检索结果-南通市图书馆

PD-YOLO:Colon Polyp Detection Model Based on Enhanced Small-Target Feature Extraction

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2025年第1期82卷 913-928页

作者： Yicong Yu Kaixin Lin Jiajun Hong Rong-Guei tsai Yuanzhi Huang New Engineering Industry College Putian UniversityPutian351100China Engineering Research Center of Big Data Application in Private Health Medicine(Putian University) Fujian Province UniversityPutian351100China Computer College Guangdong University of Science&TechnologyDongguan523668China

In recent years,the number of patientswith colon disease has increased *** polyps are the precursor lesions of colon *** not diagnosed in time,they can easily develop into colon cancer,posing a serious threat to patients’lives and health.A colonoscopy is an important means of detecting colon ***,in polyp imaging,due to the large differences and diverse types of polyps in size,shape,color,etc.,traditional detection methods face the problem of high false positive rates,which creates problems for doctors during the diagnosis *** order to improve the accuracy and efficiency of colon polyp detection,this question proposes a network model suitable for colon polyp detection(PD-YOLO).This method introduces the self-attention mechanism CBAM(Convolutional Block Attention Module)in the backbone layer based on YOLOv7,allowing themodel to adaptively focus on key information and ignore the unimportant *** help themodel do a better job of polyp localization and bounding box regression,add the SPD-Conv(Symmetric Positive Definite Convolution)module to the neck layer and use deconvolution instead of *** results indicate that the PD-YOLO algorithm demonstrates strong robustness in colon polyp *** to the original YOLOv7,on the Kvasir-SEG dataset,PD-YOLO has shown an increase of 5.44 percentage points in AP@0.5,showcasing significant advantages over other mainstream methods.

关键词： Polyp detection YOLOv7 SPD-Conv CBAM deconvolution

SPIN WAVES IN MAGNETIC SUPERLATTICES

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Chinese Physics Letters 1984年第2期1卷 92-94页

作者： MA Hong-ru tsai Chien-hua Institute of Solid State Physics Nanjing UniversityNanjing Institute of Solid State Physics Nanjing UniversityNanjingand Department of PhysicsChiao-tung UniversityShanghai

We discuss,in terms of a simple model,the spin wave excitations,spontaneous magnetization and response to alternating fields of magnetic superlattices.

关键词： excitation magnetization alternating

Regenerable adsorbent for removing ammonia evolved from anaerobic reaction of animal urine

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Environmental Sciences 2006年第6期18卷 1176-1181页

作者： CHOU Liang-hsing tsai ru-in CHANG Jen-ray LEE Maw-tien Department of Civil and Water Resource Engineering "National" Chiayi UniversityChia-YiTaiwanChina Department of Chemical Engineering "National" Chung Cheng UniversityChia-YiTaiwanChina Department of Applied Chemistry "National" Chiayi UniversityChia-YiTaiwanChina.

The waste gas evolved from biodegradation of animal urine contains ammonia causing environmental concerns. A new and effective method for removing ammonia from such waste gas using reactive adsorption is presented. In the process, activated carbon impregnated with H2SO4（H2SO4/C） is employed. Ammonia in the waste gas reacts with H2SO4 on the adsorbent instantaneously and completely to form （NIL）2SO4. The H2SO4/C adsorbent is high in NH3 adsorption capacity and regenerable. The NH3 removal capacity of this regenerable adsorbent is more than 30 times that of the adsorbents used normally in the industry. The spent H2SO4/C is regenerated by flowing low-pressure steam through the adsorbent bed to remove the （NH4）2SO4 from the adsorbent. The regeneration by-product is concentrated （NH4）2SO4 solution, which is a perfect liquid fertilizer for local use. Re-soaking the activated carbon with H2SO4 solution rejuvenates the activity of the adsorbent. Thus the H2SOJC can be reused repeatedly. In the mechanism of this reactive adsorption process, trace of H20 in the waste gas is a required, which lends itself to treating ammonia gas saturated with moisture from biodegradation of animal urine.

