检索结果-内蒙古大学图书馆

7th International Conference on Multimedia and image processing, ICMIP 2022

作者： Wan, Jiale Zhao, Li Bian, Yan Key Laboratory of Information Sensing and Intelligent Control Tianjin University of Technology and Education Tianjin300222 China

ISBN: (纸本)9781450387408

At present, offline analysis and research of lower limb motor imagery brain computer interface (MI-BCI) are relatively mature, but there are few researches on the online recognition of lower limb MI-BCI. The online recognition of the two MI tasks of the knee-Ankle adjacent joints of the right lower extremity was carried out, and the EEG signals of the two MI tasks assisted by electrical stimulation are collected. The Filter Bank Common Spatial pattern (FBCSP) is used for feature extraction, and BP neural network is used for online recognition. The average recognition accuracy rate of online motor imagery of 13 subjects reaches 79.62%, and the recognition accuracy rate of the highest person is 92.50%, which verified the feasibility and practicability of online MI-BCI in adjacent joints of lower limbs. © 2022 ACM.

关键词： Brain computer interface

来源：评论

学校读者我要写书评

暂无评论

Semantic segmentation of landslide image using DeepLabv3+ and completed local binary pattern

引用

Journal of Applied remote sensing 2025年第1期19卷

作者： Wang, Wei Zhang, Zhihua Zhu, Xinyu Yang, Shuwen Lanzhou Jiaotong University Faculty of Geomatics Gansu Lanzhou China Natl.-Loc. Jt. Eng. Res. Ctr. of Technol. and Applications for National Geographic State Monitoring Gansu Lanzhou China Key Laboratory of Science and Technology in Surveying & Mapping Gansu Province Gansu Lanzhou China

Landslide identification is an important task in the field of geologic disaster monitoring and early warning, which is of great significance for improving social safety and mitigating the impact of disasters. With the development of computer vision, deep learning is widely used in landslide recognition research. We focus on segmenting landslides from high-resolution optical satellite images using convolutional neural network. Currently, deep learning semantic segmentation models still face issues such as neglecting small objects and incorrectly segmenting terrain features with similar shapes and pixel characteristics. Considering the unbalanced distribution of categories and large differences in scene styles during the extraction of key feature information from remote sensing images, landslides have diverse and complex backgrounds. We propose a fusion DeepLabv3+ and completed local binary pattern (CLBP) landslide image semantic segmentation method (CLBP-DeepLabv3+), using the improved inverted residual block as the core structure of backbone to extract different levels of image information, and after backbone extracts landslide image features, it connects the improved DenseASPP to fuse the different levels of features to better pay comprehensive attention to local and global features and obtain contextual information at different scales. Then, the texture and edge features of the image are extracted using CLBP, and the multi-level features are merged by introducing the feature aggregation module, which constitutes the CLBP-DeepLabv3+ model. Through ablation experiments and comparative tests on a self-made dataset, the experimental results show that the proposed method performs the best on the validation set, with a mean intersection over union (mIoU) of 88.62%, mean pixel accuracy (mPA) of 94.17%, recall rate of 90.17%, and an intersection over union (IoU) for landslides of 80.53%. Compared with the original DeepLabv3+ model, the improved DeepLabv3+ increased the mI

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Real-time image recognition System based on Ultra-Low-Latency Spiking Neural Networks

Real-time Image Recognition System based on Ultra-Low-Latenc...

引用

Signal, Information and Data processing (ICSIDP), IEEE International Conference on

作者： Bin Lan Xi Zhang Heng Dong Ming Xu Beijing Institute of Technology Chongqing Innovation Center Chongqing China Beijing Institute of Technology Beijing China Sichuan TianFu New Area Beijing Institute of Technology Innovation Equipment Research Institute Chengdu China National Key Laboratory of Science and Technology on Space-Born Intelligent Information Processing(SBIIP) Beijing China

ISBN: (数字)9798331515669

ISBN: (纸本)9798331515676

remote sensing technology plays an important role in many tasks such as natural disaster detection, weather and climate monitoring and military defense. Currently, remote sensing image processing predominantly relies on Convolutional Neural Networks (CNNs) and Transformers, which require a huge amount of multiplication and addition operations, bringing pressure to the deployment of satellite platforms with strict power consumption and computing power constraints. Spiking neural network (SNN), where the signal is encoded as a spike train instead of analog values, significantly reduces energy consumption compared to traditional artificial neural network. This paper proposes an all-input parallel deployment approach that minimizes or eliminates interactions with off-chip memory, thereby accelerating inference speed and reducing power consumption. We deployed LeNet ultra-low latency SNN on the Xilinx VC709 Board. Experimental results demonstrate that the proposed accelerator achieves a processing speed of 1240 frames per second, an accuracy rate of 98.14%, and a power consumption of only 0.745 Watts, making it well-suited for the low power and real-time requirements of remote sensing tasks.

关键词： Power demand Accuracy Satellites Spiking neural networks Transformers Real-time systems Resource management remote sensing Monitoring Meteorology

来源：评论

学校读者我要写书评

暂无评论

ICIGP 2022 - Proceedings of the 2022 5th International Conference on image and Graphics processing

ICIGP 2022 - Proceedings of the 2022 5th International Confe...

