Title: Video-based assistive aid for blind people using object recognition in dissimilar frames

Authors: Hanen Jabnoun; Faouzi Benzarti; Frédéric Morain-Nicolier; Hamid Amiri

Addresses: Tunis El Manar University, National Engineering School of Tunis, LRSITI, 1002, Tunis, Tunisia ' Tunis El Manar University, National Engineering School of Tunis, LRSITI, 1002, Tunis, Tunisia ' Reims Champagne Ardenne University, STIC Research Center, University Institute of Technology of Troyes, 1000, Troyes, France ' Tunis El Manar University, National Engineering School of Tunis, LRSITI, 1002, Tunis, Tunisia

Abstract: Developing visual aids for handicapped persons is an active research area in the computer vision community. This paper presents a visual substitution tool for blind people based on object recognition in video scene. It focuses on the optimisation of the video processing using the calculation of dissimilarity between frames. The approach includes the real valued local dissimilarity map method in the frames dissimilarity measures. It uses scale invariant features transform keypoints extraction and matching to identify objects in dissimilar frames. The experiment tests showsome encouraging results in the case of finding object of interest. Thus, the proposed method can be a choice for solving the problem of blind and disabilities persons in their interaction with the surrounding environment.

Keywords: pattern recognition; video processing; visual substitution system; scale invariant features transform; SIFT; real valued local dissimilarity map; keypoints matching; features extraction; intelligent system; blind people; object detection.

DOI: 10.1504/IJAIP.2019.102967

International Journal of Advanced Intelligence Paradigms, 2019 Vol.14 No.1/2, pp.122 - 139

Received: 15 Jul 2016
Accepted: 21 Nov 2016
Published online: 14 Oct 2019 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article