• Kuchibhotla Ananta Sai Sree Sindhura
• Dr. Jitendra Jaiswal

One of the most essential human senses is vision and it has a significant role in the perception human beings make regarding the environment around them. Therefore we can say that by the development of modern aids based on electronic gadgets for the blind people, many of this topic related papers are generated suggesting a wide range of products based on computer vision and important services. This project with the assistance of the camera, helps blind and visually disabled people identify and recognize objects around them in indoor spaces. This helps the blind and visually impaired individuals maneuver indoor spaces without relying on others. Through an earpiece, voice feedback is sent that can be heard to understand what object is present in the surrounding by reducing time and effort. I aim to create an intelligent system that imitates the human eye, transferring different scenes and images into the brain. The brain, in turn, analyses the images or scenes, and the objects encircling are identified based on information previously stored. For this purpose, we use a device called, “Raspberry Pi” which performs similar to the human brain. Then the analysis process begins with long complex algorithms known as the algorithmic neural networks. YOLO is the basis of these algorithms and YOLO stands for “You Only Look Once”, this is the method pre-trained on the dataset of COCO. Over other objects detecting algorithms, YOLO has some significant benefits. In CNN type of algorithms (Convolutional neural network), FCNN which is an abbreviation of Fast convolutional neural network, the image is not fully observed in the algorithm but in case of this YOLO algorithm, whole image is visualised completely by forecasting the special bounding boxes making use of some important convolutional networks. In this way class possibilities can be determined for the boxes of these kinds and image is detected faster as compared to any other algorithm used. This network analyses the images in parts, so that, they match the most important characteristics of the objects in the images associated with the database by comparing the images. When ensuring that the characteristics match the mathematical equations programmed within the Python language, the objects inside the image are detected. Finally, the sound of each tool is called in the database, and a message is sent to tell the blind about tools in front of him/her.

Vision, Blind, Camera, Earpiece, Raspberry Pi, YOLO, COCO, Object Detection, CNN, Bounding boxes, Neural Network, Python

"Real Time Object Detection and Recognition with a Voice Feedback for the Blind", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.6, Issue 3, page no.448-455, March-2019, Available :http://www.jetir.org/papers/JETIR1903L60.pdf

"Real Time Object Detection and Recognition with a Voice Feedback for the Blind", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.6, Issue 3, page no. pp448-455, March-2019, Available at : http://www.jetir.org/papers/JETIR1903L60.pdf