检索结果-内蒙古大学图书馆

Boosting Fine-Grained Visual Anomaly Detection with Coarse-Knowledge-Aware Adversarial Learning 39

学校读者我要写书评

暂无评论

Boosting Fine-Grained Visual Anomaly Detection with Coarse-K...

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Fang, Qingqing Su, Qinliang Lv, Wenxi Xu, Wenchao Yu, Jianxing School of Computer Science and Engineering Sun Yat-sen University Guangzhou China Guangdong Key Laboratory of Big Data Analysis and Processing Guangzhou China Department of Computing The Hong Kong Polytechnic University Hong Kong School of Artificial Intelligence Sun Yat-sen University Guangdong China Pazhou Lab Guangzhou510330 China

ISBN: (纸本)157735897X

Many unsupervised visual anomaly detection methods train an auto-encoder to reconstruct normal samples and then leverage the reconstruction error map to detect and localize the anomalies. However, due to the powerful modeling and generalization ability of neural networks, some anomalies can also be well reconstructed, resulting in unsatisfactory detection and localization accuracy. In this paper, a small coarsely-labeled anomaly dataset is first collected. Then, a coarse-knowledge-aware adversarial learning method is developed to align the distribution of reconstructed features with that of normal features. The alignment can effectively suppress the auto-encoder’s reconstruction ability on anomalies and thus improve the detection accuracy. Considering that anomalies often only occupy very small areas in anomalous images, a patch-level adversarial learning strategy is further developed. Although no patch-level anomalous information is available, we rigorously prove that by simply viewing any patch features from anomalous images as anomalies, the proposed knowledge-aware method can also align the distribution of reconstructed patch features with the normal ones. Experimental results on four medical datasets and two industrial datasets demonstrate the effectiveness of our method in improving the detection and localization performance. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Adversarial machine learning

Modeling of Optimal Deep Learning Enabled Object Detection and Classification on Drone Imagery

学校读者我要写书评

暂无评论

Modeling of Optimal Deep Learning Enabled Object Detection a...

2022 International Conference on Augmented Intelligence and Sustainable Systems, ICAISS 2022

作者： Adhikari, Nirmal Behera, Nihar Ranjan Vijayakrishna Rapaka, E. Pimo, Er. S. John Chaturvedi, Vaibhav Tripathi, Vikas School of Built Environment Engineering & Computing. Leeds Beckett University LeedsLS1 3HE United Kingdom Swiss School of Business and Management Geneva Av. des Morgines 12 Petit-Lancy1213 Switzerland Data Science Indian Institute of Technology Chennai Madras India St. Xavier's Catholic College of Engineering Department of CSE Chunkankadai India Rajkiya Engineering College Department of Civil Kannauj India Graphic Era Deemed to Be University Department of Computer Science & Engineering Uttarakhand Dehradun India

ISBN: (纸本)9781665489621

Object detection in unmanned aerial vehicle (UAV) images becomes a persistent problem in the domain of computer vision. Particularly, object detection in drone images is a difficult process because of the object of different scales namely, hills, buildings, and water bodies. The study presents an execution of ensemble transfer learning to improve the efficiency of the fundamental model for multi-scale object recognition in drone imagery. This study develops an Optimal Deep Learning Enabled Object Detection and Classification on Drone Imagery (ODL-ODCDI) technique. The presented ODL-ODCDI technique can recognize and classify the objects present in the images collected by the drones. It follows a two stage process. In the first level, the ODL-ODCDI technique employed YOLO-v5 as object detector with Nadam optimizer. Next, in the latter level, the ODL-ODCDI technique makes use of random forest (RF) classifier to identify objects in the drone images. To establish the enhanced performance of the ODL-ODCDI approach, a series of experiments were performed. The experimental values depicted the improved outcomes of the ODL-ODCDI method over other DL models. © 2022 IEEE.

关键词： Object detection

Safe Reinforcement Learning for Social Motor Coordination in Two Dimensional Space 26

学校读者我要写书评

暂无评论

Safe Reinforcement Learning for Social Motor Coordination in...

