检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Shimada, Tomoyasu Murasaki, Kazuhiko Sato, Shogo Nishimura, Toshihiko Yoshida, Taiga Tanida, Ryuichi NTT Corporation Japan

Recent advances in deep learning have improved 3D point cloud registration but increased graphics processing unit (GPU) memory usage, often requiring preliminary sampling that reduces accuracy. We propose an overlapping region sampling method to reduce memory usage while maintaining accuracy. Our approach estimates the overlapping region and intensively samples from it, using a k-nearest-neighbor (kNN) based point compression mechanism with multi layer perceptron (MLP) and transformer architectures. Evaluations on 3DMatch and 3DLoMatch datasets show our method outperforms other sampling methods in registration recall, especially at lower GPU memory levels. For 3DMatch, we achieve 94% recall with 33% reduced memory usage, with greater advantages in 3DLoMatch. Our method enables efficient large-scale point cloud registration in resource-constrained environments, maintaining high accuracy while significantly reducing memory requirements. This paper is accepted for IEEE International conference on visual communications and image processing 20241 © 2024, CC BY.

关键词： Graphics processing unit

来源：评论

学校读者我要写书评

暂无评论

Low-bit-rate scalable compression of mobile wireless video

Low-bit-rate scalable compression of mobile wireless video

引用

IEEE Region 10 International conference TENCON

作者： R. Fakeh A.A.A. Ghani M.Y.M. Saman A.R. Ramli Faculty of Computer Science and Information Technology Universiti Putra Malaysia Serdang Malaysia Deptartment Of Computer & Communication Systems Engineering Faculty of Engineering Universiti Putra Malaysia Serdang Selangor Darul Ehsan Malaysia

Low bit rate video coding has triggered intensive research in both academia and industry. To address the problem of heterogeneity, and meeting a very small bandwidth of mobile wireless networks, low bit rate scalable video compression is becoming more important. This paper discusses the current approach to wireless visual communications for the need of low bit rate scalable video compression, using discrete wavelet transform techniques. The discrete wavelet transform (DWT) has provided an efficient framework of multiresolution space-frequency representation with promising applications in image processing. This codec is based on 3D wavelet coding and adaptive quantization of the subbands coefficients. A common problem with some existing quantization methods is that the inherent image structures are severely distorted with course quantization. The Moving Picture Expert Group-4 (MPEG-4), an important component of the ITU standard is referred to as the benchmark of video standards for low bit rate video transmission.

关键词： Video compression Discrete wavelet transforms Bit rate Quantization image coding Video coding Bandwidth Wireless networks Wireless communication visual communication

来源：评论

学校读者我要写书评

暂无评论

An Adaptively Polygonal Simplification Algorithm for Remote Sensing image Segmentation Contours

An Adaptively Polygonal Simplification Algorithm for Remote ...

引用

2014 International conference on Remote Sensing and Wireless communications(RSWC 2014)

作者： JIANHUA LIU ZHAOYING YANG GUOYIN CAI RONGHUA WANG MINGYI DU Key Laboratory for Urban Geomatics of National Administration of Surveying Mapping and Geoinformation

Usually,the contours of polygons generated by image segmentation technology show jagged outlines and a large amount of redundant ***,the original segmentation contours hardly conform to GIS data producing standards without *** is also difficult to automatically give a unique set of parameters per segmentation contour *** order to meet the actual requirements for GIS data producing,we present an Adaptively Improved algorithm based on Douglas-Peucker,named AIDP,which integrated criterions of vertical and radial distance restriction,and design a corresponding parameter adaptive acquisition *** proposed method AIDP was evaluated by comparing with the most widely used Douglas-Peucker algorithm implemented in the ArcGIS through visual inspection,quantitative measures and applications in water body *** experimental results showed AIDP not only can acquire simplification parameters automatically,but also greatly speed up the data processing workflow with acceptable results.

关键词： image An Adaptively Polygonal Simplification Algorithm for Remote Sensing image Segmentation Contours

来源：评论

学校读者我要写书评

暂无评论

Development of image analysis for detection of defects of BGA by using X-ray images

Development of image analysis for detection of defects of BG...

