Stefano Ghidoni

A New Approach to Urban Pedestrian Detection for Automatic Braking

This paper presents an application of a pedestrian-detection system aimed at localizing potentially dangerous situations under specific urban scenarios. The approach used in this paper differs from those implemented in traditional pedestrian-detection systems, which are designed to localize all pedestrians in the area in front of the vehicle. Conversely, this approach searches for pedestrians in critical areas only. The environment is reconstructed with a standard laser scanner, whereas the following check for the presence of pedestrians is performed due to the fusion with a vision system. The great advantages of such an approach are that pedestrian recognition is performed on limited image areas, therefore boosting its timewise performance, and no assessment on the danger level is finally required before providing the result to either the driver or an onboard computer for automatic maneuvers. A further advantage is the drastic reduction of false alarms, making this system robust enough to control nonreversible safety systems.

TerraMax Vision at the Urban Challenge 2007

This paper presents the TerraMax vision systems used during the 2007 DARPA Urban Challenge. First, a description of the different vision systems is provided, focusing on their hardware configuration, calibration method, and tasks. Then, each component is described in detail, focusing on the algorithms and sensor fusion opportunities: obstacle detection, road marking detection, and vehicle detection. The conclusions summarize the lesson learned from the developing of the passive sensing suite and its successful fielding in the Urban Challenge.

A modular tracking system for far infrared pedestrian recognition

This paper describes a modular tracking system designed to improve the performance of a pedestrian detector. The tracking system consists of two modules, a labeler and a predictor. The former associates a tracking identifier to each pedestrian, keeping memory of the past history; this is achieved by merging the detector and predictor outputs combined with data about vehicle motion. The predictor, basically a Kalman filter, estimates the new pedestrian position by observing his previous movements. Its output helps the labeler to improve the match between the pedestrians detected in the new frame and those observed in the previous shots (feedback). If a pedestrian is occluded by some obstacle for a short while, the system continues tracking its movement using motion parameters. Moreover, it is able to reassign the same tracking ID in case the occlusion disappears in a short time. This behavior helps to correct temporary mis-recognitions that occur when the detector fails. The system has been tested using a quantitative performance evaluation tool, giving promising results.

Different approaches for extracting information from the co-occurrence matrix.

In 1979 Haralick famously introduced a method for analyzing the texture of an image: a set of statistics extracted from the co-occurrence matrix. In this paper we investigate novel sets of texture descriptors extracted from the co-occurrence matrix; in addition, we compare and combine different strategies for extending these descriptors. The following approaches are compared: the standard approach proposed by Haralick, two methods that consider the co-occurrence matrix as a three-dimensional shape, a gray-level run-length set of features and the direct use of the co-occurrence matrix projected onto a lower dimensional subspace by principal component analysis. Texture descriptors are extracted from the co-occurrence matrix evaluated at multiple scales. Moreover, the descriptors are extracted not only from the entire co-occurrence matrix but also from subwindows. The resulting texture descriptors are used to train a support vector machine and ensembles. Results show that our novel extraction methods improve the performance of standard methods. We validate our approach across six medical datasets representing different image classification problems using the Wilcoxon signed rank test. The source code used for the approaches tested in this paper will be available at: http://www.dei.unipd.it/wdyn/?IDsezione=3314&IDgruppo_pass=124&preview=.

Localization and analysis of critical areas in urban scenarios

This paper presents an application of a pedestrian detection system aimed at localizing potentially dangerous situations in specific urban scenarios. The approach used in this work differs from the one implemented in traditional pedestrian detection systems, which are designed to localize all pedestrians appearing in the area in front of the vehicle. This application first locates critical areas in the urban environment, and then it searches for pedestrians in these areas only. The environment is reconstructed with a standard laser scanner system, while the following check for the presence of pedestrians is performed thanks to the fusion with a vision system. The great advantages of such an approach are that pedestrian recognition is performed on a very limited image area -therefore boosting its timing performance- and no assessment on the danger level is finally required before providing the result to either the driver or an on-board computer for automatic manoeuvres.

Performance evaluation of the 1st and 2nd generation Kinect for multimedia applications

Microsoft Kinect had a key role in the development of consumer depth sensors being the device that brought depth acquisition to the mass market. Despite the success of this sensor, with the introduction of the second generation, Microsoft has completely changed the technology behind the sensor from structured light to Time-Of-Flight. This paper presents a comparison of the data provided by the first and second generation Kinect in order to explain the achievements that have been obtained with the switch of technology. After an accurate analysis of the accuracy of the two sensors under different conditions, two sample applications, i.e., 3D reconstruction and people tracking, are presented and used to compare the performance of the two sensors.

A feature-based approach to people re-identification using skeleton keypoints.

In this paper we propose a novel methodology for people re-identification based on skeletal information. Features are evaluated on the skeleton joints and a highly distinctive and compact feature-based signature is generated for each user by concatenating descriptors of all visible joints. We compared a number of state-of-the-art 2D and 3D feature descriptors to be used with our signature on two newly acquired public datasets for people re-identification with RGB-D sensors. Moreover, we tested our approach against the best re-identification methods in the literature and on a widely used public video surveillance dataset. Our approach proved to be robust to strong illumination changes and occlusions. It achieved very high performance also on low resolution images, overcoming state-of-the-art methods in terms of recognition accuracy and efficiency. These features make our approach particularly suited for mobile robotics.

The passive sensing suite of the TerraMax autonomous vehicle

This paper presents the TerraMax autonomous vehicle, which competed in the DARPA Urban Challenge 2007. The sensing system is mainly based on passive sensors, in particular four vision subsystems are used to cover a 360deg area around the vehicle, and to cope with the problems related to complex traffic scenes navigation. A trinocular system derived from the one used during the 2005 Grand Challenge performs obstacle and lane detection, twin stereo systems (one in the front and one in the back) monitor the area close to the truck, two lateral cameras detect oncoming vehicles at intersections, and a rear view system monitors the lanes next to the truck looking for overtaking vehicles. Data fusion between laser scanners and vision will be discussed, focusing on the benefits of this approach.

A Night Vision Module for the Detection of Distant Pedestrians

This paper presents a monocular night vision system specifically developed for detecting very distant pedestrians. The focus of the system is the recognition of pedestrians that are between 40 and 100 m away from the camera. The system is intended to integrate with an existing system, which is capable of detecting pedestrians at distances less than 40 m. At very large distances, pedestrians appear at low resolution, and this requires a specific detection algorithm, rather than an adaptation of an existing one. The presented system performs best in rural environments, where it can locate pedestrians at such great distances, that the pedestrians are hardly visible even to a human driver.

Texture-Based Crowd Detection and Localisation

This paper presents a crowd detection system based on texture analysis. The state-of-the-art techniques based on co-occurrence matrix have been revisited and a novel set of features proposed. These features provide a richer description of the co-occurrence matrix, and can be exploited to obtain stronger classification results, especially when smaller portions of the image are considered. This is extremely useful for crowd localisation: acquired images are divided into smaller regions in order to perform a classification on each one. A thorough evaluation of the proposed system on a real world data set is also presented: this validates the improvements in reliability of the crowd detection and localisation.