引用

5th International Conference on image and Graphics processing, ICIGP 2022

ISBN: (纸本)9781450395465

The proceedings contain 57 papers. The topics discussed include: precision inspection and evaluation system for paper packaging of cigarettes;a lightweight network for violence detection;micro-expression detection based on action units and multi-region feature fusion;multi-scale semantic representation and supervision for remote sensing change detection;cost-effective video-based poor repertoire detection for preterm infant general movement analysis;image inpainting based on edge features and attention mechanism;spatial non-cooperative target point cloud reconstruction;fast graph-based binary classifier learning via further relaxation of semi-definite relaxation;research on handwritten digital image recognition model based on deep learning and construction of browser service platform;analysis of eyebrow motion for micro-expression recognition;and features calculation of closed curve and its application in leaf discrimination.

关键词：

来源：评论

学校读者我要写书评

暂无评论

SAR-TSCC: A Novel Approach for Long Time Series SAR image Change Detection and pattern Analysis

引用

IEEE TRANSACTIONS ON GEOSCIENCE AND remote sensing 2023年 61卷

作者： Li, Weisong Ma, Peifeng Wang, Haipeng Fang, Chaoyang Fudan Univ Key Lab Informat Sci Electromagnet Waves Shanghai 200433 Peoples R China Chinese Univ Hong Kong Inst Space & Earth Informat Sci Dept Geog & Resource Management Hong Kong Peoples R China Chinese Univ Hong Kong Shenzhen Res Inst Shenzhen 518063 Peoples R China Jiangxi Normal Univ Key Lab Poyang Lake Wetland & Watershed Res Minist Educ Nanchang 330022 Peoples R China

Change detection has played an increasingly important role in multitemporal remote sensing applications recently. Long time series analysis is providing new information of land cover changes and improving the quality and accuracy of the change information being derived from remote sensing. The purpose of this study is to dig for more change temporal information and change pattern information from synthetic aperture radar (SAR) image time series (ITS), which is of great significance for monitoring urban area changes, conducting land use surveys, and renovating illegal constructions. In the study, a novel unified framework for long time series SAR image change detection and change pattern analysis (SAR-TSCC) was proposed for land cover change mapping. To obtain the most notable change time rapidly, a fast SAR ITS change point search method based on pruned exact linear time (SAR-PELT) algorithm was adopted. Meanwhile, the deep time series classification network, named SAR time series transformer (SAR-TST), was implemented to recognize the change patterns, which is based on time series transformer (TST) architecture. Considering the lack of real training data, a novel synthetic data generation method is developed. The combination of the synthetic and real data enhanced the generalization of the classifiers. The proposed framework was used for monitoring a large urbanization area in the northwest of Hong Kong, China. The Cosmo Skymed (CSK) time series data acquired from 2013 to 2020 were exploited for land cover change analysis. Experiment results showed that our approach achieved the state-of-the-art performance, as the time accuracy reached 86% and the classification accuracy on the four main change patterns (impulse, step, cycle, and complex) is over 99%. In particular, the proposed SAR-TST model showed remarkable advantages in the presence of insufficient real data.

关键词： Time series analysis Synthetic aperture radar Monitoring Radar polarimetry pattern recognition remote sensing Urban areas Change detection deep learning (DL) synthetic aperture radar (SAR) images time series

来源：评论

学校读者我要写书评

暂无评论

Robust Zero-watermarking Algorithm for Medical images Based on Maximally Stable Extremal Regions and DCT 6

Robust Zero-watermarking Algorithm for Medical Images Based ...

引用

6th IEEE International Conference on pattern recognition and Artificial Intelligence, PRAI 2023

作者： Wang, Xue Li, Jingbing Bhatti, Uzair Aslam Huang, Mengxing Hainan University School of Information and Communication Engineering Haikou China

ISBN: (纸本)9798350325485

With the rapid development of the information age, the digitization of information media has provided us great convenience. In medical image processing, digital watermarking has made great progress. Due to the weak resistance to geometric attacks in this technology, there is a risk of information leakage. For example, patient information may be maliciously tampered with and leaked in remote medical systems. This article proposes a medical image zero-watermarking method based on MSER-DCT to address this issue. First, MSER-DCT is utilized to extract feature regions from the images, while perceptual hashing is used to obtain feature sequences. Next, chaotic mapping is employed to encrypt and restore the watermark. Finally, zero-watermarking technology is used to embed and extract the watermark. This combination of techniques allows for secure watermarking of medical images. The experimental results of this article indicate that this algorithm has strong robustness, which can effectively resist geometric attacks. © 2023 IEEE.

