BS-5-36 Evolution of Eye Movement Classification for Interactive Streaming System(BS-5. Network and Service Design, Control and Management)

スポンサーリンク

概要

• 論文の詳細を見る

• Eye gaze is now used as a content adaptation trigger in interactive media applications, such as customized advertisement in video, and bit allocation in streaming video based on region-of-interest(ROI), which could be broadly used for interactive video streaming system. However, the process of eye movement classification - separating and labeling three major types of eye movements: fixation, smooth pursuit and saccades in eye gaze trajectory - is an essential part of eye-movement data analysis and can have a significant impact on higher-level implement. In this paper, we propose an Evolved Hidden Markov Model (E-HMM) to estimate the eye movement state. Using the human key press as input to collect the "ground truth" eye movement states, and then we improved it using the well-known Forward-Backward algorithm based on initial the HMM parameters. At last, we also compared the E-HMM parameters calculated by the saliency analysis to verify the modified results.

• 2012-08-28

著者

• FENG YUNLONG

  総研大

• CHEUNG GENE

  NII

• JI YUSHENG

  NII

関連論文

• BS-5-36 Evolution of Eye Movement Classification for Interactive Streaming System(BS-5. Network and Service Design, Control and Management)

スポンサーリンク