Jun Xiang, Ruru Pan and Weidong Gao
The paper aims to propose a novel method based on deep sparse convolutional neural network (CNN) for clothing recognition. A CNN based on inception module is applied to bridge…
Abstract
Purpose
The paper aims to propose a novel method based on deep sparse convolutional neural network (CNN) for clothing recognition. A CNN based on inception module is applied to bridge pixel-level features and high-level category labels. In order to improve the robustness accuracy of the network, six transformation methods are used to preprocess images. To avoid representational bottlenecks, small-sized convolution kernels are adopted in the network. This method first pretrains the network on ImageNet and then fine-tune the model in clothing data set.
Design/methodology/approach
The paper opts for an exploratory study by using the control variable comparison method. To verify the rationality of the network structure, lateral contrast experiments with common network structures such as VGG, GoogLeNet and AlexNet, and longitudinal contrast tests with different structures from one another are performed on the created clothing image data sets. The indicators of comparison include accuracy, average recall, average precise and F-1 score.
Findings
Compared with common methods, the experimental results show that the proposed network has better performance on clothing recognition. It is also can be found that larger input size can effectively improve accuracy. By analyzing the output structure of the model, the model learns a certain “rules” of human recognition clothing.
Originality/value
Clothing analysis and recognition is a meaningful issue, due to its potential values in many areas, including fashion design, e-commerce and retrieval system. Meanwhile, it is challenging because of the diversity of clothing appearance and background. Thus, this paper raises a network based on deep sparse CNN to realize clothing recognition.
Details
Keywords
Ning Zhang, Ruru Pan, Lei Wang, Shanshan Wang, Jun Xiang and Weidong Gao
The purpose of this paper is to propose a novel method using support vector machine (SVM) classifiers for objective seam pucker evaluation. Features are extracted using wavelet…
Abstract
Purpose
The purpose of this paper is to propose a novel method using support vector machine (SVM) classifiers for objective seam pucker evaluation. Features are extracted using wavelet analysis and gray-level co-occurrence matrix (GLCM), and the samples are evaluated using SVM classifiers. The study aims to solve the problem of inappropriate parameters and large required samples in objective seam pucker evaluation.
Design/methodology/approach
Initially, seam pucker image was captured, and Edge detection and Hough transform were utilized to normalize the seam position and orientation. After cropping the image, the intensity was adjusted to the same identical level through histogram specification. Then, the standard deviations of the horizontal image and diagonal image, reconstructed using wavelet decomposition and reconstruction, were calculated based on parameter optimization. Meanwhile, GLCM was extracted from the restructured horizontal detail image, then the contrast and correlation of GLCM were calculated. Finally, these four features were imported to SVM classifiers based on genetic algorithm for evaluation.
Findings
The four extracted features reflected linear relationships among five grades. The experimental results showed that the classification accuracy was 96 percent, which catches up to the performance of human vision, and resolves ambiguity and subjective of the manual evaluation.
Originality/value
There are large required samples in current research. This paper provides a novel method using finite samples, and the parameters of the methods were discussed for parameter optimization. The evaluation results can provide references for analyzing the reason of wrinkles during garment manufacturing.
Details
Keywords
Junlong Peng and Xiang-Jun Liu
This research is aimed to mainly be applicable to expediting engineering projects, uses the method of inverse optimization and the double-layer nested genetic algorithm combined…
Abstract
Purpose
This research is aimed to mainly be applicable to expediting engineering projects, uses the method of inverse optimization and the double-layer nested genetic algorithm combined with nonlinear programming algorithm, study how to schedule the number of labor in each process at the minimum cost to achieve an extremely short construction period goal.
Design/methodology/approach
The method of inverse optimization is mainly used in this study. In the first phase, establish a positive optimization model, according to the existing labor constraints, aiming at the shortest construction period. In the second phase, under the condition that the expected shortest construction period is known, on the basis of the positive optimization model, the inverse optimization method is used to establish the inverse optimization model aiming at the minimum change of the number of workers, and finally the optimal labor allocation scheme that meets the conditions is obtained. Finally, use algorithm to solve and prove with a case.
