An Efficient Approach of Vehicle Detection Based on Deep Learning Algorithms and Wireless Sensors Networks
Cherifa Nakkach, Amira Zrelli, Tahar Ezzdine
Abstract
Machine learning is applied to analyze and classify automatically images. Artificial intelligence (AI) is considered very successful in this area. Therefore, AI is exploited to evaluate the opportunities of big data and to extract value from massive and varied data sources. In order to detect any event (person, vehicle, dog, eyes, traffic, terrorist activity), ML is explored. Hence, advanced ML techniques recur to multimedia wireless sensor networks (MWSN) to detect any event in the considered area. In this work, the authors propose an enhanced architecture MWSN, which is able fly any event detection. In this context, this paper addresses the problem of vehicle detection using convolutional neural networks using a proposed architecture MWSN. Therefore, to reach this goal, the authors assess the performance of three state-of-the-art CNN algorithms, namely faster R-CNN, which is the most popular region-based algorithm; YOLO, which is known to be the fastest one; and SSD, which takes one single shot to detect multiple objects within the image.