检索结果-内蒙古大学图书馆

Proceedings of the 38th International Conference on Neural Information Processing Systems

作者： Weixin An Yuanyuan Liu Fanhua Shang Hongying Liu Key Laboratory of Intelligent Perception and Image Understanding of Ministry of Education School of Artificial Intelligence Xidian University China College of Intelligence and Computing Tianjin University China Medical School Tianjin University China and Peng Cheng Lab Shenzhen China

ISBN: (纸本)9798331314385

Many zeroth-order (ZO) optimization algorithms have been developed to solve nonconvex minimax problems in machine learning and computer vision areas. However, existing ZO minimax algorithms have high complexity and rely on some strict restrictive conditions for ZO estimations. To address these issues, we design a new unified ZO gradient descent extragradient ascent (ZO-GDEGA) algorithm, which reduces the overall complexity to O(dε-6) to find an ε-stationary point of the function ψ for nonconvex-concave (NC-C) problems, where d is the variable dimension. To the best of our knowledge, ZO-GDEGA is the first ZO algorithm with complexity guarantees to solve stochastic NC-C problems. Moreover, ZO-GDEGA requires weaker conditions on the ZO estimations and achieves more robust theoretical results. As a by-product, ZO-GDEGA has advantages on the condition number for the NC-strongly concave case. Experimentally, ZO-GDEGA can generate more effective poisoning attack data with an average accuracy reduction of 5%. The improved AUC performance also verifies the robustness of gradient estimations.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Image Segmentation Refinement Based on Region Expansion and Minor Contour Adjustments

SSRN

引用

SSRN 2024年

作者： Yan, Li-Yue Zhang, Xing Xiong, Si-Ting Song, Qing-Yun Zhang, De-Jin School of Architecture & Urban Planning Shenzhen University Shenzhen518061 China Guangdong Key Laboratory of Urban Spatial Information Engineering Shenzhen518061 China Key Laboratory of Greater Bay Area Geographic Environment Monitoring Ministry of Natural Resources Shenzhen518061 China Shenzhen Key Laboratory of Spatial Information Intelligent Perception and Service Shenzhen518061 China Guangdong Provincial Laboratory of Artificial Intelligence and Digital Economy Shenzhen518132 China College of Computer Science and Software Engineering Shenzhen University Shenzhen518061 China

In high-precision image segmentation tasks, even slight deviations in the segmentation results can bring about significant consequences, especially in certain application areas such as in medical imaging and remote sensing image classification. The precision of segmentation has become the main factor limiting its development. Researchers typically refine image segmentation algorithms to enhance accuracy, but it is challenging for any improvement strategy to be applied effectively across different segmentation methods. It is also challenging to maintain consistent accuracy gains across different objects and image types. In order to solve this problem, in this article the further refinement of image segmentation results is proposed through the two-step refinement strategy of region expansion and minor contour adjustments. First, a dynamic gradient segmentation method dominated by closed contour position is proposed, which provides adaptive gradient constraints for the refinement algorithm. Then, region expansion is used to refine the target region iteratively with the help of color difference comparison and gradient information, while the minor contour adjustments are employed mainly to compare the intensity of gradients, assisted by regional colors, and finally adjust the contour position. This method takes into account both local region and pixel-level information, and adopts the form of dynamic judgment of adjacent region competition. The results show that the refined closed contour is closer to the ground truth. The deviation of contour turning points (corner points) has the most obvious improvement effect. The accuracy of image segmentation is improved. © 2024, The Authors. All rights reserved.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

RenderIH: A Large-scale Synthetic Dataset for 3D Interacting Hand Pose Estimation

RenderIH: A Large-scale Synthetic Dataset for 3D Interacting...

