Publications by Rita Cucchiara

In this work, we present a general purpose systemfor hierarchical structural segmentation and automaticannotation of video clips, by means of standardizedlow level features. We propose to automatically extractsome prototypes for each class with a context basedintra-class clustering. Clips are annotated followingthe MPEG-7 standard directives to provide easierportability. Results of automatic annotation and semiautomaticmetadata creation are provided.

In this work, we describe in detail the new MPEG-7 Perceptual 3D Shape Descriptor and provide a set of tests with different 3D objects databases, mainly with the Princeton Shape Benchmark. With this purpose we created a function library called Retrieval-3D and fixed some bugs of the MPEG-7 eXperimentation Model (XM). We explain how to match the Attributed Relational Graph (ARG) of every 3D model with the modified nested Earth Mover’s Distance (mnEMD). Finally we compare our results with the best found in literature, including the first MPEG-7 3D descriptor, i.e. the Shape Spectrum Descriptor.

This paper presents multimedia ontologies, where multimedia data and traditional textual ontologies are merged. A solution for their implementation for the soccer video domain and a method to perform automatic soccer video annotation using these extended ontologies is shown. HSV is a widely adopted space in image and video retrieval, but its quantization for histogram generation can create misleading errors in classification of achromatic and low saturated colors. In this paper we propose an Enhanced HSV Histogram with achromatic point detection based on a single Hue and Saturation parameter that can correct this limitation.The more general concepts of the sport domain (e.g. play/break, crowd, etc.) are put in correspondence with the more general visual features of the video like color and texture, while the more specific concepts of the soccer domain (e.g. highlights such as attack actions) are put in correspondence with domain specific visual feature like the soccer playfield and the players. Experimental results for annotation of soccer videos using generic concepts are presented.

To enhance video surveillance systems, multi-modal sensor integration can be a successful strategy. In this work, a computer vision system able to detect and track people from multiple cameras is integrated with a wireless sensor network mounting passive Pyroelectric InfraRed sensors. Thetwo subsystems are briefly described and possible cases in which computer vision algorithms are likely to fail are discussed. Then, simple but reliable outputs from the sensor nodes are exploited to improve the accuracy of the vision system. In particular, two case studies are reported: the first uses the presence detection of sensors to disambiguate between an open door and a moving person, while the second handles motion direction changes during occlusions. Preliminary results are reported and demonstrate the usefulness of the integration of the two subsystems.

In this work, we focus on two aspects of the comparison of video shots. We present a new approach to extract a variable number of key frames from a shot, by the use of a hierarchical clustering with automatic level selection, in order to provide optimal allocation of features on different parts of the shot. We then employ the Mallows distance as an effective technique to compare the discrete distributions of features, independently from the features selected for the specific application. Results and comparisons on a soccer documentary video are provided.

The present work shows a system for compressing and streaming of live videos over networks with low bandwidths (radio mobile networks), with the objective to design an effective solution for mobile video access. We present a mobile ready-to-use streaming system, that encodes video using h264 codec (offering good quality and frame rate at very low bit-rates) and streams it over the network using UDP protocol. A dynamic frame rate control has been implemented in order to obtain the best trade off between playback fluency and latency. © Springer-Verlag Berlin Heidelberg 2007.

Aim of the Visor Project [1] is to gather and makefreely available a repository of surveillance andvideo footages for the research community onpattern recognition and multimedia retrieval. Thegoal is to create an open forum and a free repositoryto exchange, compare and discuss results of manyproblems in video surveillance and retrieval.Together with the videos, the repository containsmetadata annotation, both manually annotated asground-truth and automatically obtained by videosurveillance systems. Annotation refers to a largeontology of concepts on surveillance and securityrelated objects and events. The ontology has beendefined including concepts from LSCOM andMediaMill ontologies. As well as videos andannotations, Visor provides tools for enriching theontology, annotating new videos, searching bytextual queries, composing and downloading videos.

In this paper we present a distributed videosurveillanceframework. Our end is the remote monitoringof the behavior of people moving in a scene exploitinga virtual reconstruction on low capabilitiesdevices, like PDAs and cell phones. The main noveltyof this system is the effective integration of the computervision and computer graphics modules. The first,using a probabilistic frameworks, can detect the position,the trajectory and the posture of peoples movingin the scene. The second exploits the new possibility ofboth standard 3D graphics libraries on mobile (namelyJSR184 and M3G graphic format) and new PDAsprocessing capability in order to reconstruct the remotesurveillance data in real-time.

Distributed video surveillance has a direct application in intelligent home automation or domotics (from the Latin word domus, that means “home”, and informatics); in particular, in-house videosurveillance can provide good support for people with some difficulties (e.g., elderly or disabled people) living alone and with a limited autonomy. New hardware technologies for surveillance are now affordable and provide high reliability. Problems related to reliable software solutions are not completely solved, especially concerning the application of general-purpose computer vision techniques in indoor environments. Indeed, assuming the objective is to detect the presence of people, track them, and recognize dangerous behaviours by means of abrupt changes in their posture, robust techniques must cope with non-trivial difficulties. In particular, luminance changes and shadows must be taken into account, frequent posture changes must be faced, and large and long-lasting occlusions are common due to the vicinity of the cameras and the presence of furnitureand doors that can often hide parts of the person’s body. These problems are analyzed and solutions based on background suppression, appearance-based probabilistic tracking, and probabilistic reasoning for posture recognition are described.

Pixel classifiers are often adopted in pattern recognition as a suitable method for image segmentation. A common approach to the performance evaluation of classifier systems is based on the measurement of the classification errors and, at the same time, on the computational time. In general, multiclassifiers have proven to be more precise in the classification in many applications, but at the cost of a higher computational load. This paper analyzes different classifiers and proposes an evaluation of the classifiers in the case of semi-automatic processes with human interaction. Medical imaging is a typical application, where automatic or semi-automatic segmentation can be a valuable support to the diagnosis. The paper focuses on the segmentation of cardiac images of fruit flies (genetic model for analyzing human heart's diseases). Analysis is based on M-modes, that are gray-level images derived from mono-dimensional projections of the video frames on a line. Segmentation of the M-mode images is provided by classifiers and integrated in a multiclassifier. A neural network classifier, a Bayesian classifier, and a classifier based on hidden Markov chains are joined by means of a Behavior Knowledge Space fusion rule. The comparative evaluation is discussed in terms of both accuracy and required time, in which the time to correct the classifier errors by means of human intervention is also taken into account.