Using Computer Vision to See
In this paper we propose a navigation assistant for visually impaired people, which uses computer vision techniques and is integrated on a wearable device. The system makes it possible to detect and recognize, in real-time, both static and dynamic objects
- PDF / 3,149,139 Bytes
- 16 Pages / 439.37 x 666.142 pts Page_size
- 80 Downloads / 203 Views
Telecommunication Department, Faculty of ETTI, University “Politehnica” of Bucharest, Bucharest, Romania 2 ARTEMIS Department, Institut Mines-Telecom/Telecom SudParis, UMR CNRS MAP5 8145, Evry, France {bogdan.mocanu,ruxandra.tapu,titus.zaharia}@telecom-sudparis.eu
Abstract. In this paper we propose a navigation assistant for visually impaired people, which uses computer vision techniques and is integrated on a wearable device. The system makes it possible to detect and recognize, in real-time, both static and dynamic objects existent in outdoor urban scenes without any a priori knowledge about the obstruction type or location. The detection system is based on relevant interest point extraction and tracking, background/camera motion estimation and foreground object identification through motion vectors clustering. The classification method receives as input image patches extracted by the detection module, performs global image representation using binary VLAD and prediction based on SVM. The feedback of our system is transmitted to visually impaired users through bone-conduction headphones as a set of audio warning messages. The entire system is fully integrated on a regular smartphone. The experimental evaluation performed on a set of 20 videos acquired with the help of VI users, demonstrates the pertinence of the proposed methodology. Keywords: Assistive wearable device · Obstacle localization and recognition · Acoustic feedback · Visually impaired users
1
Introduction
For people suffering of visual impairment, common daily activities such as the autonomous navigation to a desired destination, familiar face recognition or independent buying of specific products can represent an important challenge. The safety displacement in outdoor scenario is very difficult because of VI people reduced capacity to understand and perceive the environment, the continuous change of the scene [1] or possible collision with moving objects (e.g. pedestrians, cars, bicycles or animals) or static obstructions (e.g. traffic signs, waste containers, fences, trees, etc.). If for a common setting the position of static hazards can be learned, the location estimation of dynamic obstacles is particularly difficult. In an unknown setting most VI users relay on assistive devices such as the white canes or guiding dogs to acquire additional information about the potential obstructions. The white cane is effective in detecting objects situated directly in c Springer International Publishing Switzerland 2016 G. Hua and H. J´ egou (Eds.): ECCV 2016 Workshops, Part II, LNCS 9914, pp. 375–390, 2016. DOI: 10.1007/978-3-319-48881-3 26
376
B. Mocanu et al.
front of the person and it requires an actual physical contact with obstruction. However, even though the white cane is largely accessible to anyone, it shows quickly its limitations when confronted with real life situations (i.e. it cannot identify further away or overhanging objects, it cannot offer additional information about the type of obstruction and its degree of danger)[2]. Even though the trai
Data Loading...