引用

International Conference on Computer Vision (ICCV)

作者： Lijun Li Linrui Tian Xindi Zhang Qi Wang Bang Zhang Liefeng Bo Mengyuan Liu Chen Chen Alibaba Group Shanghai Artificial Intelligence Laboratory Key Laboratory of Machine Perception Shenzhen Graduate School Peking University Center for Research in Computer Vision University of Central Florida

The current interacting hand (IH) datasets are relatively simplistic in terms of background and texture, with hand joints being annotated by a machine annotator, which may result in inaccuracies, and the diversity of pose distribution is limited. However, the variability of background, pose distribution, and texture can greatly influence the generalization ability. Therefore, we present a large-scale synthetic dataset –RenderIH– for interacting hands with accurate and diverse pose annotations. The dataset contains 1M photo-realistic images with varied backgrounds, perspectives, and hand textures. To generate natural and diverse interacting poses, we propose a new pose optimization algorithm. Additionally, for better pose estimation accuracy, we introduce a transformer-based pose estimation network, TransHand, to leverage the correlation between interacting hands and verify the effectiveness of RenderIH in improving results. Our dataset is model-agnostic and can improve more accuracy of any hand pose estimation method in comparison to other real or synthetic datasets. Experiments have shown that pretraining on our synthetic data can significantly decrease the error from 6.76mm to 5.79mm, and our Transhand surpasses contemporary methods. Our dataset and code are available at https://***/adwardlee/RenderIH.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Style transfer-enabled Sim2Real framework for efficient learning of robotic transesophageal echocardiography using simulation

引用

Procedia Computer Science 2023年 226卷 99-105页

作者： keyu Li Xinyu Mao Chengwei Ye Ang Li Yangxin Xu Max Q.-H. Meng Department of Electronic Engineering The Chinese University of Hong Kong Hong Kong China Shenzhen Key Laboratory of Robotics Perception and Intelligence and the Department of Electronic and Electrical Engineering Southern University of Science and Technology Shenzhen 518055 China Shenzhen Research Institute The Chinese University of Hong Kong Shenzhen 518057 China

Recently, various machine learning techniques have been proposed to realize automatic ultrasound (US) image analysis for robotic US acquisition tasks. However, obtaining large amounts of real US images for training is usually expensive or even infeasible in some clinical applications, such as transesophageal echocardiography (TEE). An alternative is to build a simulator to generate synthetic US data for training, but the differences between simulated and real US images may result in poor model performance. This work presents a novel Sim2Real framework to learn robotic US image analysis based on simulated data to address the challenges of limited real US data for robotic TEE. A style transfer model is proposed based on unsupervised contrastive learning to convert real US images into the simulation style. Moreover, a task-relevant model is designed to combine CNNs with vision transformers and trained only on the simulated data to generate task-dependent predictions. We demonstrate the effectiveness of our method in an image regression task to predict the probe position based on US images in robotic TEE. Our results show that using only simulated US data and a small amount of unlabelled real data for training, our method can achieve comparable performance to semi-supervised and fully supervised learning methods. Moreover, the effectiveness of our previously proposed CT-based US image simulation method is also indirectly confirmed.

关键词： Medical robots and systems robotic ultrasound Transesophageal echocardiography

来源：评论

学校读者我要写书评

暂无评论

Multi-robot Path Planning with Rapidly-exploring Random Disjointed-Trees

arXiv

引用

arXiv 2023年

作者： Zhang, Biru Wang, Jiankun Meng, Max Q.-H. Shenzhen Key Laboratory of Robotics Perception and Intelligence Shenzhen China Department of Electronic and Electrical Engineering Southern University of Science and Technology Shenzhen China Jiaxing Research Institute Southern University of Science and Technology Jiaxing China

Multi-robot path planning is a computational process involving finding paths for each robot from its start to the goal while ensuring collision-free operation. It is widely used in robots and autonomous driving. However, the computational time of multi-robot path planning algorithms is enormous, resulting in low efficiency in practical applications. To address this problem, this article proposes a novel multi-robot path planning algorithm (Multi-Agent Rapidly-exploring Random Disjointed-Trees*, MA-RRdT*) based on multi-tree random sampling. The proposed algorithm is based on a single-robot path planning algorithm (Rapidly-exploring Random disjointed-Trees*, RRdT*). The novel MA-RRdT* algorithm has the advantages of fast speed, high space exploration efficiency, and suitability for complex maps. Comparative experiments are completed to evaluate the effectiveness of MA-RRdT*. The final experimental results validate the superior performance of the MA-RRdT* algorithm in terms of time cost and space exploration efficiency. Copyright © 2023, The Authors. All rights reserved.

