6. Adjusting the Difficulty of Running Game with Facial Expression Recognition Technology Using Convolutional Neural Network

Title(KR)

• CNN을 이용한 표정인식 기술에 기반한 러닝게임의 난이도 조절

Title(ENG)

• Adjusting the Difficulty of Running Game with Facial Expression Recognition Technology Using Convolutional Neural Network

Keywords(KR)

• Game Fun, Game Difficulty, Convolutional Neural Network, Expression Recognition

Keywords(ENG)

• Game Fun, Game Difficulty, Convolutional Neural Network, Expression Recognition

Author

• 왕지소, 이혜문, 이원형

Abstract(KR)

이제는 모바일 마켓순위에서 많은 게임이 높은 점유율을 차지하지만 점유율을 오랫동안 유지하는 것은 쉽지 않다. 게이머를 끌어당기는 중요한 요소는 게임 재미(Game Fun)이고, 게임을 재미있게 만드는 가장 중요한 요소는 게임 난이도이다. 하지만 게임 난이도를 디자인하는 것은 매우 어려운 일이다. 본 논문은 두 개의 연속적인 컨볼루셔널 레이어를 사용한 컨볼루셔널 신경망과 SVM 분류기를이용하여 게임 시 플레이어의 얼굴 표정을 실시간으로 검출하고 판단한다. 실험 결론은 CNN을이용한 표정 시스템은 게임 play-time 및 score를 늘릴 수 있고, 게임 재미를 증진시키기에 도와준다고 증명하였다.

Abstract(ENG)

Many games nowadays have a certain share of the market. However, to maintain the market position for long is not common. The most appealing element for gamers is Game Fun. The element that can make game interesting is the game difficulty. While the game difficulty has no uniform evaluation standard until now. The proposed paper uses a continuous convolutional neural network with SVM classifier to recognize the player's expression in real time. The system infers the player's psychological activity based on different expressions and makes adjustments to the game difficulty level to meet the user's needs. And the experiment result shows that the facial expression recognition system using deep learning could increase the play-time and score of the game, and promote the fun of games.

Page 39-46