Mahmoud M. Mahmoud; Ahmed R. Nasser
Volume 21, Issue 2 , June 2021, , Page 36-43
Abstract
Object detection of autonomous vehicles presents a big challenge forresearchers due to the requirements of accuracy and precision in real-time.This work presents a deep learning approach based on a dual architecturedesign of the network. A highly accurate multi-class network of convolutionalneural networks ...
Read More ...
Object detection of autonomous vehicles presents a big challenge forresearchers due to the requirements of accuracy and precision in real-time.This work presents a deep learning approach based on a dual architecturedesign of the network. A highly accurate multi-class network of convolutionalneural networks (CNN) is presented for input data classification. A Region-Based Convolutional Neural Networks (Faster R-CNN) network with a modifiedFeature Pyramid Networks (FPN) is used for better detection of tiny objects andYou Only Look Once (YOLOv3) network is used for general detection. Eachnetwork independently detects the existence of an object. The decision maps arethen fused and compared to decide whether an object is present or not. FasterR-CNN with FPN model reported a higher intersection over Union (IoU) andmean average precision (mAP) than the YOLOv3. This approach is reliabledemonstrating an upgrade on the existing state-of-the-art methods of fullyconnected networks.