The histogram in each patch of the input image is a useful feature applied for various development of image processing techniques. However, if the size of the input image is very large, the... Show More

The histogram in each patch of the input image is a useful feature applied for various development of image processing techniques. However, if the size of the input image is very large, the histogram construction of each patch in the image becomes very time-consuming. For applications involving the processing of several very large images, this paper proposes a superior patchwise histogram construction algorithm based on cloud-computing architecture that is faster than similar state-of-the-art approaches. Through the modern communication network, the computation cost can be easily shared to construct several patchwise histograms at the same time. The proposed algorithm is the fastest solution in the field as well as applicable to various data processing procedures related to probability distribution. Experimental results show that the proposed algorithm has the best performance compared to other related algorithms.

In this paper, the authors propose an effective scheme for association rule mining of personal hobbies in social networks. By introducing the connection and clipping techniques, the authors are able... Show More

In this paper, the authors propose an effective scheme for association rule mining of personal hobbies in social networks. By introducing the connection and clipping techniques, the authors are able to ignore unrelated items in the process of finding frequent itemsets, resulting in more accurate candidate itemsets. More specifically, set operations, which are used in the process of combining frequent itemsets, can dramatically reduce the number of databases visited. Furthermore, to explore more practical rules, interestingness level is also introduced to eliminate rules that few people are interested in. The authors' proposed association rule mapping is shown to be able to provide new insights for supporting personalized services and virtual marketing.

This study proposes a system to automatically identify multiple singers in a long audio stream that may have singing voices overlapping in time. The system is of great help in handling the rapid... Show More

This study proposes a system to automatically identify multiple singers in a long audio stream that may have singing voices overlapping in time. The system is of great help in handling the rapid proliferation of music data. To achieve this, an audio stream is segmented into a sequence of consecutive, non-overlapping, fixed-length clips using a sliding window, and then undergoes solo/duet recognition, single singer identification, and duet singer identification. Furthermore, the authors propose a triangulation-based decision approach to improve the performance of singer identification. The results of the experiment conducted using a database of solo and duet a cappella demonstrate feasibility of the proposed system.

For building the big data collection infrastructure, a vision-based smart meter-reading network and its gateway is provided for a community gas supply system that uses traditional mechanical meters.... Show More

For building the big data collection infrastructure, a vision-based smart meter-reading network and its gateway is provided for a community gas supply system that uses traditional mechanical meters. In the network architecture, the gas meter readings are captured by embedded image sensor nodes and then transmitted to a newly designed gateway for image recognition and are collected in the embedded database of gateway. The Web-based monitoring system designed using HTML5 is applicable to a mobile device which allows a user to monitor household gas consumption and history and allows a gas company to develop an effective energy management system to analyze community users' energy consumption models using the big data collected in the database.