检索结果-内蒙古大学图书馆

2021 IEEE International Conference on Real-Time Computing and Robotics, RCAR 2021

作者： Wu, Zhiheng Wu, Zhengxing Lu, Yue Wang, Jian Yu, Junzhi Institute of Automation Chinese Academy of Sciences State Key Laboratory of Management and Control for Complex Systems Beijing100190 China School of Artificial Intelligence University of Chinese Academy of Sciences Beijing100049 China State Key Laboratory for Turbulence and Complex Systems Department of Advanced Manufacturing and Robotics BIC-ESAT College of Engineering Peking University Beijing100871 China

ISBN: (纸本)9781665436786

With the rapid development of deep neural networks, underwater vision plays an increasingly important role in the underwater robotic operation. However, the scarce underwater datasets greatly limit the performance of deep learning on underwater visual tasks, further hindering the applications of underwater operation. To solve this problem, we propose an underwater image synthesis method, which can directly convert the natural light image into the synthetic underwater image end-to-end. Particularly, a pixel-level self-supervised training strategy is designed to maximize the structural similarity between the synthesized and real images, through training the real underwater images. Finally, extensive experiments are carried out, and the obtained results demonstrate the effectiveness and superiority of our methods by quantitative and qualitative comparisons. The proposed underwater image synthesis method offers a valuable sight for underwater vision and manipulating control. © 2021 IEEE.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Hierarchical Reinforcement Learning-Based Policy Switching Towards Multi-Scenarios Autonomous Driving

Hierarchical Reinforcement Learning-Based Policy Switching T...

引用

International Joint Conference on Neural Networks (IJCNN)

作者： Guo, Youtian Zhang, Qichao Wang, Junjie Liu, Shasha Chinese Acad Sci Inst Automat State Key Lab Management & Control Complex Syst Beijing Peoples R China Univ Chinese Acad Sci Sch Artif Intelligence Beijing Peoples R China

ISBN: (纸本)9780738133669

Reinforcement learning has gained a lot of attention and applications in the field of autonomous driving in recent years. However, the actual scenarios of automatic driving applications are often diverse, so the reinforcement learning algorithm using only a single driving strategy is difficult to meet the multiple requirements of efficiency and safety in the multi-scenarios autonomous driving task. To solve this challenge, we propose a hierarchical reinforcement learning structure to learn a unified top-level switching master policy between different driving styles policies. The whole framework uses a bottom-up training manner with diverse reward function designing. Through experimental comparison, our method exceeds the performance of single policy and rule-based switching strategy. Based on this framework, we won the first place in the DAI 2020 Autonomous Driving Workshop single-agent track competition.

关键词： Autonomous Driving Hierarchical Reinforcement Learning Semi-Markov Decision Process

来源：评论

学校读者我要写书评

暂无评论

A Reinforcement Learning Benchmark for Autonomous Driving in Intersection Scenarios

A Reinforcement Learning Benchmark for Autonomous Driving in...

引用

IEEE Symposium Series on Computational Intelligence (IEEE SSCI)

作者： Liu, Yuqi Zhang, Qichao Zhao, Dongbin Chinese Acad Sci Inst Automat State Key Lab Management & Control Complex Syst Beijing Peoples R China Univ Chinese Acad Sci Sch Artificial Intelligence Beijing Peoples R China

ISBN: (纸本)9781728190488

In recent years, control under urban intersection scenarios has become an emerging research topic. In such scenarios, the autonomous vehicle confronts complicated situations since it must deal with the interaction with social vehicles timely while obeying the traffic rules. Generally, the autonomous vehicle is supposed to avoid collisions while pursuing better efficiency. The existing work fails to provide a framework that emphasizes the integrity of the scenarios while deploying and testing reinforcement learning(RL) methods. Specifically, we propose a benchmark for training and testing RL-based autonomous driving agents in complex intersection scenarios, which is called RL-CIS. Then, a set of baselines consisting various algorithms are deployed. The test benchmark and baselines provide a fair and comprehensive training and testing platform for the study of RL for autonomous driving in the intersection scenario, advancing RL-based methods for autonomous driving control. The code of our proposed framework can be found at https://***/liuyuqi123/complexUrbanScenarios.

关键词： autonomous driving reinforcement learning intersection scenarios decision-making

来源：评论

学校读者我要写书评

暂无评论

Notice of Retraction: Cascade Network with Boundary-Adaptive Filter for Surgical Gesture Recognition

Notice of Retraction: Cascade Network with Boundary-Adaptive...

