A neural network has been created by Artificial Intelligence that can differentiate between criminals and non-criminals via facial recognition.
Although it seems like a scene from Minority Report, it is not. With an accuracy of 89.5 percent, offenders have been distinguished by scientists from Shangai Jiao Tong University by means of machine-vision algorithms. This study known as ‘Automated Inference on Criminality’ is the first automated work that takes account of criminality in relation to still pictures of faces.
Latest technologies are integrated by criminologists in order to identify criminals by gathering detailed data. The method is quite simple, as per Xiaolin Wu and Xi Zhang, the scientists who ran this study. First of all, ID photos of the criminals and non-criminals were taken half and half comprising 1856 Chinese men. All of these men were without facial hair and lied between 18 to 55 years old. 90 percent of the pictures were used by the scientists to develop a convolutional neural network and the rest of them were used to test the efficiency of the informed system.
The results came out to be very unnerving. It was discovered that the neural network developed by Xiaolin Wu and Xi Zhang is capable of identifying criminals correctly with 89.5 percent accuracy. Besides, a few morphologic features that are discriminative for predicting criminality were also found. These features consist of inner corner distance of the eyes, lip curvatures, and nose-mouth angle. The findings of the study are:
“Above all, the most important discovery of this research is that criminal and non-criminal face images populate two quite distinctive manifolds. The variation among criminal faces is significantly greater than that of the non-criminal faces. The two manifolds consisting of criminal and non-criminal faces appear to be concentric, with the non-criminal manifold lying in the kernel with a smaller span, exhibiting a law of normality for faces of non-criminals.”
The following characteristics were distinguished on a criminal’s face via neural network, as per the study:
- The distance between inner corners of the eyes is 6 percent shorter.
- The curvature of the upper lip is about 23 percent larger.
- The angle between two lines drawn from the corners of the mouth to the tip of the nose is 20 percent smaller.
“We are the first to study automated face-induced inference on criminality free of any biases of subjective judgments of human observers. By extensive experiments and vigorous cross validations, we have demonstrated that via supervised machine learning, data-driven face classifiers are able to make reliable inference on criminality. Furthermore, we have discovered that a law of normality for faces of non-criminals. After controlled for race, gender, and age, the general law-abiding public has facial appearances that vary in a significantly lesser degree than criminals.”
There is no doubt that the ethical controversies come with the use of Artificial Intelligence and questions like what is normal and what is not are brought up.