关键词： ammonia removal regenerable adsorbent reactive adsorption biodegradation of animal urine activated carbon

A Comparative Study on Polynomial Expansion Method and Polynomial Method of Particular Solutions

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Advances in Applied Mathematics and Mechanics 2022年第3期14卷 577-595页

作者： Jen-Yi Chang ru-Yun Chen Chia-Cheng tsai Tainan University of Technology Tainan 71002Taiwan China Department of Harbor and River Engineering Taiwan Ocean UniversityKeelung 202301Taiwan China Bachelor Degree Program in Ocean Engineering and Technology Taiwan Ocean UniversityKeelung 202301Taiwan China Center of Excellence for Ocean Engineering Taiwan Ocean UniversityKeelung 202301Taiwan China

In this study,the polynomial expansion method(PEM)and the polynomial method of particular solutions(PMPS)are applied to solve a class of linear elliptic partial differential equations(PDEs)in two dimensions with constant coefﬁ*** the solution procedure,the sought solution is approximated by the Pascal polynomials and their particular solutions for the PEM and PMPS,*** multiple-scale technique is applied to improve the conditioning of the resulted linear equations and the accuracy of numerical results for both of the PEM and *** mathematical statements are provided to demonstrate the equivalence of the PEM and PMPS bases as they are both bases of a certain polynomial vector ***,some numerical experiments were conducted to validate the implementation of the PEM and *** results demonstrated that the PEM is more accurate and well-conditioned than the PMPS and the multiple-scale technique is essential in these polynomial methods.

关键词： Pascal polynomial polynomial expansion method polynomial method of particular solutions collocation method multiple-scale technique

Tailored Finite Point Method for Numerical Solutions of Singular Perturbed Eigenvalue Problems

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Advances in Applied Mathematics and Mechanics 2014年第3期6卷 376-402页

作者： Houde Han yin-Tzer Shih Chih-Ching tsai Department of Mathematical Sciences Tsinghua UniversityBeijing 100084China Department of Applied Mathematics National Chung Hsing UniversityTaichung 40227Taiwan

We propose two variants of tailored finite point(TFP)methods for discretizing two dimensional singular perturbed eigenvalue(SPE)problems.A continuation method and an iterative method are exploited for solving discretized systems of equations to obtain the eigen-pairs of the *** study the analytical solutions of two special cases of the SPE,and provide an asymptotic analysis for the *** theoretical results are verified in the numerical *** numerical results demonstrate that the proposed schemes effectively resolve the delta function like of the eigenfunctions on relatively coarse grid.

关键词： Singular perturbation tailored finite point Schrodinger equation eigenvalue problem

Source-Aware Embedding Training on Heterogeneous Information Networks

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Data Intelligence 2023年第3期5卷 611-635页

作者： tsai Hor Chan Chi Ho Wong Jiajun Shen Guosheng yin Department of Statistics and Actuarial Science The University of Hong Kong Hong KongChina Department of Computer Science and Engineering The Hong Kong University of Science and Technology Hong KongChina TCLCorporateResearch(Hong Kong)HongKong China

Heterogeneous information networks(HINs)have been extensively applied to real-world tasks,such as recommendation systems,social networks,and citation *** existing HIN representation learning methods can effectively learn the semantic and structural features in the network,little awareness was given to the distribution discrepancy of subgraphs within a single ***,we find that ignoring such distribution discrepancy among subgraphs from multiple sources would hinder the effectiveness of graph embedding learning *** motivates us to propose SUMSHINE(Scalable Unsupervised Multi-Source Heterogeneous Information Network Embedding)-a scalable unsupervised framework to align the embedding distributions among multiple sources of an *** results on real-world datasets in a variety of downstream tasks validate the performance of our method over the state-of-the-art heterogeneous information network embedding algorithms.

关键词： Heterogeneous information network Graph representation learning Distribution alignment Recommendation system Adversarial learning Graph neural network

Sub-micron-structure machining on silicon by femtosecond laser

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

中国有色金属学会会刊：英文版 2009年第B09期19卷 171-177页

作者： Hung-yin tsai Shao-Wei LUO Chih-Wei WU Shing-Hoa WANG Department of Power Mechanical Engineering National Tsing Hua University Department of Mechanical and Mechatronic Engineering National Taiwan Ocean University

Sub-micron-structure machining on silicon substrate was studied by direct writing system of femtosecond laser with the central wavelength of 800 nm, pulse duration of 120 fs and repetition rate of 1 kHz. Three kinds of experiments were conducted: 1) the effect of photoresist on silicon; 2) machinability of different orientations of silicon; and 3) the size of micro-structure and the cross-section shape. Photoresist SU8 was coated onto silicon substrates in thicknesses of 100 μm. SU8 remained on the silicon substrate while the silicon under the machined SU8 was removed after laser machining. Orientations of (100), (110), and (111) silicon substrates were machined with the laser power of 60 μW and the scanning speed of 3 mm/min. Spike morphologies were observed on all three orientations of silicon substrates without obvious directional difference of these spikes on different silicon substrates. In addition, the ablation threshold energies were also similar. In the consideration of ablation energy, one numerical model of the machining parameters has been proposed to simulate the cross-section of the micro-structure. The predicted shape by simulation can fit the profile of the cross-section shape well.

