Blind Assistance System using Tensorflow

Abstract

Because of item recognition's cozy relationship with video investigation and picture understanding, it has drawn in much exploration consideration lately. Customary item discovery techniques are based on carefully assembled highlights and shallow teachable designs. Their exhibition effectively deteriorates by building complex outfits which join numerous low-level picture highlights with undeniable level setting from object finders and scene classifiers. With the fast improvement in profound learning, all the more amazing assets, which can learn semantic, significant level, further highlights, are acquainted with address the issues existing in conventional models. These models act contrastingly in network engineering, preparing procedure and improvement work, and so forth. In this paper, we provide a review on deep learning based object detection frameworks and how it can be used to help the visually impaired. Our audit starts with a short presentation on the historical backdrop of profound learning and its delegate apparatus, to be specific Convolutional Neural Network (CNN). We have used Tensorflow made by google which makes this task even faster and easier to implement. We have chose to dedicate this for the blind because we think they are the most neglected when it comes to project helping people with some form of disability and this group of people are at more risk of doing everyday tasks as they don’t know what is in front of them and also they can be fooled easily for example someone can give them wrong amount of currency and many more things so we have decided to help them and make them more independent.