Search

The learning‐based multiview stereo (MVS) methods for three‐dimensional (3D) reconstruction generally use 3D volumes for depth inference. The quality of the reconstructed depth maps and the corresponding point clouds is directly influenced by the spatial resolution of the 3D volume. Consequently, these methods produce point clouds with sparse local regions because of the lack of the memory required to encode a high volume of information. Here, we apply the atrous spatial pyramid pooling (ASPP) module in MVS methods to obtain dense feature maps with multiscale, long‐range, contextual information using high receptive fields. For a given 3D volume with the same spatial resolution as that in the MVS methods, the dense feature maps from the ASPP module encoded with superior information can produce dense point clouds without a high memory footprint. Furthermore, we propose a 3D loss for training the MVS networks, which improves the predicted depth values by 24.44%. The ASPP module provides state‐of‐the‐art qualitative results by constructing relatively dense point clouds, which improves the DTU MVS dataset benchmarks by 2.25% compared with those achieved in the previous MVS methods. [ABSTRACT FROM AUTHOR]

Recently, manufacturing companies have been improving quality and productivity, reducing costs, and producing customized products according to Industry 4.0. The global value chain (GVC) is also being reorganized and manufacturing companies are recovering the connectivity of value chains based on, e.g., the regional value chain (RVC) and reshoring. With the advent of Industry 4.0, many manufacturing companies are introducing smart factories. A new type of manufacturing execution system (MES), a core system of smart factories, is necessary, owing to the new technologies and the increase in collaboration between companies. Here, we present the framework, development, and application processes of a "cloud-based collaborative MES System" to support the value chain of "order-design-production-delivery" for the manufacture of personalized sportswear products in the fashion industry in Korea. To this end, first, nine future MES deployment directions and frameworks are presented. Second, we present the UML modeling, conceptual framework, and functional framework for MES system development, considering six future MES establishment directions such as cloud and collaboration. Third, the application and effect of the designed and developed cloud-based collaborative MES system are analyzed for design, fabric, printing, and sewing companies that play a role in each stage of the sportswear value chain. [ABSTRACT FROM AUTHOR]

After the Industry 4.0 discussion in Germany in 2011, much attention has been paid to smart factory in Korea. Since 2014, smart factories have been established and expanded in Korea. However, about 80% of them were established at a low level. In this paper, we analyze smart factory statuses in detail through an empirical research on 113 manufacturing companies that have established smart factories in Korea. We build a framework based on the resource-based view (RBV) and IT value creation process and analyze the results of five constructs—manufacturing strategy, organization, system, process, and performance—using basic statistical methodologies to derive the current statuses of manufacturing companies that have established smart factories. Our results show that implementing advanced technologies such as AI technology that can implement semi-finished and finished product quality inspection, manufacturing process optimization and product demand forecast is a challenge, particularly for SMEs. We also find that securing and managing facility data is a difficult problem. In addition, while output and material management ranked high, the utilization of integration systems, which is important when building a smart factory, was found to be extremely low. Lastly, the performance indicator results showed that yield management and defect rate were most important, while job creation through the introduction of smart factories was low. Based on the results of this study, the government may be able to determine effective smart factory policies and provide manufacturing companies with a guide on establishing a smart factory. [ABSTRACT FROM AUTHOR]

In a company, project management is responsible for project selection from candidates under some limited constraints to achieve the company's goal before the project begins as well as the project operations in progress. The development of new technologies and products can broaden a company's market share, and to do so, research and development (R&D) projects are significant. However, limited funds force a company to select projects that can best represent the company's interests. As projects may take a long time to develop, a number of uncertainties may occur, and the most concerning uncertainty is cost uncertainty. In this study, a robust optimization decision model for project selection considering cost uncertainty is proposed to assist the decision-making process for companies that need to select projects from a number of candidates due to limited funds. The model considers project selection in view of the total cost of ownership, which is a key factor for customers and companies in the automobile industry. The proposed model is tested in the automobile industry environment with different conservatism levels about cost uncertainty, and an analysis of expected market changes and a company's income is performed with the solutions obtained from the proposed model. The result shows that the presented model reacts to cost uncertainty robustly for assisting the decision-makers in the company. [ABSTRACT FROM AUTHOR]

In the injection mold industry, it is important for manufacturers to satisfy the delivery date for the products that customers order. The mold products are diverse, and each product has a different manufacturing process. Owing to the nature of mold, mold manufacturing is a complex and dynamic environment. To meet the delivery date of the customers, the scheduling of mold production is important and is required to be sustainable and intelligent even in the complicated system and dynamic situation. To address this, in this paper, deep reinforcement learning (RL) is proposed for injection mold production scheduling. Before presenting the RL algorithm, a mathematical model for the mold scheduling problem is presented, and a Markov decision process framework is proposed for RL. The deep Q -network, which is an algorithm for RL, is employed to find the scheduling policy to minimize the total weighted tardiness. The results of experiments demonstrate that the proposed deep RL method outperforms the dispatching rules that are presented for minimizing the total weighted tardiness. [ABSTRACT FROM AUTHOR]