Papers by Roberto Pirrone
Vision Systems: Applications, 2007
The present work investigates the problem of determining a learning path inside a suitable domain... more The present work investigates the problem of determining a learning path inside a suitable domain ontology. The proposed approach enables the user of a web learning application to interact with the system using natural language in order to browse the ontology itself. The course related knowledge is arranged as a three level hierarchy: content level, symbolic level, and conceptual level bridging the previous ones.
International Journal of Intelligent Systems, 2000
A system for part-based segmentation of range data and their interpretation as a composition of d... more A system for part-based segmentation of range data and their interpretation as a composition of deformable superquadrics is described. Segmentation and reconstruction phases are performed using the same algorithm at different scales. First, the data set is partitioned in regions corresponding approximately to simple convex objects, and then single deformable models are fitted to each region. Refinements of the model
Lecture Notes in Computer Science, 2002
A novel, two stage, neural architecture for the segmentation of range data and their modeling wit... more A novel, two stage, neural architecture for the segmentation of range data and their modeling with undeformed superquadrics is presented. The system is composed by two distinct neural networks: a SOM is used to perform data segmentation, and, for each segment, a multilayer feed-forward network performs model estimation.
Lecture Notes in Computer Science, 2005
Abstract. A very important artifact corrupting Magnetic Resonance (MR) Images is the RF inhomogen... more Abstract. A very important artifact corrupting Magnetic Resonance (MR) Images is the RF inhomogeneity, also called Bias artifact. The vi- sual effect produced by this kind of artifact is an illumination variation which afflicts this kind of medical images. In literature a lot of ...
Lecture Notes in Computer Science, 2003
A novel, two stage, neural architecture for the segmentation of range data and their modeling wit... more A novel, two stage, neural architecture for the segmentation of range data and their modeling with undeformed superquadrics is presented. The system is composed by two distinct neural networks: a SOM is used to perform data segmentation, and, for each segment, a multi-layer feed-forward network performs model estimation. The topology-preserving nature of the SOM algorithm makes this architecture suited to cluster data with respect to sudden curvature variations. The second stage is designed to model and compute the inside-outside function of an undeformed superquadric in whatever attitude, starting form the (x, y, z) data triples. The network has been trained using backpropagation, and the weights arrangement, after training, represents a robust estimate of the superquadric parameters. The whole architectural design is general, it can be extended to other geometric primitives for part-based object recognition, and performs faster than classical model fitting techniques. Detailed explanation of the theoretical approach, along with some experiments with real data are reported.
12th International Conference on Image Analysis and Processing, 2003.Proceedings., 2003
A novel approach to the detection of multiple sclerosis (MS) lesions is presented, which uses an ... more A novel approach to the detection of multiple sclerosis (MS) lesions is presented, which uses an adaptive formulation of the anisotropic diffusion and fuzzy-c-means (FCM) clustering. As opposite to previous works of the same authors, FCM runs only on PD weighted slices that, for each examination, are composed in a unique data set. Images are preprocessed with an anisotropic diffusion filter whose diffusion function has been adaptively optimized to aggregate pixels belonging to lesions and cut off all the others. Adaptivity is used to achieve significant noise reduction. Detailed description of the proposed approach is presented, along with first experimental results.
IEEE EUROCON 2009, 2009
Abstract This paper presents two methods aimed to the optic disc positioning on the retinal image... more Abstract This paper presents two methods aimed to the optic disc positioning on the retinal images. The first is based on a template of the entire image, while the second is based on the retinal vascularisation. For this second task, also a vessel extraction method is provided. Both the algorithm have been tested on test images of the drive database.
Biologically Inspired Cognitive Architectures, 2013
The behavior of an artificial agent performing in a natural environment is influenced by many dif... more The behavior of an artificial agent performing in a natural environment is influenced by many different pressures and needs coming from both external world and internal factors, which sometimes drive the agent to reach conflicting goals. At the same time, the interaction between an artificial agent and the environment is deeply affected by uncertainty due to the imprecision in the description of the world, and the unpredictability of the effects of the agent's actions. Such an agent needs meta-cognition in terms of both self-awareness and control. Self-awareness is related to the internal conditions that may possibly influence the completion of the task, while control is oriented to performing actions that maintain the internal model of the world and the perceptions aligned. In this work, a general meta-cognitive architecture is presented, which is aimed at overcoming these problems. The proposed architecture describes an artificial agent, which is capable to combine cognition and meta-cognition to solve problems in an uncertain world, while reconciling opposing requirements and goals. While executing a plan, such an agent reflects upon its actions and how they can be affected by its internal conditions, and starts a new goal setting process to cope with unforeseen changes. The work defines the concept of ''believability'' as a generic uncertain quantity, the operators to manage believability, and provides the reader with the u-MDP that is a novel MDP able to deal with uncertain quantities expressed as possibility, probability, and fuzziness. A couple u-MDPs are used to implement the agent's cognitive and meta-cognitive module. The last one is used to perceive both the physical resources of the agent's embodiment and the actions performed 2212-683X/$ -see front matter ª A v ai l abl e a t w w w . s c i e n c e d i r e c t. c o m j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / b i c a by the cognitive module in order to issue goal setting and re-planning actions.
