论文部分内容阅读
视频分析是当前最令人振奋的热点技术之一。对于安全和监控领域的应用来说,如果能够应用人工智能技术来“观察”视频,提取有用信息,并进行报警预警,那前景就非常美好了。实际上,智能技术能够表现得如此先进和神奇,以至让人以为是科幻小说。现在,有关这项技术引起的振奋已经使得科幻和现实的分割线不明显了。比如说,有人会以为视频分析技术能够在一个满是人的体育场馆里发现某个恐怖主义分子,这可是连人眼都无法完成的任务。或者说,它能够自动“看出”某个人在赌桌上出老千,而经过高度训练的监察人员都难以做得到。这就是对视频分析技术过于神化了。类似这种神化或误解还有很多,那就让我们来谈谈对于视频分析技术及其实际应用的5个最常见的误解,告诉用户其真实的情况。
Video analytics is one of the most exciting hot technologies of the moment. For applications in the security and surveillance arena, the promise of artificial intelligence to “watch” video, extract useful information, and alert early warnings is promising. In fact, intelligent technology can behave so advanced and magical, so that people think it is a science fiction. Now, the excitement about this technology has made the line between science fiction and reality less clear. For example, someone might think that video analytics can find a terrorist in a man-made stadium, something that even human eyes can not do. Or, it can automatically “see” that someone is an aging poker table, and highly trained supervisors are hard pressed to do. This is a deification of video analytics. Similar to this deification or misunderstanding there are many, let us talk about the video analysis technology and its practical application of the five most common misconceptions, tell the user the true situation.