• Smart Structures and Systems
• /
• v.29 no.1
• /
• pp.1-16
• /
• 2022

Despite recent breakthroughs in deep learning and computer vision fields, the pixel-wise identification of tiny objects in high-resolution images with complex disturbances remains challenging. This study proposes a modified U-net for tiny crack segmentation in real-world steel-box-girder bridges. The modified U-net adopts the common U-net framework and a novel Self-Attention-Self-Adaption (SASA) neuron as the fundamental computing element. The Self-Attention module applies softmax and gate operations to obtain the attention vector. It enables the neuron to focus on the most significant receptive fields when processing large-scale feature maps. The Self-Adaption module consists of a multiplayer perceptron subnet and achieves deeper feature extraction inside a single neuron. For data augmentation, a grid-based crack random elastic deformation (CRED) algorithm is designed to enrich the diversities and irregular shapes of distributed cracks. Grid-based uniform control nodes are first set on both input images and binary labels, random offsets are then employed on these control nodes, and bilinear interpolation is performed for the rest pixels. The proposed SASA neuron and CRED algorithm are simultaneously deployed to train the modified U-net. 200 raw images with a high resolution of 4928 × 3264 are collected, 160 for training and the rest 40 for the test. 512 × 512 patches are generated from the original images by a sliding window with an overlap of 256 as inputs. Results show that the average IoU between the recognized and ground-truth cracks reaches 0.409, which is 29.8% higher than the regular U-net. A five-fold cross-validation study is performed to verify that the proposed method is robust to different training and test images. Ablation experiments further demonstrate the effectiveness of the proposed SASA neuron and CRED algorithm. Promotions of the average IoU individually utilizing the SASA and CRED module add up to the final promotion of the full model, indicating that the SASA and CRED modules contribute to the different stages of model and data in the training process.

• Journal of Service Research and Studies
• /
• v.14 no.2
• /
• pp.158-177
• /
• 2024

This paper proposes and tests a theoretical model of the relational link between a novel form of customer-perceived fairness for a reward design (distributive peer justice climate) and C2C helping intention via community identification and online C2C interaction (friend-, neighboring customer-, audience-interaction) qualities in a collective consumption context (MMORPG). To test hypotheses, we amassed survey data within a collective consumption context (massively multiplayer online role-playing games, MMORPGs). We used structural equation modeling in analyzing the survey data. The results reveal that user-perceived distributive peer justice climate for a reward design enhances their C2C helping intention via community identification and C2C interactions in MMORPG contexts. Collective consumption-type service managers should focus on promoting the user-perceived distributive peer justice climate for their reward system to enhance users' present C2C co-creation experience (community identification, C2C interaction) and future C2C co-creation behavior (helping intention). By adopting an intra-unit level distributive justice concept (customer-perceived distributive peer justice climate) to a reward design in a collective consumption context (MMORPGs), this study informed collective consumption-type service managers of the importance of its management.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.3
• /
• pp.55-61
• /
• 2024

In a hybrid architecture that combines the technological advantages of P2P(peer-to-peer) and cloud computing, it is possible to efficiently supply resources and allocate loads. In other words, by appropriately utilizing the processing power of the players constituting P2P as well as the server in the cloud computing environment, MMOG(Massively Multiplayer Online Game) can be configured that considers the scale of economic cost and service quality. In fact, the computing power and communication bandwidth of servers in the cloud are important demand-based resources. The more it is used when renting, the higher the cost, while the quality of service improves. On the other hand, if the player's processing power is utilized a lot, the quality of service deteriorates relatively while the economic cost decreases. In this paper, a bandwidth adjustment technique between servers and players for MMOG based on this hybrid structure is described. When the number of players running at the same time increases, the players' actions are appropriately distributed to servers and players to effectively utilize the server's computing power and communication volume. Simulation results show that in the MMOG based on cloud and P2P hybrid architecture, the bandwidth of the server is proportionally decreased as the bandwidth directly handled by players is increased.

• Cartoon and Animation Studies
• /
• s.45
• /
• pp.397-415
• /
• 2016

This paper will investigate how animated practice can be a research form as practice-led research in an ethnography approach. This practice-led research will explore the issue of the construction of contemporary identities (based on the strange case of Dr. Jekyll and Mr. Hyde) and in particular, the Korean youth culture and identity, exemplified, for example, creation of 'avatars' in the virtual characters of animated online games such as Massively Multiplayer Online Role-playing Games (MMORPGs). In this proposed discussion, I will argue that the sudden period of change in contemporary Korea bears some resemblance to the Victorian era as explored in gothic fiction (e.g. Jekyll and Hyde). In this sense, my animation investigates the connection between the fictional Jekyll and Hyde and a real murder incident by a young Korean boy, which actually happened on the 16th November 2010, in SouthKorea.I will, therefore, construct this practice-led research to obtain the primary data consisted of online and offline practices in 'social ethnography'. These practices engage with specific Korean youth identity, comparing the 'avatar' with the real lives of participants. However, this paper will only focus on the (ethnographic) research process and strategy, using animated (visual) practices, rather than giving the meaning of the specific case of 'Korean-ness'. Eventually, I will explore the four different animated representations as it presents the distinctive animated realties or documentaries by online and offline practices. My intention is to visually interpret the issue of 'Korean-ness' within its socio-cultural context, adapting the convention and code of Jekyll and Hyde concept into an animated documentary in the 'virtual' world (auto-animated documentary by recording avatar interviews and online game footages) and the 'real' world (self-created animated documentary, based on real people and events).