Fixed DO Solfège based object detection and positional analysis for the visually impaired

Abstract

This paper proposes a novel approach towards object detection and its subsequent positional analysis for a visually impaired subject, using Fixed DO Solfège, basically sounds that we have heard multiple times, and we are aware of them and also their sequence is thus naturally perceived by us viz. DO-RE-MI-FA-SO-LA-TI This approach utilizes a concept of virtual zones superimposed on the viewport of a small wearable CMOS camera module mounted on the eyeglasses of the subject, communicating to a computing device with attached earplugs for sonic feedbacks in the Fixed DO Solfège notation. By implantation of a HAAR cascade based classifier trained system, several need to know objects are trained and fed into the recognition pool, which are thus detected in the viewport overlaid by the virtual harmonic zones further linked to the Fixed DO Solfège notations, and are mapped throughout the viewport in an incremental/sequential manner. As the subject moves his hand through the viewport, and his hands overlap the recognized objects, an audible beep basis on the sound of the zone is played on the earplugs. This not only enables the subject to know that in which direction a particular object is situated, but also, because of the sound, he can also know as to how many objects lie before/after the current object, this gives a sense of relative recognition and positional cognition of the objects.