关键词： Space research

来源：评论

学校读者我要写书评

暂无评论

Neural Ordinary Differential Equation for Irregular Human Motion Prediction

SSRN

引用

SSRN 2023年

作者： Chen, Yang Liu, Hong Song, Pinhao Li, Wenhao Key Laboratory of Machine Perception Peking University Shenzhen Graduate School Shenzhen518055 China School of Artificial Intelligence Chongqing University of Technology Chongqing401135 China Peking University China

Human motion prediction often assumes that the input sequence is of fixed frame rates. However, in real-world applications, the motion capture system may work unstably sometimes and miss some frames, which leads to inferior performance. To solve this problem, this paper leverages neural Ordinary Differential Equations and proposes a human Motion Prediction method named MP-ODE to handle irregular-time human pose series. First, a Difference Operator and a Positional Encoding are proposed to explicitly provide the kinematic and time information for the ***, we construct the encoder-decoder model with ODE-GRU unit, which enables us to learn continuous-time dynamics of human motion. Third, a Quaternion Loss transforms exponential maps to quaternion to train MP-ODE. The Quaternion Loss can avoid the discontinuities and singularities of exponential maps, boosting the convergence of the *** experiments on Human3.6m and CMU-Mocap datasets demonstrate that the proposed MP-ODE achieves promising performance in both normal and irregular-time conditions. © 2023, The Authors. All rights reserved.

关键词： Ordinary differential equations

来源：评论

学校读者我要写书评

暂无评论

MULTI-MODAL AGENT TUNING: BUILDING A VLM-DRIVEN AGENT FOR EFFICIENT TOOL USAGE

arXiv

引用

arXiv 2024年

作者： Gao, Zhi Zhang, Bofei Li, Pengxiang Ma, Xiaojian Yuan, Tao Fan, Yue Wu, Yuwei Jia, Yunde Zhu, Song-Chun Li, Qing School of Intelligence Science and Technology Peking University China State Key Laboratory of General Artificial Intelligence BIGAI China Beijing Key Laboratory of Intelligent Information Technology Beijing Institute of Technology China Guangdong Laboratory of Machine Perception and Intelligent Computing Shenzhen MSU-BIT University China Department of Automation Tsinghua University China

The advancement of large language models (LLMs) prompts the development of multi-modal agents, which are used as a controller to call external tools, providing a feasible way to solve practical tasks. In this paper, we propose a multi-modal agent tuning method that automatically generates multi-modal tool-usage data and tunes a vision-language model (VLM) as the controller for powerful tool-usage reasoning. To preserve the data quality, we prompt the GPT-4o mini model to generate queries, files, and trajectories, followed by query-file and trajectory verifiers. Based on the data synthesis pipeline, we collect the MM-Traj dataset that contains 20K tasks with trajectories of tool usage. Then, we develop the T3-Agent via Trajectory Tuning on VLMs for Tool usage using MM-Traj. Evaluations on the GTA and GAIA benchmarks show that the T3-Agent consistently achieves improvements on two popular VLMs: MiniCPM-V-8.5B and Qwen2-VL-7B, which outperforms untrained VLMs by 20%, showing the effectiveness of the proposed data synthesis pipeline, leading to high-quality data for tool-usage capabilities. Copyright © 2024, The Authors. All rights reserved.

关键词： Visual languages

来源：评论

学校读者我要写书评

暂无评论

robotic Picking Up of Polar Body for Biopsy: Localization and Approach

Robotic Picking Up of Polar Body for Biopsy: Localization an...

引用

Chinese Control Conference (CCC)

作者： Minghui Li Jinyu Qiu Ruimin Li Yidi Zhang Mingzhu Sun Xin Zhao Qili Zhao Institute of Robotics and Automatic Information System (IRAIS) Nankai University Tianjin China Tianjin Key Laboratory of Intelligent Robotic (tjKLIR) Nankai University Tianjin China Institute of Intelligence Technology and Robotic Systems Shenzhen Research Institute of Nankai University Shenzhen China

