Miao Liu

Prof. Dr. Miao Liu


Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
12 August 2024 in Applied Sciences
Verified
Journal Article
Open access
PDF file
MLMenghao Li
Menghao Li
MLMiao Liu
Miao Liu
WZWeiwei Zhang
Weiwei Zhang
Wenfeng GuoWenfeng Guo
Wenfeng Guo
ECEnqing Chen
Enqing Chen
CZCheng Zhang
Cheng Zhang
In intelligent traffic monitoring systems, the significant distance between cameras and their non-overlapping fields of view leads to several issues. These include incomplete tracking results from individual cameras, difficulty in matching targets across multiple cameras, and the complexity of inferring the global trajectory of a target. In response to the challenges above, a deep learning-based vehicle tracking algorithm called FairMOT-MCVT is proposed. This algorithm con-siders the vehicles’ characteristics as rigid targets from a roadside perspective. Firstly, a Block-Efficient module is designed to enhance the network’s ability to capture and characterize image features across different layers by integrating a multi-branch structure and depth-separable convolutions. Secondly, the Multi-scale Dilated Attention (MSDA) module is introduced to improve the feature extraction capability and computational efficiency by combining multi-scale feature fusion and attention mechanisms. Finally, a joint loss function is crafted to better distinguish between vehicles with similar appearances by combining the trajectory smoothing loss and velocity consistency loss, thereby considering both position and velocity continuity during the optimization process. The proposed method was evaluated on the public UA-DETRAC dataset, which comprises 1210 video sequences and over 140,000 frames captured under various weather and lighting conditions. The experimental results demonstrate that the FairMOT-MCVT algorithm significantly enhances multi-target tracking accuracy (MOTA) to 79.0, IDF1 to 84.5, and FPS to 29.03, surpassing the performance of previous algorithms. Additionally, this algorithm expands the detection range and reduces the deployment cost of roadside equipment, effectively meeting the practical application requirements.

2 October 2024 in Applied Sciences
Verified
Journal Article
Open access
PDF file
MLMenghao Li
Menghao Li
MLMiao Liu
Miao Liu
WZWeiwei Zhang
Weiwei Zhang
Wenfeng GuoWenfeng Guo
Wenfeng Guo
ECEnqing Chen
Enqing Chen
Maomao ZhangMaomao Zhang
Maomao Zhang
Accidents caused by vehicles changing lanes occur frequently on highways. Moreover, frequent lane changes can severely impact traffic flow during peak commuting hours and on busy roads. A novel framework based on a multi-relational graph convolutional network (MR-GCN) is herein proposed to address these challenges. First, a dynamic multilevel relational graph was designed to describe interactions between vehicles and road objects at different spatio-temporal granularities, with real-time updates to edge weights to enhance understanding of complex traffic scenarios. Second, an improved spatio-temporal interaction graph generation method was introduced, focusing on spatio-temporal variations and capturing complex interaction patterns to enhance prediction accuracy and adaptability. Finally, by integrating a dynamic multi-relational graph convolutional network (DMR-GCN) with dynamic scene sensing and interaction learning mechanisms, the framework enables real-time updates of complex vehicle relationships, thereby improving behavior prediction’s accuracy and real-time performance. Experimental validation on multiple benchmark datasets, including KITTI, Apollo, and Indian, showed that our algorithmic framework achieves significant performance improvements in vehicle behavior prediction tasks, with Map, Recall, and F1 scores reaching 90%, 88%, and 89%, respectively, outperforming existing algorithms. Additionally, the model achieved a Map of 91%, a Recall of 89%, and an F1 score of 90% under congested road conditions in a self-collected high-speed traffic scenario dataset, further demonstrating its robustness and adaptability in high-speed traffic conditions. These results show that the proposed model is highly practical and stable in real-world applications such as traffic control systems and self-driving vehicles, providing strong support for efficient vehicle behavior prediction.

