Siyi Yang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Siyi Yang has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Hui Lu
School of Electronic and Information Engineering, Beihang University, Beijing, China
2 shared publications - Meng Zhao
School of Electronic and Information Engineering, Beihang University, Beijing, China
2 shared publications - Fengjuan Guo
Shaanxi Key Laboratory of Integrated and Intelligent Navigation, Shaanxi, Xi’an, 710068, China
1 shared publication - Jingxiang Sun
Department of Radiology, The First Affiliated Hospital of Shandong First Medical University & Shandong Provincial Qianfoshan Hospital, Shandong, China; Postgraduate Department, Shandong First Medical University (Shandong Academy of Medical Sciences), Shandong, China
1 shared publication Shi Cheng
1 shared publication- Yuhui Shi
Southern University of Science and Technology
1 shared publication
Latest Publications
- 10 August 2022
- Published by Elsevier in Applied Soft Computing
- Vol. 128
A multi-robot cooperative exploration algorithm considering working efficiency and working load
- 10 August 2022
- Published by Elsevier in Applied Soft Computing
- Vol. 128
A multi-robot cooperative exploration algorithm considering working efficiency and working load
- 1 August 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 172-176
Research on Computer Aided Detection and Segmentation of Pulmonary Nodules Based on Deep Learning and Generative Adversarial Networks
Pulmonary tuberculosis assisted diagnosis technology mainly faces the problems of poor resistance to overfitting due to the small training dataset and the low accuracy of lung tissue interference detection. This paper mainly solves the current problems from two aspects. First, this paper uses the RetinaNet[l] network structure combined with the DenseNet [2] feature extraction network to perform image detection on lung CT targets. Second, this paper enhance the training dataset by increasing the size of the dataset with the method of generative adversarial network to reduce the phenomena such as overfitting in experimental results. After testing on the existing dataset, it was found that the accuracy rate of target detection based on the improved model is 87.32%, which is 70.1% on the original model. Besides, the accuracy rate can be increased to 90.1% by the enhancement of the dataset using generative adversarial network.
- 1 August 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 172-176
Research on Computer Aided Detection and Segmentation of Pulmonary Nodules Based on Deep Learning and Generative Adversarial Networks
Pulmonary tuberculosis assisted diagnosis technology mainly faces the problems of poor resistance to overfitting due to the small training dataset and the low accuracy of lung tissue interference detection. This paper mainly solves the current problems from two aspects. First, this paper uses the RetinaNet[l] network structure combined with the DenseNet [2] feature extraction network to perform image detection on lung CT targets. Second, this paper enhance the training dataset by increasing the size of the dataset with the method of generative adversarial network to reduce the phenomena such as overfitting in experimental results. After testing on the existing dataset, it was found that the accuracy rate of target detection based on the improved model is 87.32%, which is 70.1% on the original model. Besides, the accuracy rate can be increased to 90.1% by the enhancement of the dataset using generative adversarial network.
- Open Access
- 3 March 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 8, 47824-47844
The Experience-Memory Q-Learning Algorithm for Robot Path Planning in Unknown Environment
In order to solve the problem of slow convergence rate and long planned path when the robot plans a path in unknown environment by using Q-learning algorithm, we propose an Experience-Memory Q-Learning (EMQL) algorithm based on the continuous update of the shortest distance from the current state node to the start point. The autonomous learning ability of the robot is enhanced by the different role assignment of two tables in the proposed algorithm. EM table with (m * 1) dimension is designed to record the distance information, reflecting the learning process of the robot. Q table is adopted as an auxiliary guidance for the experience transfer strategy and experience reuse strategy, and these strategies enable the robot accomplish the task even if the destination is changed or the path is blocked. Further, the learning efficiency of the robot in the EMQL algorithm is improved by the dual reward mechanism consisting of static reward and dynamic reward. The static reward is designed to prevent the robot from exploring a state node excessively. The dynamic reward is responsible for helping the robot avoid searching blindly in unknown environment. We verify the effectiveness of the proposed algorithm on the grid maps and road network maps. The comparison about planning time, iteration times and path length show that the performance of the EMQL algorithm is superior to Q-learning algorithm in convergence speed and optimization ability. Additionally, the practicability of the proposed algorithm is further validated in a real-world experiment using the Turtlebot3 burger robot.
- 3 March 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 8, 47824-47844
Open AccessThe Experience-Memory Q-Learning Algorithm for Robot Path Planning in Unknown Environment
In order to solve the problem of slow convergence rate and long planned path when the robot plans a path in unknown environment by using Q-learning algorithm, we propose an Experience-Memory Q-Learning (EMQL) algorithm based on the continuous update of the shortest distance from the current state node to the start point. The autonomous learning ability of the robot is enhanced by the different role assignment of two tables in the proposed algorithm. EM table with (m * 1) dimension is designed to record the distance information, reflecting the learning process of the robot. Q table is adopted as an auxiliary guidance for the experience transfer strategy and experience reuse strategy, and these strategies enable the robot accomplish the task even if the destination is changed or the path is blocked. Further, the learning efficiency of the robot in the EMQL algorithm is improved by the dual reward mechanism consisting of static reward and dynamic reward. The static reward is designed to prevent the robot from exploring a state node excessively. The dynamic reward is responsible for helping the robot avoid searching blindly in unknown environment. We verify the effectiveness of the proposed algorithm on the grid maps and road network maps. The comparison about planning time, iteration times and path length show that the performance of the EMQL algorithm is superior to Q-learning algorithm in convergence speed and optimization ability. Additionally, the practicability of the proposed algorithm is further validated in a real-world experiment using the Turtlebot3 burger robot.