26th International Conference on Industrial Technology, ICIT 2025

作者： Yang, Fan Huang, Yuening Zhai, Chao Zhang, Hehong China University of Geosciences School of Automation Wuhan Campus China Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems China Engineering Research Center of Intelligent Technology for Geo-Exploration Ministry of Education Wuhan China Rice University George R. Brown School of Engineering and Computing HoustonTX United States Fuzhou University School of Computer and Data Science Fuzhou China

ISBN: (纸本)9798331521950

This paper presents an interactive motion control method based on reinforcement learning, designed to assist children with autism who have social motor impairments through a mirror game intervention. The virtual teacher uses the Deep Deterministic Policy Gradient (DDPG) algorithm to optimize its actions, guiding the participant to follow a Lissajous trajectory. To ensure safety, a motion-correction mechanism was developed, which automatically adjusts actions when the predicted trajectory surpasses predefined safety boundaries. The reward function considers both the distance between the virtual teacher and the target trajectory, as well as the distance between the virtual teacher and the participant, with dynamic adjustments applied by the motion-correction mechanism. Experimental results demonstrate that the virtual teacher effectively guides the participant towards the target trajectory while adhering to safety constraints. © 2025 IEEE.

关键词： Reinforcement learning

User Power Measurement Based IRS Channel Estimation via Single-Layer Neural Network

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Sun, He Mei, Weidong Zhu, Lipeng Zhang, Rui Department of Electrical and Computer Engineering National University of Singapore Singapore117583 Singapore School of Science and Engineering Shenzhen Research Institute of Big Data China Chinese University of Hong Kong Guangdong Shenzhen518172 China National Key Laboratory of Science and Technology on Communications University of Electronic Science and Technology of China 611731 China

One main challenge for implementing intelligent reflecting surface (IRS) aided communications lies in the difficulty to obtain the channel knowledge for the base station (BS)-IRS-user cascaded links, which is needed to design high-performance IRS reflection in practice. Traditional methods for estimating IRS cascaded channels are usually based on the additional pilot signals received at the BS/users, which increase the system training overhead and also may not be compatible with the current communication protocols. To tackle this challenge, we propose in this paper a new single-layer neural network (NN)-enabled IRS channel estimation method based on only the knowledge of users’ individual received signal power measurements corresponding to different IRS random training reflections, which are easily accessible in current wireless systems. To evaluate the effectiveness of the proposed channel estimation method, we design the IRS reflection for data transmission based on the estimated cascaded channels in an IRS-aided multiuser communication system. Numerical results show that the proposed IRS channel estimation and reflection design can significantly improve the minimum received signal-to-noise ratio (SNR) among all users, as compared to existing power measurement based designs. © 2023, CC BY.

关键词： Channel estimation

L2GNet: Optimal Local-to-Global Representation of Anatomical Structures for Generalized Medical Image Segmentation

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Gorade, Vandan Mittal, Sparsh Dasu, Neethi Singhal, Rekha Santosh, K.C. Jha, Debesh Department of Biomedical Engineering Northwestern University IL United States Mehta School of Data Science and AI Indian Institute of Technology Roorkee India Beth Israel Lahey Health University of Massachusetts Chan School of Medicine United States TCS Research New York United States Applied AI Research Lab Department of Computer Science University of South Dakota United States

Continuous Latent Space (CLS) and Discrete Latent Space (DLS) models, like AttnUNet and VQUNet, have excelled in medical image segmentation. In contrast, Synergistic Continuous and Discrete Latent Space (CDLS) models show promise in handling fine and coarse-grained information. However, they struggle with modeling long-range dependencies. CLS or CDLS-based models, such as TransUNet or SynergyNet are adept at capturing long-range dependencies. Since they rely heavily on feature pooling or aggregation using self-attention, they may capture dependencies among redundant regions. This hinders comprehension of anatomical structure content, poses challenges in modeling intra-class and inter-class dependencies, increases false negatives and compromises generalization. Addressing these issues, we propose L2GNet, which learns global dependencies by relating discrete codes obtained from DLS using optimal transport and aligning codes on a trainable reference. L2GNet achieves discriminative on-the-fly representation learning without an additional weight matrix in self-attention models, making it computationally efficient for medical applications. Extensive experiments on multi-organ segmentation and cardiac datasets demonstrate L2GNet’s superiority over state-of-the-art methods, including the CDLS method SynergyNet, offering an novel approach to enhance deep learning models’ performance in medical image analysis. © 2025, CC BY.