引用

IEEE Instrumentation and Measurement Technology conference

作者： T. Sumimoto T. Maruyama Y. Azuma S. Goto M. Mondou N. Furukawa S. Okada Okayama University Okayama Shikata-cho JAPAN Eastern Hiroshima Prefecture Industrial Research Institute Fukuyama JAPAN National Institute of Advanced Industrial Science and Technology Kure Hiroshima JAPAN

In the surface mount technology, Ball Grid Array (BGA) has been used in a production of PC boards, because of their excellent characters such as high density of the lead pin pitch, better lead rigidity and self-alignment during re-flow processing. This paper deals with the development of image analysis for the detection of defects at BGA solder joints in PC boards by using X-ray images. In the conventional IC boards, it is possible to detect defects of solder joints by visual inspection, because the lead of IC package is set on its outside. However, we can't detect visually defects at BGA solder joints, because they are hidden under the IC package. In a production line, the inspection of BGA in PC boards depends on the function test of electric circuits in the final process. To improve a cost performance and the reliability of PC boards, an inspection of BGA is required in the surface mount process. Types of defects at BGA solder joints are solder bridge, missing connection, solder voids, open connection and miss-registration of parts. As we can find mostly solder bridge in these defects, we pick up this to detect solder bridge in a production line. The problems of image analysis for the detection of defects at BGA solder joints are the detection accuracy and image processing time according to a line speed of production. To get design data for the development of the inspection system, which can be used easily in the surface mount process, it is important to develop image analysis techniques based on X-ray image data. At the first step of our study, we attempt to detect the characteristic of the solder bridges based on an image analysis.

关键词： image analysis X-ray detection X-ray detectors X-ray imaging Soldering Production Inspection Bridge circuits Lead Integrated circuit packaging

来源：评论

学校读者我要写书评

暂无评论

Underwater image Enhancement Based on Dehazing and Color Correction

Underwater Image Enhancement Based on Dehazing and Color Cor...

引用

IEEE International conference on Big Data and Cloud Computing (BdCloud)

作者： Hanyu Li Peixian Zhuang Wang Wei Jingjing Li Nanjing University of Information Science and Technology Nanjing China

ISBN: (数字)9781728143286

ISBN: (纸本)9781728143293

Clear and clean underwater images can provide valuable information, which are crucial for developing, exploring, and protecting the underwater resources. However, the raw underwater image seldom fulfills the requirements concerning the underwater research due to the serious image degradation. To enhance single underwater image, we propose an effective and new method combined on the dehazing and color correction algorithm. First, we obtain the dehazed image by a fusion dehazing method calculated on the difference between maximum green-blue dark channels and maximum red dark channel, and the top 0:1% bright pixels in green-blue dark channels. Second, we obtain the enhanced image via the color restoration method corresponding to the human visual system. Finally, we use an efficient and simple weight fusion strategy to incorporate the dehazed image and the enhanced image for yielding the final high quality underwater image. Subjective assessment demonstrates that our method can remove haze, correct color cast, improve image brightness, and preserve image naturalness. The enhanced results obtain the highest values in terms of entropy, average gradient, UCIQE, UIQM, and CCF, which is superior to several existing underwater enhancement methods. Moreover, our method can improve other degraded image quality, such as low-light image and haze image, and can benefit the application tests, such as key points matching and edge detection.

关键词： image color analysis Degradation image restoration image edge detection image quality Channel estimation image enhancement

来源：评论

学校读者我要写书评

暂无评论

image Compressed Sensing Reconstruction via Deep image Prior with Feature Space and Texture Information 8th

Image Compressed Sensing Reconstruction via Deep Image Prior...

