检索结果-内蒙古大学图书馆

A Forgotten strategy for Pooled Contextualized Embedding learning 2020

A Forgotten Strategy for Pooled Contextualized Embedding Lea...

2nd international conference on image processing and machine Vision / international conference on pattern recognition and machine learning (IPMV)

作者： Cai, Dunbo Huang, Zhiguo Qian, Ling Hang, Tao Feng, Qingsong China Mobile Suzhou Software Suzhou Peoples R China

ISBN: (纸本)9781450388412

string (word) embeddings are keys to advanced neural natural language understanding (NLU) models. Recently, a new embedding called contextual string embedding (CtxEmb) reached a new state of the art on many NLU tasks. Specially, a pooling variant of CtxEmb demonstrated new best results on named entity recognition (NER) tasks. While the pooling variant is good at reaching new best performance in terms of Micro F1 score, it poses questions on what kind of information in the pool is helpful. In particular, the pooling variant maintains a (possibly long) list of previously seen occurrences of a word, and computes embedding for the word in a new context based on those occurrences. In this paper, we propose a strategy to forget some previous occurrences, and thus exploring which history are beneficial in constructing embeddings for a new context of a word. The proposed forgotten strategy is designed by accounting for a distance metric defined in this paper. Preliminary experiments on the WNUT-17 task show the effectiveness of our forgotten strategy, and uncover that embeddings that are diverse in terms of cosine similarity are helpful in forming an aggregated embedding.

关键词： Word Embedding learning Contextual string Embedding Representation learning

来源：评论

学校读者我要写书评

暂无评论

A Comparative Analysis by KNN, SVM & ELM Classification to Detect Sickle Cell Anemia 1

A Comparative Analysis by KNN, SVM & ELM Classification to D...

引用

1st international conference on Robotics, Electrical and Signal processing Techniques (ICREst)

作者： Chy, Tajkia Saima Rahaman, Mohammad Anisur Chittagong Univ Engn & Technol Dept Elect & Telecommun Engn Chittagong Bangladesh

ISBN: (纸本)9781538680124

Red blood cell abnormalities involve erythrocytes that supply oxygen to all body tissues. Sometimes the formation and role of erythrocytes are hindered. Sickle cell anemia (SCA) is one kind of red blood cell disease. People carrying sickle cell anemia are increasing day by day. Sickle cell anemia shortens life expectancy. But life expectancy can be extended by diagnosing it an early stage. To identify the existence of sickle cells, an image processing procedure is developed. Blood samples are collected in the form of image format. The conversion of gray image, noise filtering and enhancement of image is done in image pre-processing. Fuzzy C means clustering is applied to determine the normal and sickle cells. Morphological operations are also applied to images. The geometrical and statistical features are used for extraction. Lastly, k nearest neighbor (km), support vector machine (svm) & extreme learning machine (elm) classifiers are implemented to test images. Comparisons among the classifiers with reliable results are presented by this system.

关键词： Sickle cell anemia gray scale image conversion median filter fuzzy c means clustering morphological operations feature extraction k nearest neighbor support vector machine extreme learning machine

来源：评论

学校读者我要写书评

暂无评论

Facial expression recognition based on LBP-EHMM

Facial expression recognition based on LBP-EHMM

引用

1st international Congress on image and Signal processing

作者： Cao, Jianqiang Tong, Can Southeast Univ Res Ctr Learning Sci Nanjing 210096 Peoples R China

ISBN: (纸本)9780769531199

Facial expression recognition can be divided into three steps: face detection, expression feature extraction and expression categorization. Facial expression feature extraction and categorization are the most key issue. To address this issue, we propose a method to combine local binary pattern (LBP) and embedded hidden markov model (EHMM), which is the key contribution of this paper This paper first gives an introduction about facial expression recognition and then describes EHMM and LBP Finally, we give out the LBP-EHMM method in facial expression recognition, and perform an experiment to obtain a comparison between LBP feature and discrete cosine transform (DCT) feature.