引用

2021 International Joint Conference on Neural Networks, IJCNN 2021

作者： Chen, Hua-Bin Ni, Zhen-Liang Li, Zhen Bian, Gui-Bin Institute of Automation Chinese Academy of Sciences State Key Laboratory of Management and Control for Complex Systems Beijing 100190 China University of Chinese Academy of Sciences Beijing 100049 China

ISBN: (纸本)9780738133669

Automatically recognizing surgical gestures plays a key role in computer-assisted surgery and automatic skill assessment. Solutions to this task that only rely on video without additional sensor hardware have attracted great attention since they are economical and feasible. However, it is still a challenging problem due to the variable gesture durations and the oversegmentation problem that occurred in a long gesture duration. To address the above problems, we propose a cascade network with boundary-adaptive filter to dynamically capture the long-range temporal dependencies and reduce the over-segmentation errors obviously. A cascade module is designed to enhance the temporal modeling abilities in different timescales, in which multiple similar temporal convolution networks are divided into cascade stages and a fusion stage. Every cascade stage with different temporal receptive fields has a frame-specific weight to adjust the distribution of each stage and they will be aggregated into the fusion stage to refine the predictions. The frame-specific model architecture may incur a more serious over-segmentation problem. A boundary-adaptive filter is designed to leverage gesture boundary information to smooth the prediction results adaptively, which maintains consistency within a gesture duration while keeping the gesture-crossing discrepancy. We evaluate our model on the JHU-ISI Gesture and Skill Assessment Working Set (JIGSAWS). The proposed method reaches the state-of-the-art frame-wise accuracy (83.79%) and comparable segment-level performance (83.35% F1@50). © 2021 IEEE.

关键词： Gesture recognition

来源：评论

学校读者我要写书评

暂无评论

Traffic Flow Synthesis Using Generative Adversarial Networks via Semantic Latent Codes Manipulation

Traffic Flow Synthesis Using Generative Adversarial Networks...

引用

IEEE Intelligent Transportation systems Conference (ITSC)

作者： Chen, Yuanyuan Lv, Yisheng Zhu, Fenghua Chinese Acad Sci Inst Automat State Key Lab Management & Control Complex Syst Beijing 100190 Peoples R China Univ Chinese Acad Sci Beijing Beijing 100190 Peoples R China

ISBN: (纸本)9781728191423

Traffic data analysis and mining are elemental functions of Intelligent Transportation systems. In recent year, tremendous sensors are deployed in order to collect big data, and equipment maintenance costs a lot. With the development of deep learning, especially especially Generative Adversarial Networks, we can generate realistic big artificial traffic flow data and use small real traffic data and synthesized traffic data in traffic data mining tasks. In this paper, we focus on discovering the semantics embedded in latent codes which are fed into Generative Adversarial Networks, and propose to use the interpolation of semantic latent code to generate semantic manipulation of traffic flow. We evaluate our approach using the publicly available data from Caltrans Performance Measurements systems (PeMS), and experimental results show the the effectiveness of the proposed method.

关键词： Training Measurement Interpolation Codes Semantics Maintenance engineering Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Benchmarking Lane-changing Decision-making for Deep Reinforcement Learning 7

Benchmarking Lane-changing Decision-making for Deep Reinforc...

引用

7th International Conference on Robotics and Artificial Intelligence (ICRAI)

作者： Wang, Junjie Zhang, Qichao Zhao, Dongbin Chinese Acad Sci Inst Automat State Key Lab Management & Control Complex Syst Beijing Peoples R China Univ Chinese Acad Sci Coll Artificial Intelligence Beijing Peoples R China

ISBN: (纸本)9781450385855

The development of autonomous driving has attracted extensive attention in recent years, and it is essential to evaluate the performance of autonomous driving. However, testing on the road is expensive and inefficient. Virtual testing is the primary way to validate and verify self-driving cars, and the basis of virtual testing is to build simulation scenarios. In this paper, we propose a training, testing, and evaluation pipeline for the lane-changing task from the perspective of deep reinforcement learning. First, we design lane change scenarios for training and testing, where the test scenarios include stochastic and deterministic parts. Then, we deploy a set of benchmarks consisting of learning and non-learning approaches. We train several state-of-the-art deep reinforcement learning methods in the designed training scenarios and provide the benchmark metrics evaluation results of the trained models in the test scenarios. The designed lane-changing scenarios and benchmarks are both opened to provide a consistent experimental environment for the lane-changing task(1).

关键词： Autonomous driving Reinforcement learning Lane change scenarios

来源：评论

学校读者我要写书评

暂无评论

Parallel medical imaging: An ACP-based approach for intelligent medical image recognition with small samples 1

Parallel medical imaging: An ACP-based approach for intellig...

