No abstract

Technological advances are changing every aspect of our lives, from the way we work, to how we learn and communicate. Advanced driver assistance systems (ADAS) have seen an increased interest due to the potential of ensuring a safer environment for all road users. This study investigates the use of a smartphone-based ADAS in terms of driving performance and driver acceptance, with the aim of improving road safety. The mobile application uses both cameras of a smartphone to monitor the traffic scene and the driver’s head orientation, and offers an intuitive user interface that can display information in a standard mode or in augmented reality (AR). A real traffic experiment consisting of two driving conditions (a baseline scenario and an ADAS scenario), was conducted in Brasov, Romania. Objective and subjective data were recorded from twenty-four participants with a valid driver’s license. Results showed that the use of the ADAS influences the driving performance, as most of them adopted an increased time headway and lower mean speeds. The technology acceptance model (TAM) questionnaire was used to assess the users’ acceptance of the proposed driver assistance system. The results showed significant interrelations between acceptance factors, while the hierarchical regression analysis indicates that the variance of behavioral intention (BI) can be predicted by attitude toward behavior.

Abstract-This paper describes an experimental system that has been designed, implemented and tested for object recognition and tracking in still, respectively dynamic imagessuccessive video frames captured in real time (live) with a web camera -based on Intel's open source computer vision functions library, OpenCV (Open Source Computer Vision). We propose a real-time object recognition system in intelligent library environments. The system consists of two key modules: feature extraction and object recognition. Specific detectors such as SIFT (Scale-Invariant Feature Transform) and SURF (Speeded Up Feature Robust) are efficient methods that provide high quality features, yet are too computational for use in real-time applications.This paper also proposes a low complexity, robust object recognition and tracking method using advanced real time feature matching. It combines Microsoft Visual Studio 2008 Express Edition C# with OpenCV Function library using SURF algorithm in Emgu CV to develop the software. The tests showed that the proposed system and method are more efficient and more robust than in most traditional applications.Index Terms-Computer vision, object recognition, object tracking, OpenCV, SURF. I. INTRODUCTIONResearch in the field of library automation is characterized by a long history of robot assistance applications such as book cataloguing, retrieval and return [1], [2]. The studies focused on solving the technical problems at fixed locations in the library. Mobile robots [3], [4] which are both autonomous and capable to detect and react to environmental factors have recently attracted the attention of libraries. Some libraries have introduced mobile robots and use them to guide student users towards the adequate shelf [5], [6]. However, there are only a few cases in which educational research has addressed the field of service robot applications [7]-[9] designed for use in libraries. These studies have emphasized the general trend which regards robots as adequate agents to ensure user guidance and involvement in learning activities.References vision due to its numerous applications.Object recognition and tracking are major tasks in several computer vision applications, such as Augmented Reality (AR), interactive systems and robotic systems.Some of the main advantages of vision systems, in robotic applications, are the simplicity of the algorithm, the low cost, and the reduced need for maintenance, while aspects such as fast and effective identification are still unsolved. Even though adequately efficient and accurate algorithms have been developed, the processing speed still fails to meet the modern manufacturing requirements [18].Most of the objects tracking approaches based on feature matching are highly computational and less robust in various environments. To efficiently track an object in a video sequence, at first, feature points are extracted from the object of interest. The extracted objects then recognize the target object, and the detected object is continuously tracked on the input stre...