A Review on Real-Time Object Detection Models Using Deep Neural Networks

EasyChair Preprint 9011

Abstract

Object detection is one of the most well-known challenges in computer vision. Many researchers have been employed in numerous application fields, including robotics, autonomous driving, and video surveillance. This paper offers a review of real-time object detection techniques using deep learning approaches. It is aimed at familiarizing the readers with the relevant knowledge, literature, and the latest updates on the state-of-art techniques. This study review records obtained electronically through the leading scientific databases (IEEE, Google Scholar, Scopus, Science Direct, Elsevier, and other journal publications) searched using three sets of keywords: Deep learning, object detection, and convolutional neural networks. Two different categories can be found in the object detection framework, traditional detectors, and deep learning-based detectors. The deep learning object detectors are divided into the two-stage detector and the one-stage detector. One-stage detectors use dense anchor boxes to perform classification and regression without establishing a sparse region of interest collection, while in two-stage detectors, sparse region proposals are created in the first stage of two-stage detectors, after which they are regressed and categorized. Object detection has been applied in crop harvesting, object detector models for blind persons, detection of pedestrians on the road, traffic sign detection and classification, text detection, and remote sensing target detection. In our future work, we propose to develop a one-stage object detection model that may help in guiding blind movements

Keyphrases: Convolutional Neural Networks, deep learning, object detection

@booklet{EasyChair:9011,
  author    = {Lenard Byenkya Nkalubo and Rose Nakibuule},
  title     = {A Review on Real-Time Object Detection Models Using Deep Neural Networks},
  howpublished = {EasyChair Preprint 9011},
  year      = {EasyChair, 2022}}