Lifelog video retrieval using emotion estimation based on recognition of spontaneous facial expressions

2016 Fiscal Year Final Research Report

• Project/Area Number: 15K15993
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Multimedia database
• Research Institution: Kyoto Institute of Technology
• Principal Investigator: Nomiya Hiroki 京都工芸繊維大学, 情報工学・人間科学系, 准教授 (80533116)
• Research Collaborator: SAKAUE Shota ; MAEDA Mitsuaki
• Project Period (FY): 2015-04-01 – 2017-03-31
• Keywords: ライフログ / 映像検索 / 感情推定 / 表情認識 / 表情強度

Outline of Final Research Achievements

A video retrieval system to efficiently retrieve and display impressive scenes from lifelog videos is developed on the basis of emotion estimation. In most of impressive scenes, a person in a video often shows a variety of facial expressions. Therefore, this system retrieves impressive scenes by means of facial expression recognition and the estimation of the intensity of facial expression using facial feature points. In addition, a method to detect emotional scenes in which a certain facial expression is observed is introduced. By using them, a lifelog video retrieval system is developed. This system works on a web browser and has a user-friendly interface so that a user easily finds impressive scenes.

Free Research Field

マルチメディアデータ工学