Loading...
Please wait, while we are loading the content...
Similar Documents
Project Identification and Tracking of Players in Sports Videos
| Content Provider | Semantic Scholar |
|---|---|
| Author | Lu, Chun-Wei Liao, Hong-Yuan Mark |
| Copyright Year | 2014 |
| Abstract | To develop an automatic basketball video analysis system is difficult for several key reasons. First, collisions between players occur frequently. Second, players on the same team (and thus in the same uniform) can look similar. And, third, players frequently per form complicated motions and may move outside the field of view of the camera . Most existing basketball video analysis systems can recognize players only in fairly close-up shots in which facial features are clear or jersey numbers are easily legible. Those criteria, however, are not commonly met in recordings of basketball games. Usually, videos of basketball games are acquired from a single pan-tiltzoom camcorder at court level. Under these circumstances, faces of players are usually indistinguishable and jersey numbers can also be difficult to recognize. Moreover, since players change their poses and orientations frequently, both face and jersey number are recognizable only in limited cases. Last but not least, since players of the same team wear identically colored jerseys and in many cases may even have similar skin and hair colors, retrieving information from colors is challenging . |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://www.iis.sinica.edu.tw/page/aboutus/FILE/2014iisupdate-E-project2.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
