Yiqun Ma
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yiqun Ma has a total of 25 co-authors. Here are the top co-authors by number of shared publications:
- Hao Bai
College of Information Engineering, Hainan Vocational University of Science and Technology, Haikou, Hainan, China
2 shared publications - Zihan Jiang
Shanghai Research Institute for Intelligent Autonomous Systems, TongJi University, Shanghai, China
2 shared publications - Bichen Shang1 shared publication
- Bingyu Shi
College of Computer and Control Engineering, Northeast Forestry University, Harbin, China
1 shared publication - Bo Jin
Institute of Systems and Robotics, Department of Electrical and Computer Engineering, University of Coimbra, Coimbra, Portugal
1 shared publication - Chengxuan Qin
School of Advanced Technology, Xian Jiaotong-Liverpool University, Suzhou, China
1 shared publication
Latest Publications
- Open Access
- 9 April 2026
- Published by Elsevier in Expert Systems with Applications
- Vol. 323
A collaborative optimization framework for efficient long-sequence Audio-Visual understanding
- 9 April 2026
- Published by Elsevier in Expert Systems with Applications
- Vol. 323
Open AccessA collaborative optimization framework for efficient long-sequence Audio-Visual understanding
- Open Access
- 14 December 2025
- Published by Association for Computing Machinery (ACM)
- p. 1588-1594
Reconstruction of Historical Digital Topography in a Mining Area Using Boundary-Constrained Linear Interpolation
- 14 December 2025
- Published by Association for Computing Machinery (ACM)
- p. 1588-1594
Open AccessReconstruction of Historical Digital Topography in a Mining Area Using Boundary-Constrained Linear Interpolation
- 6 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cybernetics
- Vol. 56 (1) , 15-28
Social Informer: Pedestrian Trajectory Prediction by Informer With Adaptive Trajectory Probability Region Optimization
Pedestrian trajectory prediction is an important research area with significant applications in autonomous driving and intelligent surveillance. However, existing studies on pedestrian trajectory prediction often suffer from a noticeable discrepancy between predicted and actual trajectories, due to incomplete extraction of pedestrian trajectory features and the randomness of the pedestrian walking process. The key objective of this article is to address this issue by proposing a method that can reasonably simulate the randomness of pedestrian walking and comprehensively extract pedestrian trajectory features. To achieve this, a novel social informer model built upon the informer model is proposed in this article. The social informer utilizes a transformer encoder-based interaction module to comprehensively extract pedestrian trajectory features, which are input into the informer model for further processing. Additionally, an adaptive variance mechanism is proposed to determine the optimal variance and accurately simulate the random nature of pedestrian walking. Finally, the proposed model is evaluated in a comparative experiment on ETH and UCY datasets, with results demonstrating that the proposed model outperforms other models, exhibiting improved accuracy and performance.
- 6 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cybernetics
- Vol. 56 (1) , 15-28
Social Informer: Pedestrian Trajectory Prediction by Informer With Adaptive Trajectory Probability Region Optimization
Pedestrian trajectory prediction is an important research area with significant applications in autonomous driving and intelligent surveillance. However, existing studies on pedestrian trajectory prediction often suffer from a noticeable discrepancy between predicted and actual trajectories, due to incomplete extraction of pedestrian trajectory features and the randomness of the pedestrian walking process. The key objective of this article is to address this issue by proposing a method that can reasonably simulate the randomness of pedestrian walking and comprehensively extract pedestrian trajectory features. To achieve this, a novel social informer model built upon the informer model is proposed in this article. The social informer utilizes a transformer encoder-based interaction module to comprehensively extract pedestrian trajectory features, which are input into the informer model for further processing. Additionally, an adaptive variance mechanism is proposed to determine the optimal variance and accurately simulate the random nature of pedestrian walking. Finally, the proposed model is evaluated in a comparative experiment on ETH and UCY datasets, with results demonstrating that the proposed model outperforms other models, exhibiting improved accuracy and performance.
- 24 June 2025
- Published by Springer Nature
- p. 196-210
Enhancing Semantic Segmentation in Open Compound Domain Adaptation Through Mixed Image and Epistemic Uncertainty
- 24 June 2025
- Published by Springer Nature
- p. 196-210
Enhancing Semantic Segmentation in Open Compound Domain Adaptation Through Mixed Image and Epistemic Uncertainty
- Open Access
- 18 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 300-304
Deep Learning Based Pepper Maturity Detection
China is one of the world's important pepper production area, the existing picking method still rely on manual picking method for picking, and pepper is a kind of batch maturity of the herb, manual picking is not only time-consuming but also very laborious, and the pepper target detection technology is the key to realize the automatic picking of pepper, the existing research for pepper target detection is relatively small, this paper intends to choose the image processing and machine learning methods for pepper to carry out target detection. First of all, the processed pepper image data are passed into Yolov5s, Yolov5m, Faster R-CNN, SSD to compare the four common deep learning methods, the comprehensive comparison in the pepper recognition stage Yolov5s is the most appropriate deep learning algorithm in the deep learning algorithms mentioned above, and in the second stage, using the binary classification method, the pepper data are divided into immature pepper and mature pepper, and directly sent to Yolov5s. The second stage uses a binary classification method to classify pepper data into immature and mature pepper, which is directly fed into the Yolov5s model with an accuracy of 97.4% for pepper maturity discrimination, providing diversified detection and corresponding technical support for the intelligent pepper picking summary.
