Cataldo Guaragnella

A visual approach for driver inattention detection

Monitoring driver fatigue, inattention, and lack of sleep is very important in preventing motor vehicles accidents. A visual system for automatic driver vigilance has to address two fundamental problems. First of all, it has to analyze the sequence of images and detect if the driver has his eyes open or closed, and then it has to evaluate the temporal occurrence of eyes open to estimate the drivers visual attention level. In this paper we propose a visual approach that solves both problems. A neural classifier is applied to recognize the eyes in the image, selecting two candidate regions that might contain the eyes by using iris geometrical information and symmetry. The novelty of this work is that the algorithm works on complex images without constraints on the background, skin color segmentation and so on. Several experiments were carried out on images of subjects with different eye colors, some of them wearing glasses, in different light conditions. Tests show robustness with respect to situations such as eyes partially occluded, head rotation and so on. In particular, when applied to images where people have eyes closed the proposed algorithm correctly reveals the absence of eyes. Next, the analysis of the eye occurrence in image sequences is carried out with a probabilistic model to recognize anomalous behaviors such as driver inattention or sleepiness. Image sequences acquired in the laboratory and while people were driving a car were used to test the driver behavior analysis and demonstrate the effectiveness of the whole approach.

A new algorithm for ball recognition using circle Hough transform and neural classifier

A large number of methods for circle detection have been studied in the last years for several image processing applications. The context application considered in this work is the soccer game. In the sequences of soccer images it is very important to identify the ball in order to verify the goal event. This domain is a challenging one as a great number of problems have to be faced, such as occlusions, shadows, objects similar to the ball, real-time processing and so on. In this work a visual framework trying to solve the above-stated problems, mainly considering real-time computational aspects, has been developed. The ball detection algorithm has to be very simple in terms of time processing and also has to be efficient in terms of false positive rate. Our framework consists of two sequential steps for solving the ball recognition problem: the first step uses a modified version of the directional circle Hough transform to detect the region of the image that is the best candidate to contain the ball; in the second step a neural classifier is applied on the selected region to confirm if the ball has been properly detected or a false positive has been found. Some tricks like background subtraction and ball tracking have been applied in order to maintain the search of the ball only in limited areas of the image. Different light conditions have been considered as they introduce strong modifications on the appearance of the ball in the image: when the image sequences are taken with natural light, as the light source is strictly directional, the ball, due to self-shades, appears as a spherical cap; this case has been taken in account and the search of the ball has been modified in order to manage this situation. A large number of experiments have been carried out showing that the proposed method obtains a high detection score.

A neural system for eye detection in a driver vigilance application

The problem of eye detection for a driver vigilance system is very important in order to monitor driver fatigue, inattention, and lack of sleep. A neural classifier has been applied to recognize the eyes in the image, selecting the couple of regions candidate to contain the eyes by using iris geometrical information and symmetry. The novelty of this work is that the algorithm works on complex images without constraints on the background, skin color segmentation and so on. Different experiments have been carried out on images of subjects with different eyes colors, some of them wearing glasses. Tests showed robustness with respect to situations such as eyes partially occluded. In particular when applied to images where people have the eyes closed the proposed algorithm correctly reveals the absence of eyes. Eyes tracking in an image sequence is applied to detect eye closure that can be dangerous if persists for a long period.

Motion compensation and multiresolution coding

Abstract Multiresolution techniques have become more and more appealing in current image coding. Image multispectral representation produces many important features such as spectral shaping of coding noise according to human eye perception, good image energy compaction, coder tuning with respect to any band characteristics, and allows for multilevel layered transmission that is one of the main targets pursued by the broadcaster. Despite these appealing capabilities, multiresolution techniques have failed to give the expected results. One of the reasons for this failure is the difficulty of exploiting the temporal redudancy present in image sequences. This paper addresses the problem of motion compensation in a multiresolution environment, considering both QMF-SBC and wavelet transform approaches. Different motion compensation schemes are derived and their efficiency is considered with regard to scalability and to the lengths of subband analysis and synthesis filters. Simulation results are used to support relevant conclusions where needed.

