14 DAY TRIAL // Guessing the state of mind of a human can be a tricky thing to do for the untrained eye, although most psychologists brag about their ability to do it, based on body (or facial, in this case) language. However, their results are not entirely accurate and most definitely not very fast either. They could never be employed, say, in an airport, in order to detect possible suspects of committing terrorist acts. A new piece of software developed by an expert of Indian origin can, though.
Dr. Prabir Bhattacharya, in collaboration with Abu Sayeed Sohail, a graduate student from the University of Concordia in Montreal, Canada, has managed to develop a piece of computer software that is able to determine a person's mood just by analyzing a scanned picture of their face. The system doesn't even require the whole face to be present in the photograph, since it can detect and classify a number of facial expressions by focusing on just 15 key points on a person's physiognomy.
The points involve certain facial muscles found under the skin and located in the close proximity of the person's eyes, nose and mouth. After assessing the data and values provided by the key points, the software quickly processes and compares them to the identifiable expression key point values in its database. Although sometimes the actual faces and expressions are very different from each other, due to individual, health, race or cultural factors, the couple of researchers have been able to depict 7 basic ones that seem to be universally valid and unchangeable by these criteria.
Since the newly-developed system is capable of impressive performances in taking random photos of large crowds, processing them in a short time, and helping operators detect some potentially dangerous individuals, it is prone to be adopted for large use in agglomerated transit places such as airports, where security is a main issue. Further details on the success, issues, and implementation of the software are yet to come.