Efficient Detection of Small and Complex Objects for Autonomous Driving Using Deep Learning

Abstract

The YOLOv2 is one of the most prominent model used for object detection, it works on the concept of anchor boxes. However, this model is prone to some problems like double anchor boxes, missing small objects, and high time complexity. In this paper, we aim to solve the problem of double anchor boxes and undetected small objects by tuning the parameters like intersection over union (IoU) and customizing non-max suppression thresholds. Also, to reduce the time complexity of the model, we have proposed the use of depth wise convolution (DW-Conv2D) instead of fundamental convolution (Conv2D) in this paper. Once we applied the proposed model to datasets like PASCAL VOC07 and VOC12, we observed significant improvements like reduced floating-point operations per second by 9.5% and better accuracy than the existing state-of-the-art models.