检索结果-内蒙古大学图书馆

49th IEEE international conference on Acoustics, Speech, and signal processing (ICASSP)

作者： Liu, Gang Li, Hongyang He, Zerui Zhong, Shenjun Harbin Engn Univ Coll Comp Sci & Technol Harbin Peoples R China Monash Univ Monash Biomed Imaging Melbourne Australia

ISBN: (纸本)9798350344868;9798350344851

Leveraging pre-trained visual language models has become a widely adopted approach for improving performance in downstream visual question answering (VQA) applications. However, in the specialized field of medical VQA, the scarcity of available data poses a significant barrier to achieving reliable model generalization. Numerous methods have been proposed to enhance model generalization, addressing the issue from data-centric and model-centric perspectives. Data augmentation techniques are commonly employed to enrich the dataset, while various regularization approaches aim to prevent model overfitting, especially when training on limited data samples. In this paper, we introduce a method that incorporates gradient-guided parameter perturbations to the visual encoder of the multimodality model during both pre-training and fine-tuning phases, to improve model generalization for downstream medical VQA tasks. The small perturbation is adaptively generated by aligning with the direction of the moving average gradient in the optimization landscape, which is opposite to the directions of the optimizer's historical updates. It is subsequently injected into the model's visual encoder. The results show that, even with a significantly smaller pre-training image caption dataset, our approach achieves competitive outcomes on both VQA-RAD and SLAKE datasets.

关键词： Medical visual question answering Model perturbation vision-language pre-training Model generalization

来源：评论

学校读者我要写书评

暂无评论

An Omni-Directional Self driving robot for Indoor Surveillance

An Omni-Directional Self driving Robot for Indoor Surveillan...

引用

2024 international conference on signal processing, Computation, Electronics, Power and Telecommunication, IConSCEPT 2024

作者： Dhanush, D. Raghul, K.R. Shri Ramya, K.R. Mukund, P.G. Ganesan, M. Amrita School of Engineering Coimbatore Amrita Vishwa Vidyapeetham Department of Electronics and Communication Engineering India

ISBN: (数字)9798331540685

ISBN: (纸本)9798331540685

Indoor surveillance stands as a critical measure for ensuring the safety and well-being of the elderly population, who often face various risks within their living spaces. To address this need effectively, this paper introduces a systematic approach to develop an omnidirectional self-driving robot tailored specifically for indoor surveillance tasks. The robot's design centers on the utilization of three omni wheels, providing holonomic motion and exceptional maneuverability within confined indoor environments. Leveraging the robot Operating System (ROS) as the underlying framework and Jetson Nano as the development board, the integration of complex robotic systems is streamlined, ensuring seamless operation and versatility. With autonomous mobility capabilities, the robot is equipped to navigate indoor spaces autonomously, facilitating continuous monitoring through an integrated camera system. Key surveillance functionalities include detecting falls among the elderly and identifying potential intrusions. Moreover, the modular design of the robotic system enables the incorporation of additional surveillance features, further enhancing its utility and adaptability. Overall, this paper presents a robust hardware solution coupled with advanced functionalities, offering a comprehensive indoor surveillance system poised to address the diverse needs of residential environments. © 2024 IEEE.

关键词： Wheels

来源：评论

学校读者我要写书评

暂无评论

Horticulture robot for Precision Plant Health Management 10

Horticulture Robot for Precision Plant Health Management

引用

10th international conference on Communication and signal processing, ICCSP 2024

作者： Hruday, Pemula Jaxa Kiran, Anakapalli Bharath Venkata Siva Prasad Reddy, V.H. Velagapudi Ramakrishna Siddhartha Engineering College Dept of Electronics and Communication Engineering Vijayawada India

