Humerus anomalies are a problem that requires rapid and accurate diagnosis to ensure immediate and efficient treatment. In this context, the main goal of this paper is to develop and analyze well-known Convolutional Neural Network models for the automatic recognition of humeral fractures, with the aim of proposing a useful tool for healthcare personnel. Specifically, three distinct architectures were implemented and compared: a three-layer untrained neural network, a network based on the ResNet18 architecture and one based on the DenseNet121 model, both of which were trained. The performance analysis highlighted a trade-off between accuracy and generalization ability, showing better accuracy in the pre-trained models - in particular, the DenseNet121 model achieved optimal accuracy across multiple runs of 85%. - which however proved more prone to suffer from overfitting compared to the non-pre-trained model. As a result, this study aims to propose the integration of deep learning tools in medical practice, laying important foundations for future developments, with the hope of improving the efficiency and accuracy of orthopedic diagnoses.

Modeling and automatically detecting complex events in different domains, such as video surveillance and healthcare, is becoming an increasingly topical issue nowadays. In fact, deriving knowledge on higher level from low-level events by combining the latter to complex structures is the task of an Event Query Language (EQL), whose main issue is the lack of formal semantics. Consequently, in order to cope with this issue, in this paper we propose $$ISEQL+$$ , an extension of ISEQL (an Interval-based Surveillance Event Query Language, that we previously defined), aimed at further improving its expressiveness. More specifically, we provide formal proofs demonstrating that the language fully covers the well-known Allen’s interval relationships, additionally supports conditional overlap ratio and conditional cardinality constraints over the interval relationships, provides robustness with respect to small variations in the intervals, and can be formalized as relational algebra extension, which will in turn allow a very efficient implementation exploiting an existing algorithm. Eventually, we also show how typical events in the healthcare domain can be easily expressed via $$ISEQL+$$ .

In this demo paper, we present SPARK (Semantic Planning with Augmented Retrieval and Knowledge), an AI-driven web application that generates personalized itineraries with natural language input. By addressing the limitations of traditional routing systems and machine learning models that lack adaptability to user intent and real-time contexts, SPARK integrates a Neo4j [6] graph that built on OpenStreetMap [8] data with semantic analysis, custom route optimization, and retrieval-augmented generation. By using large language models (LLM), the system outputs travel planning with real-time and enriched data via external APIs. Also, SPARK dynamically adapts to personalized user queries and optimizes routes based on contextual relevance. As demonstrated in an urban scenario, this demo paper highlights the potential of combining graph-based retrieval and LLMs to deliver flexible and context-aware route planning. Meanwhile, our demo shows that the proposed system is scalable and deployable for travel guidance.

Telemedicine has gained increasing relevance, particularly after COVID-19, emphasizing the need for digital tools to support clinicians and enhance access to care. In such a context, this paper presents a system for automated nevus and melanoma analysis using two methods: a pre-trained ResNet-50 CNN for dermoscopic image feature extraction, and a LinearSVC classifier for structured clinical data collected by medical doctors. The framework was evaluated on 9,024 dermoscopic images and 200 clinically documented lesions; the results demonstrate how machine learning can support skin lesion diagnosis, improving efficiency and accessibility.

The technique of recording and evaluating (processing) streams of data about occurrences and drawing conclusions from them is known as Complex Event Processing, or CEP. In recent years, due to ever-increasing security concerns, CEP is being applied increasingly broadly in heterogeneous domains in order to detect potentially dangerous events early or even prevent their occurrence; possible examples are bank robberies in the context of video surveillance, or identity theft in the domain of social network analysis. For such reasons, in this paper we propose a survey reporting the most interesting and innovative approaches to complex event processing in heterogeneous domains; specifically, here we focus on video analysis, social network analysis, and healthcare, classifying the papers into different subcategories and highlighting their advantages and disadvantages.

