Tingting Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Tingting Wang has a total of 4 co-authors. Here are the top co-authors by number of shared publications:
- Wei Zhao
School of Vehicle and Traffic Engineering, Henan University of Science and Technology, Luoyang 471003, China
2 shared publications Congcong Ren
College of Vehicle and Traffic Engineering, Henan University of Science and Technology, Luoyang, China
1 shared publicationWei Zhao
College of Vehicle and Traffic Engineering, Henan University of Science and Technology, Luoyang, China
1 shared publication- Yunqing Song
China College of Vehicle and Traffic Engineering, Henan University of Science and Technology, Luoyang, China
1 shared publication
Latest Publications
- Open Access
- 4 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 137734-137746
SAU-Net: Monocular Depth Estimation Combining Multi-Scale Features and Attention Mechanisms
Monocular depth estimation technology is widely utilized in autonomous driving for sensing and obstacle avoidance. Recent advancements in deep-learning techniques have resulted in significant progress in monocular depth estimation. However, monocular depth estimation is mainly optimized for the luminosity error of pixels, mostly disregarding the related problems of result ambiguity and boundary artifacts in the image. To address these issues, we developed an improved network model called SAU-Net. The superposition of excessive convolutional layers in conventional convolution networks impairs the network’s timeliness and results in the loss of primary information. Therefore, we propose a convolution-free stratified transformer as an image feature extractor at the network’s coding end, which limits self-attention to innumerable windows and leverages sliding windows for characterization to reduce the network delay. This study also addresses the issue of critical information loss. We connect each feature map directly to another from a different scale. In addition, an attention module is introduced to focus on the effective features, which increases the amount of target information in the depth map. We employ the gradient loss function during the training stage to improve the segmentation accuracy of the network and the smoothness of the output image. Training and testing were conducted using the KITTI dataset. To ensure the robustness of the algorithm in practical applications, we also validated the algorithm using a campus dataset that we collected. The experimental results indicated that the accuracy of the algorithm was 89.1%, 96.4%, and 98.5% under three proportional thresholds. The estimated depth map was clear in details and edges, with fewer artifacts.
- 4 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 137734-137746
Open AccessSAU-Net: Monocular Depth Estimation Combining Multi-Scale Features and Attention Mechanisms
Monocular depth estimation technology is widely utilized in autonomous driving for sensing and obstacle avoidance. Recent advancements in deep-learning techniques have resulted in significant progress in monocular depth estimation. However, monocular depth estimation is mainly optimized for the luminosity error of pixels, mostly disregarding the related problems of result ambiguity and boundary artifacts in the image. To address these issues, we developed an improved network model called SAU-Net. The superposition of excessive convolutional layers in conventional convolution networks impairs the network’s timeliness and results in the loss of primary information. Therefore, we propose a convolution-free stratified transformer as an image feature extractor at the network’s coding end, which limits self-attention to innumerable windows and leverages sliding windows for characterization to reduce the network delay. This study also addresses the issue of critical information loss. We connect each feature map directly to another from a different scale. In addition, an attention module is introduced to focus on the effective features, which increases the amount of target information in the depth map. We employ the gradient loss function during the training stage to improve the segmentation accuracy of the network and the smoothness of the output image. Training and testing were conducted using the KITTI dataset. To ensure the robustness of the algorithm in practical applications, we also validated the algorithm using a campus dataset that we collected. The experimental results indicated that the accuracy of the algorithm was 89.1%, 96.4%, and 98.5% under three proportional thresholds. The estimated depth map was clear in details and edges, with fewer artifacts.
- Open Access
- 3 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 68439-68451
Nighttime Pedestrian Detection Based on a Fusion of Visual Information and Millimeter-Wave Radar
A target detection algorithm based on the fusion of vision sensors and millimeter wave radar data can effectively improve the safety of self-driving vehicles. However, a single sensor cannot obtain comprehensive category status information on the target in the nighttime environment. To improve pedestrian detection in nighttime traffic scenarios, this paper proposes a nighttime pedestrian detection method based on the fusion of infrared vision information and millimeter wave (MMW) radar data. The lateral localization and category features of the target are obtained using the improved YOLOv5 deep learning algorithm, the distance and velocity information of the target is obtained by preprocessing the MMW radar acquisition data, the pedestrian target is tracked by using the extended Kalman filter for MMW radar detection, the projection of the valid radar target point on the IR image is completed according to the spatiotemporal fusion, and then the correlation gate method is used to correlate the data, and the visual information is inherited to the radar points to get the target multimodal information, and the success associated valid target sequences are weighted to obtain the accurate target position. Finally, a decision-level fusion algorithm framework is proposed to complete the output of pedestrian multimodal information in nighttime traffic scenes. Theoretical analysis and experimental results show that the accuracy and robustness of this method for nighttime pedestrian recognition are better than those of a single sensor.
- 3 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 68439-68451
Open AccessNighttime Pedestrian Detection Based on a Fusion of Visual Information and Millimeter-Wave Radar
A target detection algorithm based on the fusion of vision sensors and millimeter wave radar data can effectively improve the safety of self-driving vehicles. However, a single sensor cannot obtain comprehensive category status information on the target in the nighttime environment. To improve pedestrian detection in nighttime traffic scenarios, this paper proposes a nighttime pedestrian detection method based on the fusion of infrared vision information and millimeter wave (MMW) radar data. The lateral localization and category features of the target are obtained using the improved YOLOv5 deep learning algorithm, the distance and velocity information of the target is obtained by preprocessing the MMW radar acquisition data, the pedestrian target is tracked by using the extended Kalman filter for MMW radar detection, the projection of the valid radar target point on the IR image is completed according to the spatiotemporal fusion, and then the correlation gate method is used to correlate the data, and the visual information is inherited to the radar points to get the target multimodal information, and the success associated valid target sequences are weighted to obtain the accurate target position. Finally, a decision-level fusion algorithm framework is proposed to complete the output of pedestrian multimodal information in nighttime traffic scenes. Theoretical analysis and experimental results show that the accuracy and robustness of this method for nighttime pedestrian recognition are better than those of a single sensor.