Verified
Journal Article
Open access
PDF file
LHLingling Hong
Lingling Hong
MLMiao Liu
Miao Liu
Achieving a stable color output across wide temperature ranges in automotive LED applications is challenging, especially when using cost-sensitive chips with limited computational resources. This study proposes an improved temperature model that integrates Fourier heat conduction and thermal resistance concepts to more accurately capture self-heating and power dissipation effects. To accommodate the constraints of low-cost automotive-grade microcontrollers (MCUs), the associated optical algorithm is converted from floating-point to a 16.16 fixed-point format, reducing both memory usage and computational overhead. Experimental results conducted from −40 °C to 120 °C show that the improved model predicts LED temperatures within 5 °C of measured values, reducing errors by up to 30% compared to conventional PN-junction-based methods. Furthermore, by comparing the chromaticity points generated under the new and traditional models—and implementing an additional three-duty-cycle offset at 1% brightness—the improved approach reduces chromaticity drift by approximately 0.0052 in the CIE 1931 xy color space. These findings confirm the superior stability and accuracy of the new model for both thermal management and chromaticity compensation, offering a cost-effective solution for automotive LED systems requiring precise color control under constrained MCU resources.

16 January 2025 in Molecular & Cellular Biomechanics
Verified
Journal Article
Open access
PDF file
LTLiangjie Tao
Liangjie Tao
MLMiao Liu
Miao Liu
In the realm of modern applications, especially those related to human - involved scenarios such as vehicles and medical or wearable devices, the performance of RGB LED lights is of great significance. Beyond the traditional concerns of electronic - optical properties, the integration of biomechanics - related factors can bring new perspectives and optimizations. In vehicles, during prolonged illumination, the color shifts of various LED lights not only affect driving comfort but also have potential implications for driving safety from a biomechanical perspective. The human body’s visual and psychological responses to these color and luminance changes are complex. For example, sudden or inconsistent color changes can cause visual fatigue and distraction, which may impact a driver’s reaction time and decision - making ability, all of which are related to biomechanical aspects of human - machine interaction. Each LED shows deviations in wavelength and luminous intensity, and the light decay of RGB chips is inconsistent. To address these issues and ensure consistent color and luminance of all RGB ambient lights during use, this paper proposes a temperature prediction model (PT - model) based on constant current source output PWM. This model takes into account the impact of temperature changes on the color coordinates and luminous flux of LED lights. Moreover, considering the potential applications in medical and wearable devices, the model’s performance becomes even more crucial. In medical light - based treatment devices, the precise control of RGB LED temperature and light output is essential for ensuring the effectiveness of treatment while minimizing potential harm to human tissues. Biomechanics research can provide insights into how different tissues respond to light - induced heat and mechanical stress. Similarly, in wearable devices for health monitoring, the stability of RGB LED performance is related to the comfort and accuracy of the device’s operation on the human body. Experimental results show that compared to traditional models, the prediction accuracy of this model is significantly improved, with errors reduced to within 4.3%. The model’s effectiveness is verified under different ambient temperatures ranging from −40 ℃ to 120 ℃, which is crucial for ensuring its reliability in various real - world applications, especially those related to human - centered scenarios influenced by biomechanical factors.

25 July 2025
Verified
Conference Paper
Open access
LYLonghua Yuan
Longhua Yuan
MLMiao Liu
Miao Liu

21 January 2022 in Research Square
Verified
Preprint
Open access
PDF file
MLMiao Liu
Miao Liu
SKS.K. Kauh
S.K. Kauh
SLSabina Lim
Sabina Lim
Moxibustion is an important and widely used treatment method in traditional medical science. Two methods of moxibustion exist: tonification and sedation. Although these two methods have been used for thousands of years, there still has been no scientific comparative study of their effects on the skin until now. Thus, this paper aims to use a mathematical method to study moxibustion methods in skin tissue to demonstrate a difference of thermomechanics between tonification and sedation. In this paper, the temperature, burn damage, and thermal stress distributions in the skin tissue is analyzed and the results showed that the thermomechanical behavior of skin tissue between tonification and sedation is in accordance to ancient literatures of traditional medicine. Therefore, the comparative study of heat transfer and related thermomechanics in skin tissue during these treatments thermomechanics to tonification and sedation is great importance and helpful in exploring moxibustion instruments and their standardization contribute to the further developments of these medical applications.