ISBN: (纸本)9798350353068

In the ever-changing field of greenhouse agriculture, unique problems including pest infestations, fungal infections, bacterial diseases, nutrient deficits, and virus threats clearly call for the treatment of unwell plants separately. If left unchecked, these problems can jeopardize entire crops, resulting in lower yields and damaged plants. Using state-of-the-art sensor technology, this abstract presents the 'Advanced Horticulture robot,' a precise solution that can navigate, assess, and care for individual plants. A combination of the assistance of a dynamic camera and ultrasonic sensors, the robot is able to detect obstacles as well as snap clear pictures. Real-time data is interpreted by sophisticated algorithms that detect pests, bacterial infections, fungal infections, nutritional deficiencies, and possible virus threats in particular crops, such as microgreens, bell peppers, tomatoes, cucumbers, lettuce, basil, and chili plants. When the robot detects diseased plants, it implements exact treatment protocols. For example, it will only dispense targeted pesticides and optimize nutrients for chili plants that are susceptible to viruses and pests. This methodical methodology provides a revolutionary solution for precise plant health regulate while limiting the negative effects on the environment, optimizing resource utilization, and supporting sustainable greenhouse agriculture. © 2024 IEEE.

关键词： Horticulture

来源：评论

学校读者我要写书评

暂无评论

Expression recognition method based on feature redundancy optimization

引用

signal IMAGE AND VIDEO processing 2025年第4期19卷 1-11页

作者： Shao, Dangguo Zhuang, Luwei Ma, Lei Yi, Sanli Kunming Univ Sci & Technol Fac Informat Engn & Automat Kunming Peoples R China Yunnan Key Lab Comp Technol Applicat Kunming Peoples R China

Facial expression recognition (FER) plays a crucial role in domains such as healthcare and access security. Traditional models primarily utilize convolutional networks to extract features like facial landmarks and positions of facial features. However, these methods often result in feature maps with significant redundancy, contributing minimally to network performance enhancement. To address this limitation, we propose the DPConv module, which innovatively segments the channel dimension and applies dual convolutional kernel sizes. This module replaces several convolutional blocks within the POSTER++ (Mao et al. in POSTER++: A Simpler and Stronger Facial Expression Recognition Network. arXiv:2301.12149, 2023) architecture, leading to a reduction in parameters while simultaneously enhancing network efficiency and accuracy. Moreover, we propose a sliding window multi-head cross-self-attention mechanism, which is based on the sliding window multi-head self-attention (Liu et al. in Proceedings of the IEEE/CVF international conference on Computer vision, 2021) mechanism, which substitutes the conventional attention mechanism, facilitating the modeling of global dependencies and further optimizing the network's overall performance. Our model, DPPOSTER, was tested on the RAF-DB, FERPlus and SFEW datasets, and experimental comparisons were conducted with different combinations of convolution kernel sizes and channel segmentation ratios. The results showed that DPPOSTER achieved performance improvements of 0.59%, 0.37% and 2.32% over POSTER++ on the RAF-DB, FERPlus and SFEW datasets, respectively.

关键词： Deep learning Facial expression recognition Computer vision Feature redundancy Attention mechanism

来源：评论

学校读者我要写书评

暂无评论

SVTNet: Dual Branch of Swin Transformer and vision Transformer for Monocular Depth Estimation

SVTNet: Dual Branch of Swin Transformer and Vision Transform...

引用

2025 IEEE international conference on Acoustics, Speech, and signal processing, ICASSP 2025

作者： Jia, Shuwen Wang, Yongxiong Chen, Han Huang, Shuai School of Optical-Electrical and Computer Engineering University of Shanghai for Science and Technology Shanghai China

ISBN: (纸本)9798350368741

In monocular depth estimation, effective acquisition of global and local information is the key to improving accuracy. We introduce a novel dual branch network called Swin vision Transformer Net (SVTNet), where the Swin Transformer and vision Transformer are combined to learn features with global and local information. In Swin Transformer branch, we remove the Feature Pyramid Networks which reduces the noise introduced by interpolation and maintains the original characteristics of the hierarchical features extracted from each layer of Swin Transformer. In vision Transformer Branch, we remove the multi-scale convolutional structure and keep the same-scale feature maps extracted by vision Transformer for reducing the consumption of memory and extracting more global features. The experimental results on the NYU-Depth v2, KITTI, and CityScapes datasets show that compared with existing advanced methods, the SVTNet has achieved significant performance improvements in multiple evaluation metrics. © 2025 IEEE.

