Blog: A Software Application Recognizes Human Emotions From Conversation Analysis

A Software Application Recognizes Human Emotions From Conversation Analysis
Universidad Politecnica de Madrid (Spain) (11/02/10) Eduardo Martinez

Researchers at the Universidad Politecnica de Madrid have developed an application that can recognize human emotion from automated voice analysis. The program, based on a fuzzy logic tool called RFuzzy, analyzes a conversation and can determine whether the speaker is sad, happy, or nervous. If the emotion is unclear, the program can specify how close the speaker is to each emotion in terms of a percentage. RFuzzy also can reason with subjective concepts such as high, low, fast, and slow. The researchers say RFuzzy, which was written in Prolog, also could be used in conversation analysis and robot intelligence applications. For example, RFuzzy was used to program robots that participated in the RoboCupSoccer league. Because RFuzzy's logical mechanisms are flexible, its analysis can be interpreted based on logic rules that use measurable parameters, such as volume, position, distance from the ball, and speed.

View Full Article