Introduction

Background and related work

Semantic attentional bilinear network (SAS-Net)

Loss function and training

Experiments

Conclusion and future work

References

Abstract

Vehicle tracking task plays an important role on the Internet of vehicles and intelligent transportation system. Beyond the traditional Global Positioning System sensor, the image sensor can capture different kinds of vehicles, analyze their driving situation, and can interact with them. Aiming at the problem that the traditional convolutional neural network is vulnerable to background interference, this article proposes vehicle tracking method based on human attention mechanism for self-selection of deep features with an inter-channel fully connected layer. It mainly includes the following contents: (1) a fully convolutional neural network fused attention mechanism with the selection of the deep features for convolution; (2) a separation method for template and semantic background region to separate target vehicles from the background in the initial frame adaptively; (3) a two-stage method for model training using our traffic dataset. The experimental results show that the proposed method improves the tracking accuracy without an increase in tracking time. Meanwhile, it strengthens the robustness of algorithm under the condition of the complex background region. The success rate of the proposed method in overall traffic datasets is higher than Siamese network by about 10%, and the overall precision is higher than Siamese network by 8%.

Introduction

The Internet of vehicles (IOV) can improve people’s travel efficiency through urban traffic management, traffic congestion detection, path planning, road charge, and public transportation management, to alleviate traffic congestion. By the surveillance camera in the bayonet and key sections of the city, we can perform recognition and tracking of all types of vehicles. Based on the statistical analysis of the recognition and tracking results on the server, we can calculate the driving path and determine the intention of the moving vehicle. In this way, we can also analyze the real-time road conditions at the location of the sensors and then feed the results back to the user’s vehicle through a wireless sensor, guiding the next step and recommending the appropriate route. Vehicle tracking is a key technology of IOV and intelligent transportation system (ITS), in which the image sensor and wireless sensor are complementary. The accuracy and speed of vehicle tracking system directly affect the performance of IOV. In recent years, with the development of computer hardware and improving the performance of intelligent algorithms, the performance of the vehicle tracking system is also increasing. The goal of vehicle tracking task is to get the position information of the target vehicle in the first frame in a video or an image sequence. In each subsequent frame, the position of the vehicle is predicted by various operations, including the center coordinates and the width and height of the target vehicle. The difficulty of vehicle tracking is how to select effective feature extraction methods for different scenes to express the target image region, so that the tracking model can effectively learn and predict the input samples.