Browsing by Author "Lin, Chin-Hsien"

Now showing 1 - 2 of 2

COHERENT EVENT-BASED SURVEILLANCE VIDEO SYNOPSIS USING TRAJECTORY CLUSTERING
(2015)
Chou, Chien-Li
;
Lin, Chin-Hsien
;
Chiang, Tzu-Hsuan
;
Chen, Hua-Tsung
;
Lee, Suh-Yin
With the rapid development of the camera industry, surveillance systems become more and more popular in our daily life. However, it is very time-consuming to find out specific persons or objects from a mass of surveillance videos with long duration. For efficient browsing surveillance videos, numerous researchers are devoted to eliminating the inherent spatiotemporal redundancy for video synopsis. Nevertheless, too much information in a synopsis frame may distract viewers\' attention. Therefore, we propose a novel surveillance video synopsis system using coherent event classification to alleviate the above issues. Object trajectories are extracted by background subtraction, and then clustered. Coherent events containing similar actions of objects with different moving speeds are obtained by applying the longest common subsequence algorithm to measure the similarity among trajectories. The trajectories in each cluster are rescheduled and stitched onto the background to generate synopsis videos with coherent events. Comprehensive experiments conducted on various surveillance videos demonstrate the convincing performance of our proposed system.
Development of method for quantifying essential tremor using a small optical device
(2016-06-15)
Chen, Kai-Hsiang
;
Lin, Po-Chieh
;
Chen, Yu-Jung
;
Yang, Bing-Shiang
;
Lin, Chin-Hsien
Background: Clinical assessment scales are the most common means used by physicians to assess tremor severity. Some scientific tools that may be able to replace these scales to objectively assess the severity, such as accelerometers, digital tablets, electromyography (EMG) measurement devices, and motion capture cameras, are currently available. However, most of the operational modes of these tools are relatively complex or are only able to capture part of the clinical information; furthermore, using these tools is sometimes time consuming. Currently, there is no tool available for automatically quantifying tremor severity in clinical environments. New method: We aimed to develop a rapid, objective, and quantitative system for measuring the severity of finger tremor using a small portable optical device (Leap Motion). Results: A single test took 15 s to conduct, and three algorithms were proposed to quantify the severity of finger tremor. The system was tested with four patients diagnosed with essential tremor. Comparison with existing method: The proposed algorithms were able to quantify different characteristics of tremor in clinical environments, and could be used as references for future clinical assessments. Conclusions: A portable, easy-to-use, small -sized, and noncontact device (Leap Motion) was used to clinically detect and record finger movement, and three algorithms were proposed to describe tremor amplitudes. (C) 2016 Elsevier B.V. All rights reserved.