Published Paper Details:

Nowadays social Interaction between people tells us about the emotional and mental state the person is currently in. Humans tend to exhibit the emotional and mental state through the body and facial expressions. To predict the emotion of a person, the facial expression is majorly considered while body movements are majorly ignored. In most of the cases, body movements/expressions also reveal the emotional and mental state the person is currently in. In Numerous researches/experiments, only facial expressions are considered to predict the emotional state the person is currently in. Humans tend to fake facial expression, while it is difficult to fake the body expression. This proposed project mainly focuses on the body movements to analyze the emotional state rather than concentrating on facial recognition. To recognition this the human activity is identified from the UCI HAR Smartphone dataset. This data has data that was recorded from 30 volunteers with the age difference of 19-48 years and with 561 measurement variables. The human activity thus identified is compared with the 2-D image and the activity is matched with the 2-D image. This process is done to make sure that the activity matches the activity that has been identified from the UCI HAR smartphone dataset. Each activity has many possible positions, these positions/postures related to the activity that has been identified in the smartphone dataset help us in identifying the emotions that a human could possess while doing the activity.