关键词： Local binary pattern

来源：评论

学校读者我要写书评

暂无评论

A context-sensitive technique based on Support Vector machines for image classification

引用

1st international conference on pattern recognition and machine Intelligence

作者： Bovolo, F Bruzzone, L Univ Trent Dept Informat & Commun Technol I-38050 Povo TN Italy

ISBN: (纸本)3540305068

In this paper, a novel context-sensitive classification technique based on Support Vector machines (CS-SVM) is proposed. This technique aims at exploiting the promising SVM method for classification of 2-D (or n-D) scenes by considering the spatial-context information of the pixel to be analyzed. ne context-based architecture is defined by properly integrating SVMs with a Markov Random Field (MRF) approach. In the design of the resulting system, two main issues have been addressed: i) estimation of the observation term statistic (class-conditional densities) with a proper multiclass SVM architecture;ii) integration of the SVM approach in the framework of MRFs for modeling the prior model of images. Thanks to the effectiveness of the SVM machine learning strategy and to the capability of MRFs to properly model the spatial-contextual information of the scene, the resulting context-sensitive image classification procedure generates regularized classification maps characterized by a high accuracy. Experimental results obtained on Synthetic Aperture Radar (SAR) remote sensing images confirm the effectiveness of the proposed approach.

关键词： image processing

来源：评论

学校读者我要写书评

暂无评论

Exploring the Potential of Convolution Neural Network Based image Classification 1

Exploring the Potential of Convolution Neural Network Based ...

引用

2023 1st DMIHER international conference on Artificial Intelligence in Education and Industry 4.0, IDICAIEI 2023

作者： Sharma, Ritika Bashir, Saimul Tiwary, Vinay Nath Kumar, Sanjeev Chandigarh University Computer Science and Engineering Punjab India

ISBN: (纸本)9798350338423

This research examines the utilization of Convolutional Neural Networks (CNNs) in image classification, shedding light on recent progress in machine learning algorithms. Recognized as a unique component of deep learning, CNNs have earned significant acclaim in a myriad of visual tasks, attributed to their capacity for hierarchical pattern recognition and extraction of stable features. In this investigation, an enhanced CNN model is introduced, which employs cutting-edge preprocessing and augmentation strategies to elevate image readability, thereby streamlining the learning mechanism. key machine learning principles were employed, equipping the CNN model to identify and distinguish intricate structures within images with amplified precision. The model's efficiency is assessed using multiple benchmark datasets, achieving optimistic outcomes concerning precision, recall, and F1-score. The paper also touches upon inherent obstacles such as overfitting and computational expenses linked with CNNs, suggesting remedies to boost scalability and efficiency. The observations suggest that with meticulous planning and execution, CNNs can serve as a potent instrument for automated image classification, paving the way for a wide array of applications from medical imaging to self-driving vehicles. This research augments the expanding literature in machine learning, offering perspectives that could inform the design of more effective and efficient image classification systems. © 2023 IEEE.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

Does the learning principle help us improve hand-crafted finger-knuckle-print features? 1

Does the learning principle help us improve hand-crafted fin...

引用

1st international conference on Communications, Control Systems and Signal processing, CCSSP 2020

作者： Haouam, Mohamed Yassine Meraoumia, Abdallah Bendib, Issam Laimeche, Lakhdar Algeria

ISBN: (纸本)9781728158358

One of the most critical steps in pattern recognition systems is to extract the most distinctive features of the image. Therefore, the accuracy of these systems is strongly related to the efficiency of this step to represent the image. In general, hand-craft image features are obtained using feature extraction methods that often have no prior knowledge about the image type, which may result in poor features. In this paper, a hand-craft image feature learning method, based on Local Binary pattern (LBP), for Finger-Knuckle-Print representation and recognition is applied. Thus, in the learning phase, each training finger image is first divided into several non-overlapping regions, and the feature filter and the codebook of each region are individually learned. For the enrolment phase, the obtained learned filter and codebook are applied to each image to extract the histogram feature of each region. Then, the different histograms are concatenated in a single feature vector to represent the finger features. Finally, an SVM classifier is used for the identification phase. The experimental results clearly show that the learning principle offers a good gain in the system accuracy, which can provide an excellent identification rate and thus improve security. © 2020 IEEE.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

An analytical method for predicting the performance of parallel image processing operations

引用

JOURNAL OF SUPERCOMPUTING 1998年第1-2期12卷 157-174页

作者： Juhasz, Z Univ Veszprem Dept Informat Syst H-8201 Veszprem Hungary

This paper presents an analytical performance prediction model and methodology that can be used to predict the execution time, speedup, scalability and similar performance metrics of a targe set of image processing operations running on a p-processor parallel system. The model which requires only a few parameters obtainable on a minimal system can help in the systematic design, evaluation and performance tuning of parallel image processing systems. Using the model one can reason about the performance of a parallel image processing system prior to implementation. The method can also support programmers in detecting critical parts of an implementation and system designers in predicting hardware performance and the effect of hardware parameter changes on performance. The execution of parallel image processing operations was studied and operations were arranged in three main problem classes based on data locality and the communication patterns of the algorithms. The core of the method is the derivation of the overhead function, as it is the overhead that determines the achievable speedup. The overheads were examined and modelled for each class. The use of the method is illustrated by four class-representative image processing algorithms: image-scalar addition, convolution, histogram calculation and the Fast Fourier Transform. The developed performance model has been validated on a 16-node parallel machine and it has been shown that the model is able to predict the parallel run-time and other performance metrics of parallel image processing operations accurately.

