AI Based Object Classifier For Blind People With Voice

Abstract

As we can see, there are numerous blind persons 
nearby who encounter various challenges, such as 
difficulty in crossing roads and identifying objects 
in their environment. With the advancement in 
technology in several fields, human life is evolving 
to better standards. Unfortunately, those who are 
blind are unable to fully enjoy this kind of lifestyle. 
So, this project is one strategy for introducing blind 
individuals to a new way of living that makes them 
independent on others. The major goal of this 
project is to create a deep-learning algorithm that 
can be used to analyse the environment for people 
who are blind by using the rapidly evolving 
technology. We\'ll accomplish this using object 
detection and transform the data into speech alerts 
and warnings. Real-time object detection is one of 
the more challenging tasks since it requires 
continuous processing and takes a long time. The 
convolution neural network is the main backbone 
for any type of object detection (CNN). We can 
create algorithms based on photos and videos by 
employing a convolution neural network. We 
utilise the YOLO technique for object detection 
because it is simple and quick to process. In 
addition, for the voice warnings, we employed Text 
to Speech (TTS). The dataset used in this technique 
is the COCO dataset, which contains the names of 
things and objects in our daily lives. These 
algorithms have been thoroughly trained by the 
over 90 outdoor objects that we view every day in 
our daily lives.