- 18 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 300-304
Open AccessDeep Learning Based Pepper Maturity Detection
China is one of the world's important pepper production area, the existing picking method still rely on manual picking method for picking, and pepper is a kind of batch maturity of the herb, manual picking is not only time-consuming but also very laborious, and the pepper target detection technology is the key to realize the automatic picking of pepper, the existing research for pepper target detection is relatively small, this paper intends to choose the image processing and machine learning methods for pepper to carry out target detection. First of all, the processed pepper image data are passed into Yolov5s, Yolov5m, Faster R-CNN, SSD to compare the four common deep learning methods, the comprehensive comparison in the pepper recognition stage Yolov5s is the most appropriate deep learning algorithm in the deep learning algorithms mentioned above, and in the second stage, using the binary classification method, the pepper data are divided into immature pepper and mature pepper, and directly sent to Yolov5s. The second stage uses a binary classification method to classify pepper data into immature and mature pepper, which is directly fed into the Yolov5s model with an accuracy of 97.4% for pepper maturity discrimination, providing diversified detection and corresponding technical support for the intelligent pepper picking summary.
- 2 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Artificial Intelligence
- Vol. 5 (11) , 5575-5588
Social NSTransformers: Low-Quality Pedestrian Trajectory Prediction
This paper introduces a novel model for low-quality pedestrian trajectory prediction, the Social Non-stationary Transformers (NSTransformers), that merges the strengths of NSTransformers and Spatio-Temporal graph transformer (STAR). The model can capture social interaction cues among pedestrians and integrate features across spatial and temporal dimensions to enhance the precision and resilience of trajectory predictions. We also propose an enhanced loss function that combines diversity loss with logarithmic root mean squared error (log-RMSE) to guarantee the reasonableness and diversity of the generated trajectories. This design adapts well to complex pedestrian interaction scenarios, thereby improving the reliability and accuracy of trajectory prediction. Furthermore, we integrate a Generative Adversarial Network (GAN) to model the randomness inherent in pedestrian trajectories. Compared to the conventional standard Gaussian distribution, our GAN approach better simulates the intricate distribution found in pedestrian trajectories, enhancing the trajectory prediction’s diversity and robustness. Experimental results reveal that our model outperforms several state-of-the-art methods. This research opens the avenue for future exploration in low-quality pedestrian trajectory prediction.
- 2 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Artificial Intelligence
- Vol. 5 (11) , 5575-5588
Social NSTransformers: Low-Quality Pedestrian Trajectory Prediction
This paper introduces a novel model for low-quality pedestrian trajectory prediction, the Social Non-stationary Transformers (NSTransformers), that merges the strengths of NSTransformers and Spatio-Temporal graph transformer (STAR). The model can capture social interaction cues among pedestrians and integrate features across spatial and temporal dimensions to enhance the precision and resilience of trajectory predictions. We also propose an enhanced loss function that combines diversity loss with logarithmic root mean squared error (log-RMSE) to guarantee the reasonableness and diversity of the generated trajectories. This design adapts well to complex pedestrian interaction scenarios, thereby improving the reliability and accuracy of trajectory prediction. Furthermore, we integrate a Generative Adversarial Network (GAN) to model the randomness inherent in pedestrian trajectories. Compared to the conventional standard Gaussian distribution, our GAN approach better simulates the intricate distribution found in pedestrian trajectories, enhancing the trajectory prediction’s diversity and robustness. Experimental results reveal that our model outperforms several state-of-the-art methods. This research opens the avenue for future exploration in low-quality pedestrian trajectory prediction.
- Open Access
- 6 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 108746-108760
Stable Image Encryption Algorithm Based on Expanded One-Dimensional Chaotic Jumping and Parallel Encoding Operation Grouping
The current algorithm of encrypting images uses low-dimensional chaotic systems with a limited key space and poor security, while high-dimensional chaotic systems are difficult to implement and inefficient. Additionally, if image encryption algorithms use fixed DNA encoding rules, they could be cracked; dynamic coding can easily deviate from optimal solutions, thus causing algorithm instability. Based on this, a novel technique for image encryption is presented in this article. It is possible to avoid linear correlation of the chaotic sequence by randomly jumping on two uncorrelated one-dimensional chaoses. Meanwhile, numerous encryption results are generated in parallel by randomly grouping DNA encoding groups and encoding operations, and the optimal solution is then selected based on the encryption results, thereby minimizing ciphertext image instability to some extent. According to the experimental results, this algorithm produces high-security, plaintext-sensitive, and non-cracking encrypted images. Furthermore, this algorithm produces excellent encryption results both on color and sized images, and is applicable to a wide variety of applications.
- 6 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 108746-108760
Open AccessStable Image Encryption Algorithm Based on Expanded One-Dimensional Chaotic Jumping and Parallel Encoding Operation Grouping
The current algorithm of encrypting images uses low-dimensional chaotic systems with a limited key space and poor security, while high-dimensional chaotic systems are difficult to implement and inefficient. Additionally, if image encryption algorithms use fixed DNA encoding rules, they could be cracked; dynamic coding can easily deviate from optimal solutions, thus causing algorithm instability. Based on this, a novel technique for image encryption is presented in this article. It is possible to avoid linear correlation of the chaotic sequence by randomly jumping on two uncorrelated one-dimensional chaoses. Meanwhile, numerous encryption results are generated in parallel by randomly grouping DNA encoding groups and encoding operations, and the optimal solution is then selected based on the encryption results, thereby minimizing ciphertext image instability to some extent. According to the experimental results, this algorithm produces high-security, plaintext-sensitive, and non-cracking encrypted images. Furthermore, this algorithm produces excellent encryption results both on color and sized images, and is applicable to a wide variety of applications.