Abstract This paper describes an intelligent ecosystem that can continuously monitor patients’ health conditions, whether at home, at work or during recreational activities, by leveraging a creative blend of wearable medical devices, intelligent agents (IA) and complex event processing (CEP). With the help of a smart application, linking wearable devices and the power of IA and CEP, patients will be constantly and actively supervised during their daily activities. This can even save their lives in case they experience sudden or gradual problems. Thanks to our system, patients with chronic illnesses that are not serious but potentially unstable will no longer overburden first aid services. This is also helpful in containing the spread of COVID-19. Specifically, in this paper, we focus on automatic monitoring of vital parameters, electrocardiogram analysis and psoriasis detection. Experimental results conducted on real patients show how promising our approach is.

Our work describes a smart-ecosystem able to mon-itor patients' health condition, even at home or at work, by ex-ploiting a creative blend of Medical Wearables, Intelligent Agents, Complex Event Processing and Image Processing. With the help of a smart application, that links together the Wearables and the power of Artificial Intelligence, patients will be continuously and actively supervised during their daily activities. This can even save their lives, in case sudden or gradual issues should occur. Using our system, patients with non-severe though potentially unstable chronic diseases will no longer overburden first aid services. This is also useful for containing the spread of COVID-19. Specifically, in this paper we focus on automated vitals monitoring, electrocardiogram (ECG) analysis, and Psoriasis detection.

For Point of Interest (POI) recommendations or touristic orienteering applications, it is essential to explore the user preference or potential interests to predict potential POIs or touristic routes. One of the state-of-the-art approaches for inferring the user’s interests is social sensing, which is based on implicit feedback and semantic similarities to extract the user preference. In this paper, we profile the user preferences by social sensing to exploit the similarity between users’ reviews and POI descriptions. The experiment is based on the "Yelp!" dataset and provides the labels associated with different businesses considered in the "Yelp!" Social Network. The preliminary results show that the proposed model can automatically estimate the user interests and improve the effectiveness of the user profiling procedure. The results also indicate that our proposed model can offer a foundational approach for ranking POIs and touristic routes.

Abstract We develop a family of efficient plane-sweeping interval join algorithms for evaluating a wide range of interval predicates such as Allen’s relationships and parameterized relationships. Our technique is based on a framework, components of which can be flexibly combined in different manners to support the required interval relation. In temporal databases, our algorithms can exploit a well-known and flexible access method, the Timeline Index, thus expanding the set of operations it supports even further. Additionally, employing a compact data structure, the gapless hash map, we utilize the CPU cache efficiently. In an experimental evaluation, we show that our approach is several times faster and scales better than state-of-the-art techniques, while being much better suited for real-time event processing.

Manga are one of the most popular forms of comics consumed on a global level. Unfortunately, this kind of media was not designed for digital consumption, and consequently its format does not fit well into small areas, such as smartphone screens. In order to cope with this issue, in this paper we propose a novel approach to comics segmentation and sequencing by taking advantage of existing machine learning concepts which are used to generate an artificial intelligence (AI) capable of correctly detecting panels within an image. The regions proposed by the AI are then used to generate a grid that acts as anchor points for a mobile application guiding the reader during navigation and enabling full Manga responsiveness. The developed approach achieves overall better performances in terms of precision and recall, as well as higher fault tolerance than state-of-the-art approaches. The reliability of this method is also considered largely satisfactory for real-world scenarios, so that we are about to finalize an app implementing the method to be spread soon; additionally, future work will be devoted to generalize our approach to all the comics formats.

We present a framework for high-level event detection in video streams based on a novel temporal extension of relational algebra. With the help of intuitive and interactive graphical user interfaces, a user can have a look at the different layers of our system to gain insights into the inner workings of the system, as well as create new events on the fly and track their processing through the system. As a proof-of-concept we have predefined events on three video surveillance data sets, but we also plan to run a demo with a live video stream generated by a local webcam.

In this paper, we present a framework supporting information retrieval over corpora of documents using an automatic sematic query expansion approach. The main idea is to expand the set of words used as query terms exploiting the notion of semantic similarity between the concepts related to the search terms. We leverage existing lexical resources and similarity metrics computed among terms to generate - by a proper mapping into a vectorial space - an index for the fast retrieval of a set of terms "semantically correlated" to a given query term. The vector of expanded terms is then exploited in the query stage to retrieve documents that are significantly related to specific combinations of the query terms. Preliminary experimental results concerning efficiency and effectiveness of the proposed approach are reported and discussed.

