The RML emotion database contains 720 audiovisual emotional expression samples that were collected at Ryerson Multimedia Lab. Six basic human emotions are expressed: Anger, Disgust, Fear, Happiness, Sadness, Surprise. A digital video camera was used to record the samples in a quiet and bright environment, with a simple background. Our experimental subjects were provided with a list of emotional sentences and were directed to express their emotions as naturally as possible by recalling the emotional happening, which they had experienced in their lives. A total number of ten different sentences were provided for each emotional class. Examples of these sentences are: I am so happy today (Happiness); Why do you always cheat on me (Anger); I messed up the mid-term (Sadness); What? Is that true (Surprise); Please, please do not beat me (Fear); That bad fish has a disgusting smell (Disgust). However, these sentences were used as reference only. Some subjects expressed their emotions by using the same sentence structure, while others opted to use variations or different sentences according to their cultural background.
The RML emotion database is language and cultural background independennt. The video samples were collected from eight human subjects, speaking six different languages (English, Mandarin, Urdu, Punjabi, Persian, Italian). Different accents of English and Chinese were also included. The samples were recorded at a sampling rate of 22050 Hz using a single channel 16-bit digitization, and a frame rate of 30fps. Each video clip has a length of about 3-6 seconds with one emotion being expressed. All the video files are in AVI format, with a total size of 4.2GB.
The RML emotion database is suitable for audio-based, static image-based, and video-based 2D and 3D dynamic analysis and recognition. It can be distributed free of charge for research purpose and non-commercial use only. To obtain the database, please send your request and signed agreement form to zhibing.xie (at) ryerson.ca. We will reply to you with a downloading link as soon as possible.