引用

8th International conference on Cognitive Systems and Information processing, ICCSIP 2023

作者： Peng, Zhao Jinchan, Wang Huanqing, Peng Fei, Xiang Liwen, Zhang College of Information Engineering Henan University of Science and Technology Luoyang471023 China

ISBN: (纸本)9789819980178

Deep learning has yielded remarkable achievements in compressed sensing image reconstruction in recent years. However, when the sampling rate is low, the reconstruction image is blurred and lacks texture details. In this paper, we design a dual-path compressed sensing reconstruction network based on image feature space information flow and texture information, which can solve the above problems to a certain extent. The feature space information flow path employs proximal gradient descent algorithms to map essential structural information from the pixel space to the feature space, effectively reducing data dimensionality and redundancy. On the other hand, after combining the attention mechanism, the texture information path restores the texture details of the image in the frequency domain, which can effectively enhance the quality and visual effects of the reconstruction image. During training, a unified loss function guides the alternating optimization of both paths. Evaluation of prominent benchmark datasets, including Set5, Set11, and BSD68, reveals that our proposed method outperforms traditional iterative approaches and existing deep learning-based methodologies in terms of both reconstructed image quality and network robustness under low sampling rates. © 2024, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： image reconstruction

来源：评论

学校读者我要写书评

暂无评论

QEVC: Quality Enhancement-Oriented Video Coding

QEVC: Quality Enhancement-Oriented Video Coding

引用

International conference on Computers, communications, and Systems (ICCCS)

作者： Hao Li Weimin Lei Wei Zhang School of Computer Science and Engineering Northeastern University Shenyang P. R. China

ISBN: (数字)9781728161365

ISBN: (纸本)9781728161372

Quality enhancement (QE) is an important post-processing technology for high-resolution video services at low bit rates, which can effectively improve the quality of compressed video. The application of deep learning methods to the quality enhancement task has achieved great success in the past few years. However, the existing schemes are usually coding-independent, which still leaves room for further development of related technologies. Therefore, in this paper, we propose a quality enhancement-oriented video coding scheme. By analyzing the features of different video regions, a deep reinforcement learning model is used to determine the distortions of regions. Then during the video reconstruction, convolutional neural network (CNN)-based quality enhancement networks with different scales are selected to improve the video quality according to the distortion of different regions. Experimental results show that the proposed scheme outperforms the HEVC anchor in case of bits saving, bits allocation, and shows good visual quality especially at low bit rates.

关键词： image coding Distortion Indexes Training Bit rate Task analysis Encoding

来源：评论

学校读者我要写书评

暂无评论

FREQ-MIP-AA: Frequency Mip Representation for Anti-Aliasing Neural Radiance Fields

FREQ-MIP-AA: Frequency Mip Representation for Anti-Aliasing ...

引用

IEEE International conference on image processing

作者： Youngin Park Seungtae Nam Cheul-Hee Hahm Eunbyung Park Department of Digital Media and Communications Engineering Sungkyunkwan University Visual Display Division Samsung Electronics Department of Artificial Intelligence Sungkyunkwan University Department of Electrical and Computer Engineering Sungkyunkwan University

ISBN: (数字)9798350349399

ISBN: (纸本)9798350349405

Neural Radiance Fields (NeRF) have shown remarkable success in representing 3D scenes and generating novel views. However, they often struggle with aliasing artifacts, especially when rendering images from different camera distances from the training views. To address the issue, Mip-NeRF proposed using volumetric frustums to render a pixel and suggested integrated positional encoding (IPE). While effective, this approach requires long training times due to its reliance on MLP architecture. In this work, we propose a novel antialiasing technique that utilizes grid-based representations, usually showing significantly faster training time. In addition, we exploit frequency-domain representation to handle the aliasing problem inspired by the sampling theorem. The proposed method, FreqMipAA, utilizes scale-specific low-pass filtering (LPF) and learnable frequency masks. Scale-specific low-pass filters (LPF) prevent aliasing and prioritize important image details, and learnable masks effectively remove problematic high-frequency elements while retaining essential information. By employing a scale-specific LPF and trainable masks, FreqMipAA can effectively eliminate the aliasing factor while retaining important details. We validated the proposed technique by incorporating it into a widely used grid-based method. The experimental results have shown that the FreqMipAA effectively resolved the aliasing issues and achieved state-of-the-art results in the multi-scale Blender dataset. Our code is available at https://***/yi0109/FreqMipAA.