关键词： dual branch Feature Pyramid Networks monocular depth estimation Same-scale feature Swin Transformer vision Transformer

来源：评论

学校读者我要写书评

暂无评论

Exploring Convolutional Autoencoder Efficacy in Noise Removal for Image processing and Computer vision: A Study Using the MNIST Dataset 10

Exploring Convolutional Autoencoder Efficacy in Noise Remova...

引用

10th international conference on Control, Decision and Information Technologies (CoDIT)

作者： Kandic, Elma Akagic, Amila Bohlouli, Mandi Univ Pecs Fac Engn & Informat Technol Boszorkany Ut 2 H-7624 Pecs Hungary Univ Sarajevo Dept Comp Sci & Informat Zmaja Bosne BbUniv Campus Sarajevo 71000 Bosnia & Herceg Petanux GmbH Res & Innovat Dept Dottendorfer Str 86 D-53129 Bonn Germany

ISBN: (纸本)9798350373981;9798350373974

Noise removal in image processing and computer vision is a crucial preprocessing step employing a spectrum of techniques. In recent years, autoencoders exhibit remarkable efficacy in mapping noisy images to clean counterparts, capturing intricate relationships for effective noise removal. Motivated by the challenges posed by noise in real-world images, this research focuses on the denoising preprocessing step, crucial for tasks like object detection and segmentation. The study explores the application of autoencoders in removing artificially added noise from images within the MNIST dataset. The MNIST dataset's simplicity and historical significance facilitate focused examinations on specific aspects, such as the impact of different types and levels of noise. The efficacy of autoencoders for noise removal is assessed through the evaluation of results using various metrics, including SSIM, PSNR, MSE, and RMSE. In one remarkable instance, the reconstruction process achieved an impressive peak SSIM score of 99.06%, showcasing the efficacy of the method in preserving image fidelity despite the challenging presence of noise. This comprehensive analysis provides valuable insights into the performance and effectiveness of autoencoders in the context of noise reduction in various domains.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Research on the positioning technology of curved weld seam based on laser vision 4

Research on the positioning technology of curved weld seam b...

引用

4th international conference on Advanced Algorithms and signal Image processing, AASIP 2024

作者： Dong, Quancheng Cao, Lin Shao, Mingxuan Tian, Lianfa Miao, Chunming Xu, Ziyao University Of Ji'nan School of Mechanical Engineering Shandong Ji'nan China MH Robot & Automation Co. LTD. Shandong Weifang China

ISBN: (数字)9781510682818

ISBN: (纸本)9781510682801

As the demand for intelligent and autonomous welding becomes more and more extensive, in order to realise the degree of autonomy of welding robotic arm for welding, thus the research of weld seam positioning technology is carried out. The weld positioning system is built by combining the line structured light vision sensor and the robotic arm, and the Gaussian filtering method, Otsu threshold segmentation method and image morphology closed-circuit algorithm are carried out sequentially by acquiring the image of the weld seam light bar, solving the interference caused by the high reflective noise of the welding piece to the narrow weld seam, and obtaining the complete weld seam characteristic topographic information. Determine the scanning speed of 6mm/s, realize the global scanning of the curved weld seam of the non-regular weldment, obtain the pixel coordinates of the light bar image of the weld seam in each frame, assign the value of 0 to the Z coordinate, obtain the linear model of RANSAC, obtain the linear model with the highest number of collected points for feature point extraction, and judge the current frame image, obtain the weld seam centreline position point set and remove the weld seam reconstruction of the workbench. For the extracted non-regular curve weld centreline, NURBS curve interpolation and control of two-point gap straight line interpolation are combined to obtain a smooth and dense complex weld centreline. From the experimental analysis, the error range between the extracted weld centreline and the actual position point in each axis direction is 0.2~0.7mm, which meets the welding positioning accuracy requirements. © 2024 SPIE.

关键词： robotic arms

来源：评论

学校读者我要写书评

暂无评论

A Navigation System based on Event Camera for People Living with Blindness and vision Impairment 4

A Navigation System based on Event Camera for People Living ...

