Detecting Voice Features for Criminal Case

Abstract

Emotion detection research area is very popular in medical informatics and criminal technology. The voice is the one of the ways to detect the human inner emotions such as general feelings or psychiatric feelings. Humans has been expressed their emotions by varying different vocal features during speech generation. Voice and speech analysis help to become secure communications among people and human cannot easily lie their emotions to other. At the law court, emotion analysis can detect a culprit speak was lie or truth. The propose system is used to detect the emotions (normal, sad) of human through speech analysis by using Pitch, Mel Frequency Cepstral Coefficients (MFCC) and Short Time Energy (STE) feature extraction techniques to increase the accuracy and reliability. This system will help in court of law, when the lawyer questioning to the witness or plaintiff or defense to get the truth of the crime scene.