Object Identification and Recognition System For Blind Persons

Abstract

This paper helps with the ability to inform blind individuals about things and their spatial palcements, this initiative contibute to the transformation of the visual world into the audible world. The scene’s identifiable objects are recognized by their lables and rendered as speech. Utilizing 3D binarual source modeling, the 2-channel audio is used to encode their spatial positions. The system has many components. A portable camera device will be used to record viedo, which will then sent to the server for real-time object identification and picture recognition (OpenCV).The position and size of the bounding boxes from the detection technique are used to estimate the 3D sound creation program bulit on the unity game engine. With wireless earbuds, the user receives the sound. When the detected item varies from the previous one, or every few seconds, whichever come first, the sound is produced. With the help of the device, the user will able to successfully detect objects that are three to five meters away. Potential issues for the current prototype include information overload when the system tries to tell users of too many things and detect failure when objects are too close or far away