关键词： image compression

来源：评论

学校读者我要写书评

暂无评论

A Novel Perspective Invariant Dissociaton Hash Method for remotely Sensed imagery Plusieurs Target Classification

引用

IEEE ACCESS 2021年 9卷 31994-32007页

作者： Zhang, Zhifeng Cui, Xiao Ji, Xiaohui Zhengzhou Univ Light Ind Software Engn Coll Zhengzhou 450001 Peoples R China

Plusieurs target classification based on remotely sensed imagery is a hot and difficult-ridden topic in recent years. Specifically, the time and space complexity of the classical pattern recognition model in the classification task based on remote sensing image is usually high, and the remotely sensed imagery is usually not normally captured. For the former, we introduce the low delay and low storage hash method into the plusieurs target classification of remotely sensed imagery. Aiming at the latter, in order to improve the effectiveness of the proposed model in monitoring perspective transformation data, a dissociation perspective invariant model is constructed. By fusing these two solutions, a perspective invariant dissociaton hash model for remotely sensed imagery plusieurs target classification is obtained. By adding perspective invariant constraint to the supervised dissociaton hash method, our method forces the same type of target to share the same binary code, increases the similarity of the hash code of the same type of target, and thus improves the performance. In order to verify the validity and universality of our method, two different data sets were used for experiments, in which different hash methods and different classification methods are compared. The experimental results show that, compared with the comparison methods, the proposed method improves the accuracy of plusieurs target classes with a small number of samples, thus achieving a higher overall classification accuracy. In addition, the advantages of hash method in low storage make the speed of proposed method improved compared with the classical pattern recognition method.

关键词： remote sensing Indexes Satellites Classification algorithms Training Kernel Data models Perspective invariant hash method plusieurs target classification

来源：评论

学校读者我要写书评

暂无评论

A remote sensing image target recognition method based on improved Mask-RCNN model 2

A remote sensing image target recognition method based on im...

引用

2nd IEEE International Conference on Big Data, Artificial Intelligence and Internet of Things Engineering, ICBAIE 2021

作者： Huiming, Yu Fuxin, Xu Central South University School of Physics and Electronics Changsha410083 China

ISBN: (纸本)9780738131221

As an important branch in the field of machine vision, target detection and recognition technology have important applications in various fields. In view of the fact that the general target detection and recognition model does not perform well in remote sensing images, and there're large amount of different size targets need to be recognized. On the basis of Mask R-CNN model, one of the classic target detection and recognition model, this paper proposes A target recognition and detection method that integrates data enhancement processing and improves the Mask R-CNN backbone network FPN to RFP. When inputting image data, perform random enhancement processing on the data, randomly cut the image, and randomly merge;when extracting features, improve the FPN in the original Mask R-CNN model to RFP to make the extracted features better Reflecting the original picture information, experiments have shown that this model has a good performance in fine-grained target recognition in remote sensing images. On the same data set, its evaluation indicators are superior to other comparison algorithms in all aspects. © 2021 IEEE.

关键词： remote sensing

来源：评论

学校读者我要写书评

暂无评论

Robust Approach for Butterfly Species Classification Using a Single Spatio-Hyperspectral image

Robust Approach for Butterfly Species Classification Using a...

引用

Workshop on Hyperspectral image and Signal processing: Evolution in remote sensing, WHISPERS

作者： Erick A. Adjé Gilles Delmaire Arnaud S. R. M. Ahouandjinou Matthieu Puigt Gilles Roussel Univ. Littoral Côte d'Opale LISIC – UR 4491 Longuenesse France Univ. d'Abomey-Calavi LRSIA Cotonou Benin

ISBN: (数字)9798331513139

ISBN: (纸本)9798331513146

Distinguishing between pest and pollinator butterfly species is a major challenge in precision agriculture. However, traditional RGB cameras, capturing only shape and surface color, are insufficient for detailed insect analysis. This work explores the rich spectral information provided by hyperspectral imaging for effective butterfly species identification. For this purpose, we use a single spatio-spectral image that provides partial spectral information to identify the butterfly species. The proposed classification approach consists of a convex combination of the probabilistic decisions obtained by the Gaussian Naive Bayes and Z-score methods for each butterfly reflectance. Compared to traditional classification models, this approach showed higher robustness and performance.

关键词： Reflectivity Precision agriculture image recognition Shape Signal processing Probabilistic logic Cameras Robustness Hyperspectral imaging Signal to noise ratio

来源：评论

学校读者我要写书评

暂无评论

MORPHOLOGICAL TEXTURE DESCRIPTION FOR HYPERSPECTRAL imageS: pattern SPECTRUM 11

MORPHOLOGICAL TEXTURE DESCRIPTION FOR HYPERSPECTRAL IMAGES: ...

引用

11th Workshop on Hyperspectral Imaging and Signal processing - Evolution in remote sensing (WHISPERS)

作者： Deborah, H. Richard, N. Norwegian Univ Sci & Technol Dept Comp Sci Teknologivn 22 Gjovik Norway Univ Poitiers Lab XLIM JRU CNRS 7252 Futuroscope France

ISBN: (纸本)9781665436014

A metrological extension of morphological granulometry for the hyperspectral domain is introduced in this work. This development is enabled by the latest study of a suitable ordering relation for hyperspectral images. With granulometry as a texture descriptor, a suitable similarity measure for it is also introduced. In addition to providing validation experiments to the extension, a preliminary result in a texture discrimination task can also be found in this work.

关键词： Texture mathematical morphology pattern spectrum hyperspectral

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：