Object Detection and Recognition in Autonomous Vehicles using Fast Region-Convolutional Neural Network

Object detection is a computer vision approach whose fundamental intent is to detect objects and it can be detected from any video or image feeds. The accurate and timely identification of traffic signs plays a vital role in autonomous vehicles. However, the traffic sign’ s size represents a small part of the input image. In this research, the Fast Region-Convolutional Neural Network (R-CNN) is proposed for object detection and recognition in autonomous vehicles using deep learning. Initially, the data is obtained by the KITTI dataset and data augmentation is employed to increase the generalization capability of the model. Then, the Gray Level Co-occurrence Matrix (GLCM) and ResNet-50 are used for feature extraction. The Attention-guided Context Feature Pyramid Network (ACFPN) is employed for feature fusion and then the Fast RCNN approach is utilized for detection and recognition purposes. The proposed Fast RCNN achieves a better mean Average Precision (mAP) of 99.26% compared to the existing methods respectively.