective since it provides a smart way to compensate the noise artifacts coming

The widespread use of digital cameras, as well as the increasing popularity of online photo sharing has led to the proliferation of networked photo collections. Handling such a huge amount of media, without imposing complex and time consuming archiving procedures, is highly desirable and poses a number of interesting research challenges to the media community. In particular, the definition of suitable content based indexing and retrieval methodologies is attracting the effort of a large number of researchers worldwide, who proposed various tools for automatic content organization, retrieval, search, annotation and summarization. In this thesis, we will present and discuss three different approaches for content-and-context based retrieval. The main focus will be put on personal photo albums, which can be considered one of the most challenging application domains in this field, due to the largely unstructured and variable nature of the datasets. The methodologies that we will describe can be summarized into the following three points: i. Stochastic approaches to exploit the user interaction in query-by-example photos retrieval. Understanding the subjective meaning of a visual query, by converting it into numerical parameters that can be extracted and compared by a computer, is the paramount challenge in the field of intelligent image retrieval, also referred to as the “semantic gap” problem. An innovative approach is proposed that combines a relevance feedback process with a stochastic optimization engine, as a way to grasp user's semantics through optimized iterative learning providing on one side a better exploration of the search space, and on the other side avoiding stagnation in local minima during the retrieval. ii. Unsupervised event collection, segmentation and summarization. The need for automatic tools able to extract salient moments and provide automatic summary of large photo galleries is becoming more and more important due to the exponential growth in th

The widespread use of digital cameras, as well as the increasing popularity of online photo sharing has led to the proliferation of networked photo collections. Handling such a huge amount of media, without imposing complex and time consuming archiving procedures, is highly desirable and poses a number of interesting research challenges to the media community. In particular, the definition of suitable content based indexing and retrieval methodologies is attracting the effort of a large number of researchers worldwide, who proposed various tools for automatic content organization, retrieval, search, annotation and summarization. In this thesis, we will present and discuss three different approaches for content-and-context based retrieval. The main focus will be put on personal photo albums, which can be considered one of the most challenging application domains in this field, due to the largely unstructured and variable nature of the datasets. The methodologies that we will describe can be summarized into the following three points: i. Stochastic approaches to exploit the user interaction in query-by-example photos retrieval. Understanding the subjective meaning of a visual query, by converting it into numerical parameters that can be extracted and compared by a computer, is the paramount challenge in the field of intelligent image retrieval, also referred to as the “semantic gap” problem. An innovative approach is proposed that combines a relevance feedback process with a stochastic optimization engine, as a way to grasp user's semantics through optimized iterative learning providing on one side a better exploration of the search space, and on the other side avoiding stagnation in local minima during the retrieval. ii. Unsupervised event collection, segmentation and summarization. The need for automatic tools able to extract salient moments and provide automatic summary of large photo galleries is becoming more and more important due to the exponential growth in th

The widespread use of digital cameras, as well as the increasing popularity of online photo sharing has led to the proliferation of networked photo collections. Handling such a huge amount of media, without imposing complex and time consuming archiving procedures, is highly desirable and poses a number of interesting research challenges to the media community. In particular, the definition of suitable content based indexing and retrieval methodologies is attracting the effort of a large number of researchers worldwide, who proposed various tools for automatic content organization, retrieval, search, annotation and summarization. In this thesis, we will present and discuss three different approaches for content-and-context based retrieval. The main focus will be put on personal photo albums, which can be considered one of the most challenging application domains in this field, due to the largely unstructured and variable nature of the datasets. The methodologies that we will describe can be summarized into the following three points: i. Stochastic approaches to exploit the user interaction in query-by-example photos retrieval. Understanding the subjective meaning of a visual query, by converting it into numerical parameters that can be extracted and compared by a computer, is the paramount challenge in the field of intelligent image retrieval, also referred to as the “semantic gap” problem. An innovative approach is proposed that combines a relevance feedback process with a stochastic optimization engine, as a way to grasp user's semantics through optimized iterative learning providing on one side a better exploration of the search space, and on the other side avoiding stagnation in local minima during the retrieval. ii. Unsupervised event collection, segmentation and summarization. The need for automatic tools able to extract salient moments and provide automatic summary of large photo galleries is becoming more and more important due to the exponential growth in th

The automatic understanding of human activity is probably one of the most challenging problems for the scientific community. Several application domains would benefit of such an analysis, from context-aware computing, to area monitoring and surveillance, to assistive technologies for elderly or disabled, and more. In a broad sense, we can define the activity analysis as the problem of finding an explanation coherent with a set of observations. These observations are typically influenced by several factors from different disciplines, such as sociology or psychology, but also mathematics and physics, making the problem particularly hard. In the last years, also the computer vision community focused its attention on this area, producing the latest advances in the acquisition and understanding of human motion data from image sequences. Despite the increasing effort spent in this field, there still exists a consistent gap between the numerical low-level pixel information that can be observed and measured, and the high abstraction level of the semantic that describes a given activity. In other words, there exist a conceptual ambiguity between the image sequence observations and their possible interpretations. Although several factors are involved, the activity modeling and the comparison strategy play crucial roles. In this proposal, a correlation between activity and corresponding path has been assumed. In light of this, the work carried out tackles two strictly related issues: (i) obtaining a proper representation of human activity; (ii) define an effective tool for reliably measuring the similarity between activity instances. In particular, the object activity is modeled with a signature obtained through a symbolic abstraction of its spatio-temporal trace, allowing the application of particular high-level reasoning for computing the activity similarity. This representation is particularly effective since it provides a smart way to compensate the noise artifacts coming