关键词：亚微米结构激光加工硅基底飞秒激光直接写入断面形状微观结构模拟预测

An Intelligent Recommendation System for Real Estate Commodity

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Computer Systems Science & Engineering 2022年第9期42卷 881-897页

作者： Tsung-yin Ou Guan-Yu Lin Hsin-Pin Fu Shih-Chia Wei Wen-Lung tsai Department of Marketing and Distribution Management National Kaohsiung University of Science and TechnologyKaohsiung824Taiwan College of Management National Kaohsiung University of Science and TechnologyKaohsiung824Taiwan Department of Information Management Asia Eastern University of Science and TechnologyNew Taipei220Taiwan

Most real estate agents develop new objects by visiting unfamiliar clients, distributing leaflets, or browsing other real estate trading website platforms,whereas consumers often rely on websites to search and compare prices when purchasing real property. In addition to being time consuming, this search processrenders it difficult for agents and consumers to understand the status changes ofobjects. In this study, Python is used to write web crawler and image recognitionprograms to capture object information from the web pages of real estate agents;perform data screening, arranging, and cleaning;compare the text of real estateobject information;as well as integrate and use the convolutional neural networkof a deep learning algorithm to implement image recognition. In this study, dataare acquired from two business-to-consumer real estate agency networks, i.e., theSinyi real estate agent and the Yungching real estate agent, and one consumer-toconsumer real estate agency platform, i.e., the, FiveNineOne real estate agent. Theresults indicate that text mining can reveal the similarities and differences betweenthe objects, list the number of days that the object has been available for sale onthe website, and provide the price fluctuations and fluctuation times during thesales period. In addition, 213,325 object amplification images are used as a database for training using deep learning algorithms, and the maximum image recognition accuracy achieved is 95%. The dynamic recommendation system for realestate objects constructed by combining text mining and image recognition systems enables developers in the real estate industry to understand the differencesbetween their commodities and other businesses in approximately 2 min, as wellas rapidly determine developable objects via comparison results provided by thesystem. Meanwhile, consumers require less time in searching and comparingprices after they have understood the commodity dynamic information, therebyallowing the

关键词： Real estate agency web crawler image comparison text mining deep learning real estate object dynamic recommendation system

A HYBRID MODEL OF KNOWLEDGE MANAGEMENT FOR NEW SERVICE DEVELOPMENT

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Systems Science and Systems Engineering 2016年第4期25卷 424-447页

作者： August tsai Department of Industrial Management College of Management Chien Hsin University of Science and Technology Taoyuan32097 Taiwan China

Although many theories regarding the implementation of knowledge management （KM） in organizations have been proposed and studied, most applications tend to stand alone without incorporating the business processes. Different categories of knowledge provide different benefits and how to integrate various categories of KM into a hybrid approach as an effective KM manner remains strategically important, and yet is still understudied. Therefore, in this paper a hybrid model that integrates principal KM applications for new service development （NSD） and the measurement of the resulting f＇mancial benefits have been developed. The proposed KM model incorporates newsgroups, knowledge forums, knowledge asset management and knowledge application processes as a hybrid means for sharing organizational knowledge along two axes, explicit vs. implicit and individual vs. collective. One of the largest management consulting companies in Taiwan, China, whose process model of NSD was standing alone with KM applications, was selected for the case study. A set of hybrid KM processes was developed to implement the proposed KM model, and it illustrates an application with greater financial benefits for integrating hybrid KM practices into the business process. Based on knowledge value added （KVA） validation, the proposed KM model provides a new operating system for sharing NSD knowledge within an organization. Through the case study by measuring the achieved financial results, the proposed KM model is found to provide an exclusive hybrid platform with an empirical process model to address innovative approaches and practical values of KM within an organization.

关键词： Knowledge management new service development (NSD) newsgroup knowledge forum knowledge asset knowledge application process knowledge value added (KVA)