PROBLEM TO BE SOLVED: To improve emotion recognition accuracy in an emotion recognition device for recognizing human emotions. SOLUTION: The authentication device 1 uses an image feature amount calculation unit that calculates a first emotion feature amount from image information of a subject captured by a camera 32, and voice information obtained by collecting a voice produced by this subject with a microphone 31. A voice feature amount calculation unit that calculates a second emotion feature amount, and a recognition unit that recognizes the emotion of the subject based on the emotion feature amount that is a combination of the first emotion feature amount and the second emotion feature amount are provided. ing. [Selection diagram] Figure 1【課題】人の感情を認識する感情認識装置において、感情の認識精度を向上する。【解決手段】認証装置1は、カメラ32が撮影した被写体の画像情報から第1の感情特徴量を算出する画像特徴量算出部と、この被写体が発する音声をマイク31で集音した音声情報から第2の感情特徴量を算出する音声特徴量算出部と、第1の感情特徴量と第2の感情特徴量を合成した感情特徴量に基づき、この被写体の感情を認識する認識手段とを備えている。【選択図】図1