引用

1st IEEE International Conference on Digital Twins and Parallel Intelligence, DTPI 2021

作者： Shen, Tianyu Gou, Chao Wang, Jiangong Huang, Jun He, Yonglan Xue, Huadan Jin, Zhengyu Wang, Fei-Yue Institute of Automation Chinese Academy of Sciences State Key Laboratory of Management and Control for Complex Systems Beijing China Shenzhen Campus of Sun Yat-sen University Sun Yat-sen University Guangzhou China Peking Union Medical College Hospital Chinese Academy of Medical Sciences Peking Union Medical College Department of Radiology Beijing China Institute of Automation Chinese Academy of Sciences State Key Laboratory of Management and Control for Complex Systems Beijing China

ISBN: (纸本)9781665433372

The deep learning methods trained with large-scale manually annotated datasets have led to significant breakthroughs in medical image community. However, obtaining such datasets remains a challenging work in medical domain. In this paper, we propose an ACP-based approach named parallel medical imaging (PMI) for analyzing and addressing the small sample problem of medical image recognition. Firstly, we present the basic framework and key techniques of PMI, in which the Artificial imaging systems are constructed to effectively augment the 'medical small sample', the Computational experiments are carried out for designing and evaluating the visual models trained with small samples, and the Parallel execution is conducted to achieve closed-loop interaction and optimization in the domain-knowledge guidance. Furthermore, we illustrate a preliminary implementation and application of PMI approach through the case study of mammogram analysis. © 2021 IEEE.

关键词： Medical imaging

来源：评论

学校读者我要写书评

暂无评论

A Prototype-Based Generalized Zero-Shot Learning Framework for Hand Gesture Recognition 25

A Prototype-Based Generalized Zero-Shot Learning Framework f...

引用

25th International Conference on Pattern Recognition (ICPR)

作者： Wu, Jinting Zhang, Yujia Zhao, Xiaoguang Chinese Acad Sci Inst Automat State Key Lab Management & Control Complex Syst Beijing Peoples R China Univ Chinese Acad Sci Sch Artificial Intelligence Beijing Peoples R China

ISBN: (纸本)9781728188089

Hand gesture recognition plays a significant role in human-computer interaction for understanding various human gestures and their intent. However, most prior works can only recognize gestures of limited labeled classes and fail to adapt to new categories. The task of Generalized Zero-Shot Learning (GZSL) for hand gesture recognition aims to address the above issue by leveraging semantic representations and detecting both seen and unseen class samples. In this paper, we propose an end-to-end prototype-based GZSL framework for hand gesture recognition which consists of two branches. The first branch is a prototype-based detector that learns gesture representations and determines whether an input sample belongs to a seen or unseen category. The second branch is a zero-shot label predictor which takes the features of unseen classes as input and outputs predictions through a learned mapping mechanism between the feature and the semantic space. We further establish a hand gesture dataset that specifically targets this GZSL task, and comprehensive experiments on this dataset demonstrate the effectiveness of our proposed approach on recognizing both seen and unseen gestures.

关键词： Zero-shot learning

来源：评论

学校读者我要写书评

暂无评论

Guest Editorial on "Computational intelligence in analysis and integration of complex systems"

引用

complex & INTELLIGENT systems 2022年第3期8卷 1819-1821页

作者： Zhao, Bo Zeng, Wenyi Gao, Weinan Zhang, Qichao Beijing Normal Univ Sch Syst Sci Beijing 100875 Peoples R China Beijing Normal Univ Sch Artificial Intelligence Beijing 100875 Peoples R China Florida Inst Technol Coll Engn & Sci Melbourne FL 32901 USA Chinese Acad Sci Inst Automat State Key Lab Management & Control Complex Syst Beijing 100190 Peoples R China

来源：评论

学校读者我要写书评

暂无评论

Detection of Driver Fatigue state using Deep Neural Network 16

Detection of Driver Fatigue State using Deep Neural Network

引用

16th IEEE Conference on Industrial Electronics and Applications (ICIEA)

作者： Anwar, Noreen Xiong, Gang Guo, Miao Ye, Peijun Ali, Hub Wei, Qinglai Chinese Acad Sci Inst Automat State Key Lab Management & Control Complex Syst Beijing Peoples R China Beijing Univ Technol Sch Artificial Intelligence & Automat Beijing Peoples R China

ISBN: (纸本)9781665422482

Image-based driver fatigue detection remains a challenging problem due to occlusion of face, the variation of head poses and illuminations. This paper implements an effectual technique for investigating the driver's fatigue state by using infrared image of an eye in the open or closed condition. In this method we use the deep learning technique to monitor the change, i.e., open and closed conditions of eye state. We integrate ResNet and depthwise convolution network together and use as the core of the structure of the network to perform face and facial landmark detection tasks. After acquiring the eye region, we perform the eye state identification task by using its coordinates of feature points. To determine fatigue, we use PERCLOS method and the results confirm accuracy and effectiveness of the algorithm by comparing with other existing methods. We can reach an accuracy of 97.2% and the average time is 31.20 milliseconds, represent that this driver monitoring inference system has significant importance for both the traffic and driver's safety.

关键词： driver fatigue detection deep neural network ResNet depthwise convolution network PERCLOS

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：