关键词： Image segmentation

Cross-Modal Few-Shot Learning with Second-Order Neural Ordinary Differential Equations

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Zhang, Yi Cheng, Chun-Wun He, Junyi He, Zhihai Schönlieb, Carola-Bibiane Chen, Yuyan Aviles-Rivero, Angelica I. Harbin Institute of Technology China Department of Electrical and Electronic Engineering Southern University of Science and Technology China Department of Applied Mathematics and Theoretical Physics University of Cambridge United Kingdom Pengcheng Laboratory China Shanghai Key Laboratory of Data Science School of Computer Science Fudan University China Yau Mathematical Sciences Center Tsinghua University China

We introduce SONO, a novel method leveraging Second-Order Neural Ordinary Differential Equations (Second-Order NODEs) to enhance cross-modal few-shot learning. By employing a simple yet effective architecture consisting of a Second-Order NODEs model paired with a cross-modal classifier, SONO addresses the significant challenge of overfitting, which is common in few-shot scenarios due to limited training examples. Our second-order approach can approximate a broader class of functions, enhancing the model’s expressive power and feature generalization capabilities. We initialize our cross-modal classifier with text embeddings derived from class-relevant prompts, streamlining training efficiency by avoiding the need for frequent text encoder processing. Additionally, we utilize text-based image augmentation, exploiting CLIP’s robust image-text correlation to enrich training data significantly. Extensive experiments across multiple datasets demonstrate that SONO outperforms existing state-of-the-art methods in few-shot learning performance. Copyright © 2024, The Authors. All rights reserved.

关键词： Zero-shot learning

Exploring Reinforcement Learning in Large-Scale data Processing

学校读者我要写书评

暂无评论

Exploring Reinforcement Learning in Large-Scale Data Process...

Smart Generation Computing, Communication and Networking (SMART GENCON), International Conference on

作者： Pawan Bhambu Ritesh Kumar P. Sharmila Vivek D. Patil Shivam Khurana Vivek V Department of Computer Science & Engineering Vivekananda Global University Jaipur India Maharishi School of Engineering and Technology Maharishi University of Information Technology Uttar Pradesh India Department of Science and Humanities Prince Shri Venkateshwara Padmavathy Engineering College Chennai Department of Artificial Intelligence & Data Science Vishwakarma Institute of Information Technology Pune India Centre of Interdisciplinary Research in Business and Technology Chitkara University Institute of Engineering and technology Chitkara University Punjab India Computer Science and Engineering Faculty of Engineering and Technology JAIN (Deemed-to-be University) Bangalore Karnataka India

Reinforcement mastering (RL) is vital for big-scale fast processing. RL utilizes the availability of large datasets and dealers skilled in this information to remedy a wide variety of obligations, from automated diagnostics and self-reliant using structures to clinical information analysis. Using exploiting the traits of the modern-day significant information era, RL allows efficient and correct decision-making in massive-scale information processing and answers complicated troubles. This paper explores the use of reinforcement learning (RL) in large-scale data processing. RL is a subfield of machine learning that focuses on decision making in dynamic environments. With the increasing demand for processing large amounts of data, traditional techniques such as rule-based systems and batch processing are becoming inadequate. This has led to the emergence of RL as a potential solution due to its ability to learn and adapt to changing environments. Through a comprehensive review of existing literature, this paper discusses the applications of RL in large-scale data processing and its potential benefits. Furthermore, it examines various challenges and their potential solutions in implementing RL for data processing tasks. The paper concludes by highlighting the future research directions in this domain and the potential impact of RL on improving data processing efficiency and accuracy. It similarly examines some of the demanding situations and possibilities related to using RL for huge-scale records processing and the capability programs of RL in the future.

