Methods and equipment for detecting and visualizing level of attention in videos


With the spread of video delivery systems such as YouTube, the necessity of methods for transmitting information from the viewer in real time, and the collection of user information for marketing is mounting. The information obtained by the server at present is limited to the number of accesses and PC identification numbers (IP address, MAC address, etc.), as well as that provided by the viewer. It is difficult, for example, to determine whether or not the viewer is actually in front of the screen or what aspects of the video they are interested or concerned with.


This invention involves a method for quantitating details on the viewer during playback, such as “what parts does the viewer tend to look at” and their “level of attention” (existing eye movement measuring devices are also used), and methods, equipment and systems for visualizing time-series statistical probability distribution information, then consecutively updating and delivering accordingly.


●Quantification of the effects of advertising and publicity in videos
●It can be adapted to use in game systems, robots, and sport lesson content


With the present invention, as well as information on the direction of sight, by measuring pupil diameter and assigning a weight, time-series statistical probability distributions can be determined. Because low-cost infrared emitters and filters have been added to eye movement measuring devices, and it is possible to adapt the CMOS cameras already installed in notebook PCs and mobile terminals widely available on the market, it is easy to introduce the present system into the viewing environment. 

purpose of providing seeds

Sponsord research, Collaboration research, Technical consultation


posted: 2014/05/21