Reasoning with temporal data has attracted the attention of many researchers from different backgrounds including artificial intelligence, database management, computational linguistics and biomedical informatics. More specifically, activity detection is a very important problem in a wide variety of application domains such as video surveillance, cyber security, fault detection, but also clinical research. Thus, in this paper we present a prototype architecture designed and developed for activity detection in the medical context. In more detail, we first acquire data in real time from a cricothyrotomy simulator, when used by medical doctors, then we store the acquired data into a scientific database and finally we use an Activity Detection Engine for finding expected activities, corresponding to specific performances obtained by the medical doctors when using the simulator. Some preliminary experiments using real data show the approach efficiency and effectiveness. Eventually, we also received positive feedbacks by the medical personnel who used our prototype.

There are numerous applications where we wish to discover unexpected activities in a sequence of time-stamped observation data--for instance, we may want to detect inexplicable events in transactions at a website or in video of an airport tarmac. In this paper, we start with a known set $({\cal A})$ of activities (both innocuous and dangerous) that we wish to monitor. However, in addition, we wish to identify "unexplained" subsequences in an observation sequence that are poorly explained (e.g., because they may contain occurrences of activities that have never been seen or anticipated before, i.e., they are not in $({\cal A})$). We formally define the probability that a sequence of observations is unexplained (totally or partially) w.r.t. $({\cal A})$. We develop efficient algorithms to identify the top-$(k)$ Totally and partially unexplained sequences w.r.t. $({\cal A})$. These algorithms leverage theorems that enable us to speed up the search for totally/partially unexplained sequences. We describe experiments using real-world video and cyber-security data sets showing that our approach works well in practice in terms of both running time and accuracy.

The extraordinary technological progress we have witnessed in recent years has made it possible to generate and exchange multimedia content at an unprecedented rate. As a consequence, massive collections of multimedia objects are now widely available to a large population of users. As the task of browsing such large collections could be daunting, Recommender Systems are being developed to assist users in finding items that match their needs and preferences. In this article, we present a novel approach to recommendation in multimedia browsing systems, based on modeling recommendation as a social choice problem. In social choice theory, a set of voters is called to rank a set of alternatives, and individual rankings are aggregated into a global ranking. In our formulation, the set of voters and the set of alternatives both coincide with the set of objects in the data collection. We first define what constitutes a choice in the browsing domain and then define a mechanism to aggregate individual choices into a global ranking. The result is a framework for computing customized recommendations by originally combining intrinsic features of multimedia objects, past behavior of individual users, and overall behavior of the entire community of users. Recommendations are ranked using an importance ranking algorithm that resembles the well-known PageRank strategy. Experiments conducted on a prototype of the proposed system confirm the effectiveness and efficiency of our approach.

Consider a video surveillance application that monitors some location. The application knows a set of activity models (that are either normal or abnormal or both), but in addition, the application wants to find video segments that are unexplained by any of the known activity models - these unexplained video segments may correspond to activities for which no previous activity model existed. In this paper, we formally define what it means for a given video segment to be unexplained (totally or partially) w.r.t. a given set of activity models and a probability threshold. We develop two algorithms - FindTUA and FindPUA - to identify Totally and Partially Unexplained Activities respectively, and show that both algorithms use important pruning methods. We report on experiments with a prototype implementation showing that the algorithms both run efficiently and are accurate.

In the classical theory of social choice, a set of voters is called to rank a set of alternatives and a social ranking of the alternatives is generated. In this paper, we model recommendation in the context of browsing systems as a social choice problem, where the set of voters and the set of alternatives both coincide with the set of objects in the data collection. We then propose an importance ranking method that strongly resembles the well known PageRank ranking system, and takes into account both the browsing behavior of the users and the intrinsic features of the objects in the collection. We apply the proposed approach in the context of multimedia browsing systems and show that it can generate effective recommendations and can scale well for large data collections.