International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 10 Issue: 06 | June 2023
p-ISSN: 2395-0072
www.irjet.net
A Smart Assistance for Visually Impaired Rohith Reddy Appidi1, Ummagani Avinash2 1Dept. Of Electronis and Communication Engineering, VNR Vignana Jyothi Institute of Engineering and Technology
Hyderabad, India
2Dept. Of Electronis and Communication Engineering, VNR Vignana Jyothi Institute of Engineering and Technology
Hyderabad, India ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Independent living is one of the major concerns
on the image collected, the You Only Look Once Object detection algorithm was employed. We used the weights of the model trained on the COCO dataset to do inferences on the captured image in this project. The COCO dataset, which stands for Common Things in Context, is aimed to illustrate a vast group of objects that humans see regularly. The COCO Dataset contains 121,408 pictures, 883,331 object annotations, and 80 object types(categories). Another aspect of our system is the ability to search Wikipedia and send emails. The entire system was built using the Python programming language.
for people around the globe. But when it comes to people who are visually challenged, they always rely on other people to get their daily things done. In today's advanced technical environment, the need for self-sufficiency is recognized in the situation of visually impaired people who are socially restricted. More than 90% of individuals with blindness and low vision are mostly seen in developing countries. Visually impaired people always need the help of others, and they rely on others for essential needs. In this project, we developed a system that uses a conglomeration of technologies like Object detection, Speech Recognition, etc, to solve the problems faced by blind people to a certain extent. It makes use of a camera module that acts as a virtual eye for the visually impaired and helps them recognize objects surrounding them. This system also includes other features like searching Wikipedia and sending emails. All the features are implemented using a headset that provides voice assistance for an easier lifestyle for the visually impaired. The whole system was developed in the Python programming language.
2. PROPOSED METHODOLOGY The smart assistant that was built for the visually impaired will be capable of sending emails (also, SOS alerts), searching Wikipedia, and gives the object classes that are being detected by the YOLO Algorithm in the form of voice output. The Fig.1 given below best describes the process flow of this sub-task. Initially, the user speaks out a pre-defined command and if the command matches the string email, then the function corresponding to this sub-task will be invoked. If the command is SOS, then the camera captures an image and attaches the image as an email attachment, and this information is sent to the specific mail ids provided by the user.
Key Words: Visually Impaired, Text to Speech, Speech to Text, SMTP Protocol, Wikipedia, API, Object Detection, YOLO, COCO dataset.
1. INTRODUCTION Human vision is a gift to humanity that allows us to visually experience and interact with objects across the planet. When it comes to a person with a visual impairment, it becomes a difficult situation that they must deal with every second of their lives. It is the most prevalent type of disability identified in people (about 285 million according to WHO). Visual impairment has a significant impact on several abilities related to vision: Daily activities (which necessitate vision), Interaction over the Internet (which necessitate interaction with others over the Internet), Emergencies (which necessitate safety). A Smart Assistance Application for the Visually Impaired is an application that helps visually impaired persons get more familiar with their surroundings and recognize objects. We used a system that integrated technologies like object recognition, speech-to-text, and textto-voice conversion in this project to help blind individuals overcome some of the challenges they confront. The object detection method is used on the saved image, which aids in the detection of an object in the image. To execute inference
© 2023, IRJET
|
Impact Factor value: 8.226
Fig -1:User Interaction flow for sending emails
|
ISO 9001:2008 Certified Journal
|
Page 636