引用

4th IEEE international conference on Electrical, Computer, and Energy Technologies, ICECET 2024

作者： Ma, Maria Sabrina Kavehei, Omid School of Biomedical Engineering The University of Sydney Sydney Australia

ISBN: (纸本)9798350395914

Neuromorphic sensing has drawn significant interest in automated sensing for its low latency, low power consumption, and high throughput signal processing. Leveraging these advantages, in this work, we propose a proof-of-concept navigation system for people with vision impairment based on an event camera with physical and optical filtering that greatly helps our real-time denoising and backend signal processing. We used a 940±2 nm filter directly on the event camera with our own controlled series of 940 nm LEDs for scene exposure when and where required. Our event processing algorithm is developed to potentially provide real-time directional feedback to the person through bone-conduction vibrations that do not interfere with the normal hearing pathway;hence, the person is able to continue to hear the surrounding environment around them. © 2024 IEEE.

关键词： Navigation systems

来源：评论

学校读者我要写书评

暂无评论

Eagle Strategy based Improved Adaptive Control System for an Inverted Pendulum - Mobile robot 3

Eagle Strategy based Improved Adaptive Control System for an...

引用

3rd IEEE international conference on Intelligent Techniques in Control, Optimization and signal processing, INCOS 2024

作者： Jarin, T. Deepanraj, B. Beula, Annie Steffy Vignesh, K. Ezhil Jyothi Engineering College Department of Eee Kerala Thrissur India Prince Mohammad Bin Fahd University Department of Mechanical Engineering Al-Khobar Saudi Arabia Stella Mary's College of Engineering Department of Eee Tamil Nadu Aruthenganvilai India

ISBN: (纸本)9798350361186

Navigation robot is required in industries to perform the assigned task. Therefore the robot should choose the shortest path to reach the given target. The computation time and cost should be less while making the decision. This work explains the position balancing of the Two Wheel Inverted Pendulum-Mobile robot (TWIP-MR) system using the nature-inspired algorithm Eagle Strategy with Particle Swarm Optimization (ESPSO). According to stability concerns, this is an unstable open loop system exhibiting nonlinear behavior. The design of the system focuses on the balanced movement of the robot while moving towards the intended position. Two loops are offered while designing TWIP-MR, one for balancing the linear displacement and the other for maintaining the appropriate angular movement. This research uses a mathematical technique to provide an analysis of the TWIP-MR system. The random global search, intensive local search and a levy walk is adopted from Eagle search strategy. Particle Swarm Optimization is included in the local search stage of the ES. The PID controller is used in order to reduce the positional errors. © 2024 IEEE.

关键词： Adaptive control systems

来源：评论

学校读者我要写书评

暂无评论

IOT Based Virtual Telepresence robot 3

IOT Based Virtual Telepresence Robot

引用

3rd IEEE international conference on Intelligent Techniques in Control, Optimization and signal processing, INCOS 2024

作者： Pon Pushpa, S. Ewins Jayasree, T. Mariammal, K. Selwyn, Chris Jonathan Solomon, S. Benny Raj, S. Harish College of Engineering Anna University Department of Ece Guindy Chennai India Madras Institute of Technology Anna University Department of Electronics Engg. Chrompet Chennai India

ISBN: (纸本)9798350361186

This paper, marks an innovative exploration of robotics, IoT, and telepresence technologies, aiming to create a versatile and user-friendly robot, for remote control and real-time presence in distant locations. The key components and functionalities of this work, include the ESP32 micro controller, which serves as the central brain, interfacing with an analog joystick for user control. The Raspberry Pi camera, integrated through Message Queuing Telemetry Protocol, enhances capabilities, with high-quality real-time video, streaming to provide an immersive experience. The L298N motor driver facilitates precise control over the robot's mobility, allowing it to navigate in any direction. ESP32's role in interfacing with the analog joystick, for movement control is emphasized. As a conclusion, the integration of hardware and software components, in the development of the IoT based Virtual telepresence robot, represents a significant leap forward in the realm of remote robotics. © 2024 IEEE.

关键词： Internet of things

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：