Spectral clustering is a well-known graph-theoretic clustering algorithm. Although spectral clustering has several desirable advantages (such as the capability of discovering non-convex clusters and applicability to any data type), it often leads to incorrect clustering results because of high sensitivity to noise points. In this study, we propose a robust spectral clustering algorithm known as KNN-SC that can discover exact clusters by decreasing the influence of noise points. To achieve this goal, we present a novel approach that filters out potential noise points by estimating the density difference between data points using $k$ -nearest neighbors. In addition, we introduce a novel method for generating a similarity graph in which various densities of data points are effectively represented by expanding the nearest neighbor graph. Experimental results on synthetic and real-world datasets demonstrate that KNN-SC achieves significant performance improvement over many state-of-the-art spectral clustering algorithms.

In this paper, we propose a dance pose evaluation method for a dance learning application using a smartphone. In the past, methods for classifying and comparing dance gestures through 3-D joint information obtained through a 3-D camera have been proposed, but there is a problem in using them for accurate dance pose evaluation. That is, these methods simply compare the similarity between the dance gestures without evaluation of the exact dance pose. To solve this problem, we propose a new method that can be operated on a smartphone for exact dance pose evaluation that simultaneously performs an affine transformation and an evaluation method to compare the joint position and joint angle information. In addition, we prove that the proposed method is suitable for dance learning applications through comparative experiments on a smartphone with real-world datasets.

In this study, we propose a variety of dance pose evaluation methods required for a smartphone-based application that helps dance self-learning. Previously, various dance-pose-evaluation methods and dance-learning applications have been proposed for the same purpose, but some problems were identified in the actual application process, such as the absence of joints due to occlusion, the overfitting of the torso, and the evaluation of rotational poses. To mitigate these problems, this study proposes three methods to a) solve the view occlusion problem that precludes evaluation by creating a state variable referred to as visibility, b) solve the torso overfitting problem by creating a keypoint referred to as the core, and c) estimate the rotation of the body based on a vector obtained between the core and torso joints. Furthermore, we conducted experiments using a dance self-learning application running on a real smartphone and various real-world datasets. Based on the experimental results, we demonstrate that the newly proposed method can provide effective feedback to students who try to learn to dance alone, by performing robust dance pose evaluations in different scenarios.

The skyline query has recently attracted a considerable amount of research interest in several fields. The query conducts computations using the domination test, where “domination” means that a data point does not have a worse value than others in any dimension, and has a better value in at least one dimension. Therefore, the skyline query can be used to construct efficient queries based on data from a variety of fields. However, when the number of dimensions or the amount of data increases, naïve skyline queries lead to a degradation in overall performance owing to the higher cost of comparisons among data. Several methods using index structures have been proposed to solve this problem but have not improved the performance of skyline queries because their indices are heavily influenced by the dimensionality and data amount. Therefore, in this study, we propose HI-Sky, a method that can perform quick skyline computations by using the hash index to overcome the above shortcomings. HI-Sky effectively manages data through the hash index and significantly improves performance by effectively eliminating unnecessary data comparisons when computing the skyline. We provide the theoretical background for HI-Sky and verify its improvement in skyline query performance through comparisons with prevalent methods.