Findings
The case study shows that this method can effectively achieve the extremely short duration goal of the engineering project at the minimum cost, and provide the basis for the decision-making of the engineering project.
Originality/value
The contribution of this paper to the existing knowledge is to carry out a preliminary study on the relatively blank field of the current engineering project with a very short construction period, and provide a path for the vast number of engineering projects with strict requirements on the construction period to achieve a very short construction period, and apply the inverse optimization method to the engineering field. Furthermore, a double-nested genetic algorithm and nonlinear programming algorithm are designed. It can effectively solve various optimization problems.
Details
Keywords
Hong-Yu Yao, Xiang-Jun Kong, Ya-Jie Shi, Xian-Bo Xiao and Ning-Ning Le
Engineered material arresting systems (EMASs) are dedicated to stopping aircraft that overrun the runway before they enter dangerous terrain. The system consists of low-strength…
Abstract
Purpose
Engineered material arresting systems (EMASs) are dedicated to stopping aircraft that overrun the runway before they enter dangerous terrain. The system consists of low-strength foamed concretes. The core component of the arresting system design is a reliable simulation model. Aircraft test verification is required before the practical application of the model. This study aims to propose a simulation model for the arresting system design and conducts serial verification tests.
Design/methodology/approach
Six verification tests were conducted using a Boeing 737 aircraft. The aircraft was equipped with an extra inertia navigation system and a strain gauge system to measure its motion and the forces exerted on the landing gears. The heights of the arrestor beds for these tests were either 240 or 310 mm, and the entering speeds of the aircraft ranged from 23.9 to 60.6 knots.
Findings
Test results revealed that both the aircraft and the pilots on board were safe after the tests. The maximum transient acceleration experienced by the dummies on board was 2.5 g, which is within the human tolerance. The model exhibited a satisfied accuracy to the field tests, as the calculation errors of the stopping distances were no greater than 7 per cent.
Originality/value
This study proposes a simulation model for the arresting system design and conducts serial verification tests. The model can be used in EMAS design.
Details
Keywords
Michelle Thompson, Leonie Cassidy, Bruce Prideaux, Anja Pabel and Allison Anderson
This research looks at the significance of friends and relatives as an information source for consumers planning holidays. Recent research has largely ignored friends and…
Abstract
This research looks at the significance of friends and relatives as an information source for consumers planning holidays. Recent research has largely ignored friends and relatives as destination information sources and has focused instead on the Internet. Two categories of friends and relatives are identified, friends and relatives who live in a destination and friends and relatives who have visited a destination of interest. An exit survey of 1,203 tourists departing a major international destination in Australia found that while the Internet was an important source of information, friends and relatives were as important, if not more, regardless of country of origin and age. These findings indicate that information from friends and relatives and the Internet are complementary rather than exclusive in the minds of consumers.
Details
Keywords
Qianjin Dong, Xueshan Ai, Guangjing Cao, Yanmin Zhang and Xianjia Wang
The purpose of this paper is to obtain risk indicators of water security of drought periods in which the indices of reliability, resiliency, and vulnerability are integrated.
Abstract
Purpose
The purpose of this paper is to obtain risk indicators of water security of drought periods in which the indices of reliability, resiliency, and vulnerability are integrated.
Design/methodology/approach
It is not reasonable that weight coefficients of different risk indices are often determined subjectively in conventional procedures, so the entropy weight method is introduced and chosen to solve the problem. Entropy weight method can get the weight coefficients of different risk indices objectively and is valid from the case study.
Findings
The feasibility and validity of entropy weight methods to determine weight coefficients of different risk indices objectively are recognized.
Research limitations/implications
Accessibility and availability of data are the main limitations.
Practical implications
The paper provides a more objective risk indicator of water security of drought periods for water resources managers.
Originality/value
This paper determines the weight coefficients of different risk indices for risk assessment of water security of drought periods based on hazard entropy. The paper is aimed at water resources managers and relative researchers, especially those who deal with risk assessment of water security of drought periods.