A kinect-based gesture recognition approach for a natural human robot interface

In this paper, we present a gesture recognition system for the development of a human-robot interaction (HRI) interface. Kinect cameras and the OpenNI framework are used to obtain real-time tracking of a human skeleton. Ten different gestures, performed by different persons, are defined. Quaternions of joint angles are first used as robust and significant features. Next, neural network (NN) classifiers are trained to recognize the different gestures. This work deals with different challenging tasks, such as the real-time implementation of a gesture recognition system and the temporal resolution of gestures. The HRI interface developed in this work includes three Kinect cameras placed at different locations in an indoor environment and an autonomous mobile robot that can be remotely controlled by one operator standing in front of one of the Kinects. Moreover, the system is supplied with a people re-identification module which guarantees that only one person at a time has control of the robot. The systems performance is first validated offline, and then online experiments are carried out, proving the real-time operation of the system as required by a HRI interface.

Archaeological trace extraction by a local directional active contour approach

Archaeological trace extraction in aerial or satellite data is a difficult issue for automatic algorithms due to the traces similarity to other image artifacts or to their poor boundary information, discontinuities and so on. We propose in this paper a modified region based active contour approach for archaeological trace identification that overcomes the limits of standard methods of region uniformity and different consistencies with respect to the background. The proposed approach introduces a directional energy model in the minimization of the conventional energy term used in the existing active contour approaches. The local trace direction is estimated automatically after an initial unconstrained evolution of the region. Then, an iterative block based directional procedure has been introduced to limit the application of the modified method to local and adjacent areas and to allow the processing of large images in which the traces may have complex intersections or follow a curved trajectory. Finally, in order to reduce the initialization dependance problem, we propose the use of one seed point for each trace as the initial curve. Tests on the extraction of archaeological traces such as centuriations and ancient roads, visible as crop marks, have demonstrated that the proposed method and the developed MATLAB-based Graphical User Interface (GUI) facilitate unskilled/semi-skilled users in their archaeologic traces mapping operations and improve their detection precisions.

Health Care Improvement: Comparative Analysis of Two CAD Systems in Mammographic Screening

Technological innovations have produced remarkable results in the health care sector. In particular, computer-aided detection (CAD) systems are becoming very useful and helpful in supporting physicians for early detection and control of some diseases such as neoplastic pathologies. In this paper, two different CAD systems able to detect and to localize microcalcification clusters in mammographic images are implemented. The two methods utilize an artificial neural network and a support vector machine, respectively, as classifier. Adopting the MIAS database as procedure testing, the performance of the two implemented systems are compared in terms of sensitivity, specificity, accuracy, free-response operating characteristic curves, and Cohens kappa coefficient. The obtained values for the previous parameters show the efficiency of both methods to operate as “second opinion” in microcalcification cluster detection, improving the screening process efficiency.

A Survey of Automatic Event Detection in Multi-Camera Third Generation Surveillance Systems

Third generation surveillance systems are largely requested for intelligent surveillance of different scenarios such as public areas, urban traffic control, smart homes and so on. They are based on multiple cameras and processing modules that integrate data coming from a large surveillance space. The semantic interpretation of data from a multi-view context is a challenging task and requires the development of image processing methodologies that could support applications in extensive and real-time contexts. This paper presents a survey of automatic event detection functionalities that have been developed for third generation surveillance systems with a particular emphasis on open problems that limit the application of computer vision methodologies to commercial multi-camera systems.

Simple nonlinear dual-window operator for edge detection

We propose a nonlinear edge detection technique based on a two-concentric-circular-window operator. We perform a preliminary selection of edge candidates using a standard gradient and use the dual-window operator to reveal edges as zero-crossing points of a simple difference function depending only on the minimum and maximum values in the two windows. Comparisons with other well-established techniques are reported in terms of visual appearance and computational efficiency. They show that detected edges are surely comparable with Cannys and Laplacian of Gaussian algorithms, with a noteworthy reduction in terms of computational load.

Object oriented motion estimation by sliced-block matching algorithm

An algorithm for object-oriented motion estimation is presented. The algorithm initially determines a macro-block partition on the basis of the computed current frame difference, using the hidden information that a foreground moving object produces high absolute frame difference values in the neighborhood of the object boundaries. An inter-frame coding algorithm, adopting a modified version of classical block matching is then applied on separate slices of each macro-block. Resulting data are used to obtain a preliminary object segmentation. The algorithm further splits each macro-block if characterized by the presence of more than one motion vector into sub-areas. The approach allows us to obtain a global object segmentation-with the possibility of tracking-and lower prediction errors with respect to classical block matching, without increasing the computational complexity.