Open AccessComparative Study of Vehicle Detection using SSD and Faster RCNN
Author(s) -
K. Mohan Krishna,
Arcot Sowmya,
D. Jerusha,
D. Susmitha
Publication year - 2021
Publication title -
international journal of computer science and mobile computing
Language(s) - English
Resource type - Journals
ISSN - 2320-088X
DOI - 10.47760/ijcsmc.2021.v10i07.004
Subject(s) - computer science , artificial intelligence , detector , focus (optics) , deep learning , law enforcement , intelligent transportation system , machine learning , pattern recognition (psychology) , engineering , telecommunications , transport engineering , physics , law , political science , optics
With the recent developments in Artificial Intelligence, vehicle detection systems have become an essential part of many sectors like transport, automobile, security, law enforcement, and traffic management. This increased the requirement for an efficient system for vehicle detection. The main focus of our work is to find the best algorithm which can be used to design a vehicle detection system. For this, we compare two well-known deep learning algorithms which are Faster R-CNN and Single Shot Detector (SSD) algorithms. Both of the Pre-trained models of Tensorflow were tested on a dataset of hundred images with cars in them. It was found that Faster R-CNN is better with an accuracy score of 82.75 but was slower than SSD, whereas SSD had an accuracy score of 80.58 but was faster compared to Faster R-CNN.