Details
Keywords
Jun Yong Xiang, Zhen He, Yung Ho Suh, Jae Young Moon and Ya Fen Liu
The purpose of this paper is to analyze the causal relationships among categories in the China Quality Award (CQA) model based on the Malcolm Baldrige National Quality Award model.
Abstract
Purpose
The purpose of this paper is to analyze the causal relationships among categories in the China Quality Award (CQA) model based on the Malcolm Baldrige National Quality Award model.
Design/methodology/approach
The paper identifies seven factors from CQA categories: leadership, strategic planning, human resource focus, process management, customer and market focus, information and analysis, and results. Extending the basic Baldrige theory “Leadership drives the system that creates results,” this paper identifies driver (leadership), direction (strategic planning), foundation (information and analysis), system (human resource focus, process management, and customer and market focus), and results(business results). Structural equation model (SEM) is used to analyze the empirical data and estimate the path coefficients among CQA categories.
Findings
First, driver has not only a direct influence on results, but also has an indirect influence on results through system. Leadership has a great influence on foundation and direction. Second, direction affects human resource focus and customer and market focus of system while it has no influence on process management. Third, human resource focus and customer and market focus both affect process management, and process management has a significant impact on results. Fourth, foundation affects direction and all of the categories of system.
Originality/value
There are few studies which try to analyze the causal relationships among categories in the CQA model.
Details
Keywords
Xu Zhang, Mingling Zhai, Yanyan Wang, Yulei Gao, Haoliang Zhao, Xiang Zhou and Jun Gao
In order to verify the feasibility of different techniques, this chapter further studies the adaptability of two massive straw biomass applications in rural areas in China.
Abstract
Purpose
In order to verify the feasibility of different techniques, this chapter further studies the adaptability of two massive straw biomass applications in rural areas in China.
Methodology/approach
The methods of assessing biomass power generation project with Life Cycle Assessment (LCA), survey and field test of one biogas station, and game-theoretic analysis are adopted.
Findings
The following conclusions can be drawn: The air pollution costs account for more than 60% of the total environmental cost, followed by depreciation expense and maintenance fee of 18%, compared to that of biomass power generation at 0.01711 CNY/kWh. The adopted greenhouse sunlight technology of Solar Biogas Plant in Xuzhou, China, raises the inside average temperature by 11.0 °C higher than outside and keeps the pool temperature above 16 °C in winter, ensuring a gas productivity of biogas project in winter up to 0.5–0.7 m3/m3 by volume. This chapter also analyzes the information cost incurred by asymmetric information in biomass power generation via game theory method and illustrates the information structure with game results. It provides not only a foundation for the policy research in promoting straw power generation but also theoretical framework to solve the problem of straw collection.
Social implications
These studies will propose solutions to relevant problems arisen in the running process.
Originality/value
These studies are all based on real cases, field research, and appropriate theoretical analyses, so, they can reduce the relevant costs and promote the application of relevant technologies.
Details
Keywords
Solomon W. Polachek and Jun (Jeff) Xiang
In this paper, we define a tractable procedure to measure worker incomplete information in the labor market. The procedure, which makes use of earnings distribution skewness, is…
Abstract
In this paper, we define a tractable procedure to measure worker incomplete information in the labor market. The procedure, which makes use of earnings distribution skewness, is based on econometric frontier estimation techniques, and is consistent with search theory. We apply the technique to 11 countries over various years, and find that incomplete information leads workers to receive on average about 30–35% less pay than they otherwise would have earned, had they information on what each firm paid. Generally, married men and women suffer less from incomplete information than the widowed or divorced; and singles suffer the most. Women suffer more from incomplete information than men. Schooling and labor market experience reduce these losses, but institutions within a country can reduce them, as well. For example, we find that workers in countries that strongly support unemployment insurance (UI) receive wages closer to their potential, so doubling UI decreases incomplete information and results in 5% higher wages. A more dense population reduces search costs leading to less incomplete information. A more industrial economy disseminates wage information better, so workers exhibit less incomplete information and higher wages. Finally, we find that foreign worker inflows increase incomplete information, and at the same time reduce average wage levels, at least in the short run.