International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 10 Issue: 07 | JULY 2023
p-ISSN: 2395-0072
www.irjet.net
Enhancing Blind Assistance Applications with Optical Neural Networks: A Comparative Analysis of Electron-based Deep Learning and Photonbased Optical Computing SABBINENI LAKSHMI GOPI KOUSHIK1, JAGADESHWAR REDDY CHILAKAMARRI2 ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - With the use of deep learning and neural
a tiny delay in recognizing an object might have serious repercussions.
networks, blind aid technologies have advanced significantly in recent years. The current research explores the use of optical neural networks, which utilise photons for concurrent processing, as an innovative way to significantly shorten the blind person's reaction time. These optical neural networks are able to process visual input 3000 times faster than conventional neural networks, utilising the speed advantages of photons over electrons. This allows for the real-time detection of potentially life-threatening objects that cross a blind person's path. The technical features of optical neural networks along with the way they are used in applications for blind assistance are explored in this research, with a focus on how they might increase safety and enable rapid object detection.
The use of optical neural networks results in a quantum jump in parallel computation and processing performance. Optical neural networks can handle enormous quantities of visual input in parallel because they can carry out operations concurrently and independently using photons, the building blocks of light. Compared to its conventional electron-based equivalents, optical neural networks can process information up to 3000 times quicker because to this special property. With instantaneous feedback on items and dangers in the surrounding area, such unmatched speed offers up new applications for blind aid. The idea of optical neural networks is examined in this research study along with its potential to transform technology for the blind. We examine the technological foundations of optical neural networks and describe how they use the characteristics of photons to perform parallel processing and real-time object identification. We also look into the networks' architectural layout, emphasising how well they fit with current blind aid systems already in place.
Key Words: Deep Learning, Neural Networks, Blind Aid Technologies, Optical Neural Networks, Photons, Concurrent Processing, Reaction Time.
1. INTRODUCTION Millions of individuals all over the world are severely disabled by blindness, which restricts their mobility and liberty. Researchers and creators have been working for years to create modern technologies that can help blind people and enhance their quality of life in general. Through the provision of real-time object identification and spatial awareness, recent developments in artificial intelligence (AI) and deep learning have provided fascinating prospects to transform applications for blind assistance. Among these developments, optical neural networks a cutting-edge method which makes use of photons for computation have emerged as a potentially useful tool for enhancing blind people's safety and reducing response times.
The primary objective of this research is to demonstrate how optical neural networks affect the safety and autonomy of blind people in the real world. Optical neural networks can allow blind people to travel with confidence, lowering the danger of future accidents and increasing their level of independence overall. This is done by enabling immediate object detection. Through case studies and user experiences, we hope to demonstrate with compelling evidence how the usage of optical neural networks may significantly improve the daily lives of people who are blind or visually impaired.
2. OVERVIEW OF DEEP LEARNING
The use of traditional blind aid devices to help the visually impaired navigate their surroundings has evolved drastically With the help of electron-based neural networks, object identification skills have been significantly improved. Now, computers are able to analyse visual data and provide it to the user through the auditory process or tactile feedback. The reliance on sequential processing of electron-based networks, which can cause observable delays in real-time applications, is a crucial constraint of these networks. Blind people may not have enough time to respond to potentially dangerous barriers in dynamic surroundings, therefore even
© 2023, IRJET
|
Impact Factor value: 8.226
Recent years have seen the emergence of deep learning as a cutting-edge technology with a wide range of applications in a variety of industries. Drawing on the principles of artificial intelligence and drawing inspiration from the human brain, the deep learning algorithms have fundamentally altered the way in which machines learn and process data. The fundamental components of deep learning are artificial neural networks, which are computational models that replicate the interconnected neurons of the brain. These neural networks are able to learn from large amounts of
|
ISO 9001:2008 Certified Journal
|
Page 1192