关键词： Training image coding Three-dimensional displays Codes Filtering Frequency-domain analysis Low-pass filters

来源：评论

学校读者我要写书评

暂无评论

Genre Classification of Movie Trailers using Spectrogram Analysis and Machine Learning

Genre Classification of Movie Trailers using Spectrogram Ana...

引用

International Black Sea conference on communications and Networking (BlackSeaCom)

作者： Porawat Visutsak Fuangfar Pensiri Ponrudee Netisopakul Navapoom Punsathit Pongkorn Rojsuwan Maywadee Phuthong Worapol Khanthawat Wittawat Prommanee Kanthong Jearaphan Tawichai Saekua Tichaporn Anantaprueksa Napas Kavalee Naravitch Sukkokee Nanticha Kingket Prapassorn Teeravas Patchara Nimmonrat Patcharada Yungyuen Pattareeya Nurltes Ratchanon Wongpanti Sekson Hompangwhai Haritchaya Sutthikawee Apichai Siangchin Department of Computer and Information Science Faculty of Applied Science KMUTNB Bangkok Thailand Department of Computer Science and Information Faculty of Science at Sriracha Kasetsart University Chonburi Thailand Faculty of Information Technology KMITL Bangkok Thailand

ISBN: (数字)9798350351859

ISBN: (纸本)9798350351866

This study investigates the application of machine learning for movie trailer genre classification using spectrogram analysis. Spectrograms, visual representations of a signal's frequency content over time, are extracted from movie trailers. Convolutional Neural Networks (CNNs), known for their image recognition capabilities, are employed to analyze these spectrograms and identify patterns that distinguish between different genres. Our approach is compared to a Random Forest model for performance evaluation. The models are trained on a dataset of movie trailers categorized into five genres: action, romance, drama, comedy, and thriller. Librosa, a Python library, is utilized for audio pre-processing, while the overall training process is conducted within the Python environment. This paper explores the potential of machine learning in conjunction with spectrogram analysis for effective movie trailer genre classification.

关键词： Training Performance evaluation visualization Time-frequency analysis Humanities image recognition Motion pictures

来源：评论

学校读者我要写书评

暂无评论

Transfer learning based multi-class malware classification using VGG16 feature extractor and SVM classifier

Transfer learning based multi-class malware classification u...

引用

International conference on communications and Signal processing

作者： Harman Saggu Er. Karandeep Singh T. N. Suresh Babu Department of Computer Science and Engineering Punjabi University Patiala India Department of Electronics and Communication Engineering Adhiparasakthi Engineering College Tamil Nadu India

ISBN: (数字)9798350353068

ISBN: (纸本)9798350353075

Malicious programs are becoming a severe problem in today’s internet-connected world. Several works have explored image analysis-based malware detection in recent years. The image analysis method converts the malicious code into gray-scale images, which enables the use of deep learning models in the domain of malware detection. This paper proposes a new architecture using a fine-tuned visual Geometry Group(VGG16) model with a support vector machine (SVM) classifier to achieve high-performance accuracy in detecting and classifying evolving malware. At first, the malicious code is represented in gray-scale images, which are further reshaped and resized to be fed into a fine-tuned VGG16 model as end-to-end feature extraction with the least dependence on human-crafted feature engineering. The extracted textural feature maps are further correlated to remove the highly correlated features before using them to train a support vector machine (SVM) classifier. The experiments are performed on public bench-marked MalImg datasets and actual malware samples. The VGG 16+ SVM model obtained 99.25% classification accuracy for the MalImg and 80.29% for real-world malware, which signifies that the model can detect real-world instances of malware. The proposed method outperforms similar existing works and is suitable for deployment in an IoT environment due to the least cost involved in feature engineering.

关键词： Support vector machines Geometry Deep learning visualization image analysis Costs Transfer learning

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：