Research on Pedestrian Detection Method Based on Receptive Field Module

With the continuous development of technology, autonomous driving has become a popular subject of discussion in the field of artificial intelligence. Accurate detection of passing pedestrians has been identified as an important problem in unmanned driving. Given this, this paper proposes an improved...

Full description

Saved in:
Bibliographic Details
Published in2021 International Conference on Artificial Intelligence, Big Data and Algorithms (CAIBDA) pp. 107 - 112
Main Authors Kuang, Xianyan, Zhang, Jianhua
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.05.2021
Subjects
Autonomous driving
Big Data
Deep learning
Detectors
ECA mudule
Feature extraction
pedestrian detection
Real-time systems
RFB module
Robustness
Safety
Online AccessGet full text

Cover

More Information
Summary:With the continuous development of technology, autonomous driving has become a popular subject of discussion in the field of artificial intelligence. Accurate detection of passing pedestrians has been identified as an important problem in unmanned driving. Given this, this paper proposes an improved Single Shot MultiBox Detector (SSD) algorithm. The algorithm is improved from two aspects. Firstly, a Receptive Field Block (RFB) module is assembled to the top of the SSD to enhance the feature extraction capability and robustness of the model. Secondly, an Efficient Channel Attention (ECA) module is added to improve the performance of the deep neural network and enhance its feature representation capability. The experimental results on the PASCAL VOC 2007 (VOC) dataset show that the proposed algorithm can significantly improve the accuracy and speed of pedestrian detection compared with the traditional SSD algorithm.
DOI:10.1109/CAIBDA53561.2021.00030