Shallow Convolution Neural Network for an Industrial Robot Real Time Vision System

Dumindu Eranda Jayakody; J.A.K.S. Jayasinghe; D. C. Wijewardene1

1

Publication Date: 2021/03/06

Abstract: Recent advancement in the deep learningbased object detection techniques have significantly improved the detection speed and accuracy. Now, it is possible to execute a complex deep neural network having a large number of layers in real time using desktop computers equipped with graphic processing units (GPU’s). Further, shallow neural network architectures are also developed for embedded systems with less processing power. This paper proposes shallow neural network architecture capable of real-time object detection as a vision system of an industrial robot. The proposed shallow neural network is executed on Raspberry Pi 3B+ having limited resources compared to a desktop computer equipped with GPU’s.

Keywords: Computer Vision, Object Detetion, Yolo

DOI: No DOI Available

PDF: https://ijirst.demo4.arinfotech.co/assets/upload/files/IJISRT21FEB537.pdf

REFERENCES

No References Available