Advances in Intelligent Systems and Computing, 2013
Multistage decision-making in robots involved in real-world tasks is a process affected by uncert... more Multistage decision-making in robots involved in real-world tasks is a process affected by uncertainty. The effects of the agent's actions in a physical environment cannot be always predicted deterministically and in a precise manner. Moreover, observing the environment can be a too onerous for a robot, hence not continuos. Markov Decision Processes (MDPs) are a well-known solution inspired to the classic probabilistic approach for managing uncertainty. On the other hand, including fuzzy logics and possibility theory has widened ...
Lecture Notes in Computer Science, 2011
ABSTRACT For the preceding conference see Zbl 1175.68021.
Lecture Notes in Computer Science, 1993
Neural architectures may offer an adequate way to deal with early vision since they are able to l... more Neural architectures may offer an adequate way to deal with early vision since they are able to learn shape features or classify unknown shapes, generalising the features of a few meaningful examples, with a low computational cost after the training phase. Two different neural approaches are proposed by the authors: the first one consists of a cascaded architecture made up by a first stage named BWE (Boundary Webs Extractor) which is aimed to extract a brightness gradient map from the image, followed by a backpropagation ...
Lecture Notes in Computer Science, 1993
Shape from Shading is perhaps the most difficult topic to deal with in Artificial Vision: several... more Shape from Shading is perhaps the most difficult topic to deal with in Artificial Vision: several researchers have faced it using different approaches. The most part of these methods are based on the Horn algorithm so they require very heavy regularity assumptions about the perceived objects' shape and are computationally expensive.
Robotics and Autonomous Systems, 2001
An autonomous robot involved in long and complex missions should be able to generate, update and ... more An autonomous robot involved in long and complex missions should be able to generate, update and process its own plans of action. In this perspective, it is not plausible that the meaning of the representations used by the robot is given form the outside of the system itself. Rather, the meaning of internal symbols must be firmly anchored to the world through the perceptual abilities and the overall activities of the robot. According to these premises, in this paper we present an approach to action representation, that is based on a "conceptual" level of representation, acting as an intermediate level between symbols and data coming from sensors. Symbolic representations are interpreted by mapping them on the conceptual level through a mapping mechanism based on artificial neural networks. Examples of the proposed framework are reported, based on experiments performed on a RWI-B12 autonomous robot.
Journal of Intelligent Systems, 2000
A system for part-based segmentation of range data and their interpretation as a composition of d... more A system for part-based segmentation of range data and their interpretation as a composition of deformable superquadrics is described. Segmentation and reconstruction phases are performed using the same algorithm at different scales. First, the data set is partitioned in regions corresponding approximately to simple convex objects, and then single deformable models are fitted to each region. Refinements of the model can be achieved by recursively applying the method. The proposed Moving Target (MT) algorithm is an original variation of the well known approach by Solina, ideated to avoid the classical inconvenient of the minimization procedure where the solution escapes the global minimum and/or gets stuck in a local one. Motivations of the proposed approach along with its theoretical formulation are presented and discussed in detail. The whole system has been validated using a several range images.
Journal of Clinical Monitoring and Computing, 2006
Computer Methods and Programs in Biomedicine, 2008
c o m p u t e r m e t h o d s a n d p r o g r a m s i n b i o m e d i c i n e 9 2 ( 2 0 0 8 ) 35-... more c o m p u t e r m e t h o d s a n d p r o g r a m s i n b i o m e d i c i n e 9 2 ( 2 0 0 8 ) 35-53 Bias artifact Illumination correction MR image Homomorphic filter a b s t r a c t RF-inhomogeneity correction is a relevant research topic in the field of magnetic resonance imaging (MRI). A volume corrupted by this artifact exhibits nonuniform illumination both inside a single slice and between adjacent ones. In this work a bias correction technique is presented, which suppresses this artifact on MR volumes scanned from different body parts without any a priori hypothesis on the artifact model. Theoretical foundations of the method are reported together with experimental results and a comparison is presented with both the 2D version of the algorithm and other techniques that are widely used in MRI literature. (O. Gambino).
A novel neural architecture aimed to estimate superquadrics parameters form range data is present... more A novel neural architecture aimed to estimate superquadrics parameters form range data is presented. The network topology is designed to model and compute the inside-outside function of an undeformed superquadric in whatever attitude, starting from the (x, y, z) data triples. The network has been trained using backpropagation, and the weights arrangement, after training, represents a robust estimate of the superquadric parameters. The architectural approach is general, it can be extended to other geometric primitives for ...
Uploads
Papers by Roberto Pirrone