关键词： performance prediction image processing analytical model overheads message-passing communication pattern

来源：评论

学校读者我要写书评

暂无评论

FAst HAND DETECTION AND GEstURE recognition 14

FAST HAND DETECTION AND GESTURE RECOGNITION

引用

international conference on machine learning and Cybernetics (ICMLC)

作者： Wang, Yuh-Rau Syu, Jia-Liang Li, Hsin-Ting Yang, Ling St Johns Univ Dept Comp Sci & Informat Engn Taipei Taiwan Tamkang Univ Dept Elect Engn Taipei Taiwan St Johns Univ Dept Mkt & Logist Management Taipei Taiwan

ISBN: (纸本)9781467372206

A fast hand detection and gesture recognition method is proposed in this paper. To reduce the computation time, we employ symmetric mask-based discrete wavelet transform (SMDWT) to reduce the image resolution and then utilize the extracted characteristics to perform hand detection and gesture recognition. The proposed method reduces about 66% of the overall computation time. Experimental results show that the proposed real time hand detection and gesture recognition methods can detect fast and accurately, and can be implemented on embedded system easily. The average gesture recognition rate is around 97.5%.

关键词： Discrete wavelet transform Hand detection Skin color detection Gesture recognition

来源：评论

学校读者我要写书评

暂无评论

OPTIMIZING FEATURE INTEGRATION USING A machine learning ADAPTIVE SYNTHESIS LAYER

OPTIMIZING FEATURE INTEGRATION USING A MACHINE LEARNING ADAP...

引用

1st CONF ON SCIENCE OF ARTIFICIAL NEURAL NETWORKS

作者： BIELAK, D Westinghouse Electric Corp. (United States)

ISBN: (纸本)0819408751

The classification process for pattern recognition uses sensors to read measurements from input examples. A feature function next reduces and quantizes these measurements into feature vectors (combinations of feature data). Finally, using the feature vectors, a decision function classifies the current example by comparison to a statistical model of feature vector data. Which particular feature vectors are made available to the decision function has usually been determined during the design phase by the person constructing the system depending on the hardware available for the sensors. With an adaptive synthesis layer, however, a collective learning automata learns which feature vectors are contributing to correct classification and dynamically adjusts the decision function accordingly. A weighted average scheme is used to combine multiple subhypotheses of the example's class (known as rank hypotheses) into a single output hypothesis (known as the super hypothesis). Updating the weights depends on two factors: an evaluation score and a feature vector compensation. The score is a collective measure of the weighted average combination of rank hypotheses. The feature vector compensation is an individual measure of each feature vector's contribution to the overall decision based on a history of detected patterns. This two-layer approach is one of the most efficient methods in multi-objective programming, yet the application of this approach to machine learning as proposed in this dissertation is unique. In particular, a collective learning automation is used to enhance the combination of a number of candidate class subhypotheses into as single, unique classification. This process is refereed to as adaptive synthesis. This approach has been applied to black and white character recognition and grey scale block classification using the Adaptive learning image Analysis System (ALIAS) at the George Washington University and the Research Institute for Applied Knowledge Process

关键词： Sensors image classification pattern recognition machine learning Artificial neural networks image processing Classification systems Neural networks Data modeling image analysis

来源：评论

学校读者我要写书评

暂无评论

Enhancement of the image Sequence Using Anisotropic Diffusion of the Deformable Object

Enhancement of the Image Sequence Using Anisotropic Diffusio...

引用

4th international conference on Computer Science and Informatics, JCIS 1998, 1st international Workshop on High Performance, 1st international Workshop on Computer Vision, pattern recognition and image processing Volume 4

作者： Tang, Ming Ma, SongDe Natl. Lab. of Pattern Recognition Institute of Automation Chinese Academy of Sciences P.O.Box 2728 Beijing 100080 China

The image sequence is often met in the researches of computer vision and image processing. Exploring the information concealed in the sequence of images may make some algorithms, such as enhancement ones, more effective and efficient. This paper derives a noval formula to enhance the image sequence. Having considered the heat diffusion equation of the deformable object, we derive a practical one for the enhancement of the sequence of images. Our method may converge faster while still keeps the locations of edges precise and sharp.

关键词： image enhancement

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：