The widespread use of digital cameras, as well as the increasing popularity of online photo sharing has led to the proliferation of networked photo collections. Handling such a huge amount of media, without imposing complex and time consuming archiving procedures, is highly desirable and poses a number of interesting research challenges to the media community. In particular, the definition of suitable content based indexing and retrieval methodologies is attracting the effort of a large number of researchers worldwide, who proposed various tools for automatic content organization, retrieval, search, annotation and summarization. In this thesis, we will present and discuss three different approaches for content-and-context based retrieval. The main focus will be put on personal photo albums, which can be considered one of the most challenging application domains in this field, due to the largely unstructured and variable nature of the datasets. The methodologies that we will describe can be summarized into the following three points: i. Stochastic approaches to exploit the user interaction in query-by-example photos retrieval. Understanding the subjective meaning of a visual query, by converting it into numerical parameters that can be extracted and compared by a computer, is the paramount challenge in the field of intelligent image retrieval, also referred to as the “semantic gap” problem. An innovative approach is proposed that combines a relevance feedback process with a stochastic optimization engine, as a way to grasp user's semantics through optimized iterative learning providing on one side a better exploration of the search space, and on the other side avoiding stagnation in local minima during the retrieval. ii. Unsupervised event collection, segmentation and summarization. The need for automatic tools able to extract salient moments and provide automatic summary of large photo galleries is becoming more and more important due to the exponential growth in th

In the framework of the synthesis of monopulse array antennas for search-and-track applications, the thesis focuses on the development and the analysis of a method based on the sub-arraying technique aimed at generating an optimal sum and compromise difference patterns through an excitation matching procedure. By exploiting some properties of the solution space, the synthesis problem is reformulated as a combinatorial one to allow a considerable saving of computational resources. Thanks to a graph-based representation of the solution space, the use of an efficient path-searching algorithm to speed-up the convergence of the procedure for the synthesis of large array antennas as well as the use of the Ant Colony Optimizer (ACO) to benefit of its hill-climbing properties in dealing with the non-convexity of the sub-arraying problem are considered. Moreover, a hybrid approach is developed to individually control the level of the secondary lobes. In particular, the sub-array configuration is determined at the first step by exploiting the knowledge of the optimum difference mode coefficients and in the second step, the sub-array weights are computed by means of a quadratic programming procedure. In the numerical validation, a set of representative examples concerned with both pattern matching problems and pattern-feature optimization are reported in order to assess the effectiveness and flexibility of the proposed approach. Comparisons with previously published results are reported and discussed, as well.

In the framework of the synthesis of monopulse array antennas for search-and-track applications, the thesis focuses on the development and the analysis of a method based on the sub-arraying technique aimed at generating an optimal sum and compromise difference patterns through an excitation matching procedure. By exploiting some properties of the solution space, the synthesis problem is reformulated as a combinatorial one to allow a considerable saving of computational resources. Thanks to a graph-based representation of the solution space, the use of an efficient path-searching algorithm to speed-up the convergence of the procedure for the synthesis of large array antennas as well as the use of the Ant Colony Optimizer (ACO) to benefit of its hill-climbing properties in dealing with the non-convexity of the sub-arraying problem are considered. Moreover, a hybrid approach is developed to individually control the level of the secondary lobes. In particular, the sub-array configuration is determined at the first step by exploiting the knowledge of the optimum difference mode coefficients and in the second step, the sub-array weights are computed by means of a quadratic programming procedure. In the numerical validation, a set of representative examples concerned with both pattern matching problems and pattern-feature optimization are reported in order to assess the effectiveness and flexibility of the proposed approach. Comparisons with previously published results are reported and discussed, as well.

In the framework of the synthesis of monopulse array antennas for search-and-track applications, the thesis focuses on the development and the analysis of a method based on the sub-arraying technique aimed at generating an optimal sum and compromise difference patterns through an excitation matching procedure. By exploiting some properties of the solution space, the synthesis problem is reformulated as a combinatorial one to allow a considerable saving of computational resources. Thanks to a graph-based representation of the solution space, the use of an efficient path-searching algorithm to speed-up the convergence of the procedure for the synthesis of large array antennas as well as the use of the Ant Colony Optimizer (ACO) to benefit of its hill-climbing properties in dealing with the non-convexity of the sub-arraying problem are considered. Moreover, a hybrid approach is developed to individually control the level of the secondary lobes. In particular, the sub-array configuration is determined at the first step by exploiting the knowledge of the optimum difference mode coefficients and in the second step, the sub-array weights are computed by means of a quadratic programming procedure. In the numerical validation, a set of representative examples concerned with both pattern matching problems and pattern-feature optimization are reported in order to assess the effectiveness and flexibility of the proposed approach. Comparisons with previously published results are reported and discussed, as well.

In the framework of the synthesis of monopulse array antennas for search-and-track applications, the thesis focuses on the development and the analysis of a method based on the sub-arraying technique aimed at generating an optimal sum and compromise difference patterns through an excitation matching procedure. By exploiting some properties of the solution space, the synthesis problem is reformulated as a combinatorial one to allow a considerable saving of computational resources. Thanks to a graph-based representation of the solution space, the use of an efficient path-searching algorithm to speed-up the convergence of the procedure for the synthesis of large array antennas as well as the use of the Ant Colony Optimizer (ACO) to benefit of its hill-climbing properties in dealing with the non-convexity of the sub-arraying problem are considered. Moreover, a hybrid approach is developed to individually control the level of the secondary lobes. In particular, the sub-array configuration is determined at the first step by exploiting the knowledge of the optimum difference mode coefficients and in the second step, the sub-array weights are computed by means of a quadratic programming procedure. In the numerical validation, a set of representative examples concerned with both pattern matching problems and pattern-feature optimization are reported in order to assess the effectiveness and flexibility of the proposed approach. Comparisons with previously published results are reported and discussed, as well.