关键词：

Beyond CNNs: Exploiting Further Inherent Symmetries in Medical Image Segmentation

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Pang, Shuchao Du, Anan Orgun, Mehmet A. Wang, Yan Sheng, Quan Z. Wang, Shoujin Huang, Xiaoshui Yu, Zhenmei The School of Cyber Science and Engineering Nanjing University of Science and Technology Jiangsu Nanjing210094 China The School of Computing Macquarie University SydneyNSW2109 Australia The School of Electrical and Data Engineering University of Technology SydneyNSW2007 Australia The Data Science Institute University of Technology SydneyNSW2007 Australia Shanghai AI Laboratory Shanghai200433 China The School of Data and Computer Science Shandong Women's University Jinan250014 China

Automatic tumor or lesion segmentation is a crucial step in medical image analysis for computer-aided diagnosis. Although the existing methods based on Convolutional Neural Networks (CNNs) have achieved the state-of-the-art performance, many challenges still remain in medical tumor segmentation. This is because, although the human visual system can detect symmetries in 2D images effectively, regular CNNs can only exploit translation invariance, overlooking further inherent symmetries existing in medical images such as rotations and reflections. To solve this problem, we propose a novel group equivariant segmentation framework by encoding those inherent symmetries for learning more precise representations. First, kernel-based equivariant operations are devised on each orientation, which allows it to effectively address the gaps of learning symmetries in existing approaches. Then, to keep segmentation networks globally equivariant, we design distinctive group layers with layer-wise symmetry constraints. Finally, based on our novel framework, extensive experiments conducted on real-world clinical data demonstrate that a Group Equivariant Res-UNet (named GER-UNet) outperforms its regular CNN-based counterpart and the state-of-the-art segmentation methods in the tasks of hepatic tumor segmentation, COVID-19 lung infection segmentation and retinal vessel detection. More importantly, the newly built GER-UNet also shows potential in reducing the sample complexity and the redundancy of filters, upgrading current segmentation CNNs and delineating organs on other medical imaging modalities. © 2022, CC BY-NC-ND.

关键词： computer aided diagnosis

RETRACTED ARTICLE: DVT: a recent review and a taxonomy for oral and maxillofacial visualization and tracking based augmented reality: image guided surgery

学校读者我要写书评

暂无评论

Multimedia Tools and Applications 2023年第1期83卷 685-729页

作者： Alsadoon, Abeer AlSallami, Nada Rashid, Tarik A. Gosper, Jeffrey J. Prasad, P. W. C. Haddad, Sami School of Computing Mathematics and Engineering Charles Sturt University (CSU) Albury-Wodonga Australia School of Computer Data and Mathematical Sciences Western Sydney University (WSU) Sydney Australia Asia Pacific International College (APIC) Sydney Australia Computer Science Department Worcester State University (WSU) Worcester USA Computer Science and Engineering University of Kurdistan Hewler Erbil Iraq Department of Oral and Maxillofacial Services Greater Western Sydney Area Health Services Mount Druitt Australia Department of Oral and Maxillofacial Services Central Coast Area Health Gosford Australia

Augmented reality (AR) navigation system is increasingly being integrated into Image Guided Surgery systems. The use of AR registration and tracking system in operating rooms (OR) for oral and maxillofacial surgery (OMS) can result in reducing medical errors and decreasing total operation times. As such, AR systems are assisting to reduce some of the surgical complexities associated with OMS, making it easier for the surgeons to view the operation. Although AR systems has been implemented in OR for many years, there are still several factors that are less than optimal and can cause complications, such as inadequate system accuracy, poor image and video quality, high operating time and cost, and significant complexity of the system required to achieve accurate views of the surgical target of OMS. The aim of this research is to improve the use of AR in OMS using the proposed review taxonomy which incorporated data, Visualization, and Tracking (DVT). DVT taxonomy defines the major components that are required to implement in an AR navigation system. Those components are validated and evaluated considering the clear and accurate output or view during craniofacial surgery for the end user. The proposed DVT taxonomy have been considered comparison of system, completeness of system and acceptance of the system as the major criteria. DVT is evaluated and validated our DTV taxonomy by analysing and classifying the 33 state of art publications which work in the AR navigation. This work presents a review over navigational approach towards surgery through AR which highlights the features and usefulness of AR compared to the existing surgical process in terms of processing time, accuracy, efficiency, and feasibility in surgery.

关键词：