Polar body biopsy has been widely applied in the preimplantation genetic diagnosis and oocyte screen for assisted reproductive technology. The key step in the polar body biopsy is picking up the polar body using a micropipette. However, the almost transparent appearance of the polar body and its dynamic motion with the micropipette moving inside the oocyte makes it a challenging task to pick up the polar body. In this paper, a defocus imaging method is proposed to locate the polar body with an almost transparent appearance. Then, the dynamic drift of the polar body with the micropipette moving inside the cell is analyzed to determine a trajectory of the micropipette suitable for picking up the polar body. Finally, a PID-based motion control of the micropipette is developed to move the micropipette tip along the above trajectory to approach the polar body. The experimental results on porcine oocytes show that the success rate of polar body localization based on defocus imaging technology reaches 90%. Moving along the determined trajectory inside the oocyte, the micropipette is capable of approaching the polar body with an average error less than 5 μm, success rate can reach 90% (18/20), paving the solid ground for robotic picking up of the polar body.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Robust Pseudo-label Learning with Neighbor Relation for Unsupervised Visible-Infrared Person Re-Identification 24

Robust Pseudo-label Learning with Neighbor Relation for Unsu...

引用

32nd ACM International Conference on Multimedia, MM 2024

作者： Yin, Xiangbo Shi, Jiangming Zhang, Yachao Lu, Yang Zhang, Zhizhong Xie, Yuan Qu, Yanyun School of Informatics Xiamen University Xiamen China Institute of Artificial Intelligence Xiamen University Xiamen China Tsinghua Shenzhen International Graduate School Tsinghua University Shenzhen China Key Laboratory of Multimedia Trusted Perception and Efficient Computing Ministry of Education of China Xiamen University Xiamen China School of Computer Science and Technology East China Normal University Shanghai China East China Normal University Chongqing Institute of East China Normal University Shanghai China

ISBN: (纸本)9798400706868

Unsupervised Visible-Infrared Person Re-identification (USVI-ReID) presents a formidable challenge, which aims to match pedestrian images across visible and infrared modalities without any annotations. Recently, clustered pseudo-label methods have become predominant in USVI-ReID, although the inherent noise in pseudo-labels presents a significant obstacle. Most existing works primarily focus on shielding the model from the harmful effects of noise, neglecting to calibrate noisy pseudo-labels usually associated with hard samples, which will compromise the robustness of the model. To address this issue, we design a Robust Pseudo-label Learning with Neighbor Relation (RPNR) framework for USVI-ReID. To be specific, we first introduce a straightforward yet potent Noisy Pseudo-label Calibration module to correct noisy pseudo-labels. Due to the high intra-class variations, noisy pseudo-labels are difficult to calibrate completely. Therefore, we introduce a Neighbor Relation Learning module to reduce high intra-class variations by modeling potential interactions between all samples. Subsequently, we devise an Optimal Transport Prototype Matching module to establish reliable cross-modality correspondences. On that basis, we design a Memory Hybrid Learning module to jointly learn modality-specific and modality-invariant information. Comprehensive experiments conducted on two widely recognized benchmarks, SYSU-MM01 and RegDB, demonstrate that RPNR outperforms the current state-of-the-art GUR with an average Rank-1 improvement of 10.3%. The code is available at https://***/XiangboYin/RPNR. © 2024 ACM.

关键词： Unsupervised learning

来源：评论

学校读者我要写书评

暂无评论

Disturbance Rejection Control for Autonomous Trolley Collection Robots with Prescribed Performance

arXiv

引用

arXiv 2023年

作者： Xi, Rui-Dong Lu, Liang Zhang, Xue Xiao, Xiao Xia, Bingyi Wang, Jiankun Meng, Max Q.-H. Shenzhen Key Laboratory of Robotics Perception and Intelligence the Department of Electronic and Electrical Engineering Southern University of Science and Technology Shenzhen518055 China Jiaxing Research Institute Southern University of Science and Technology Jiaxing China Yuanhua Robotics Perception & AI Technologies Ltd Shenzhen518055 China

Trajectory tracking control of autonomous trolley collection robots (ATCR) is an ambitious work due to the complex environment, serious noise and external disturbances. This work investigates a control scheme for ATCR subjecting to severe environmental interference. A kinematics model based adaptive sliding mode disturbance observer with fast convergence is first proposed to estimate the lumped disturbances. On this basis, a robust controller with prescribed performance is proposed using a backstepping technique, which improves the transient performance and guarantees fast convergence. Simulation outcomes have been provided to illustrate the effectiveness of the proposed control scheme. Copyright © 2023, The Authors. All rights reserved.

关键词： Skids

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：