Image/video Indexing, Retrieval and Summarization Based on Eye Movement

Information retrieval is one of the most fundamental functions in this era information. There is ambiguity in the scope of interest of users, regarding image/video retrieval, since an image usually contains one or more main objects in focus, as well as other objects which are considered as ̳background‘. This ambiguity often reduces the accuracy of image-based retrieval such as query by image example. Gaze detection is a promising approach to implicitly detect the focus of interest in an image or in video data to improve the performance of image retrieval, filtering and video summarization. In this paper, image/video indexing, retrieval and summarization based on gaze detection are described.