In this paper, the authors provide a complete hardware/software solution for checking the movements of a trainee with a trainer, using an interactive dance game. The authors propose 3D motion capture sensors as the core hardware element, and a “progress block matching” algorithm as the core movement-matching algorithm.
Though the study is interesting, the authors should have considered the following points. First, they should have provided stronger reasons to support the idea of using 3D motion capture sensors. Many other types of sensors, such as the accelerometer and the gyroscope, can also do the job. Second, the literature survey portion of the paper should have included more studies since, in general, many studies related to this research exist. Third, in Subsection 4.2.2, the authors criticize the dynamic time warping (DTW) algorithm, which is not applicable to their approach. However, their reason--“We do not know when the player finishes the current move”--needs more referencing and proof. Some studies exist that segment movements and use DTW for movement matching.
Finally, they could have discussed a bit about how the system could be more adaptive to users.