Lie Detection Using Nonlinear Psychological Signals

Purpose: One of the issues that many judicial and security authorities are concerned about is the speech accuracy confirmation. All of the designed lie detection systems have not been able to confirm the speech accuracy completely. Methods: These methods include sound stress measurement, thermal imaging, and deception detection which use blinking. The drawback of the above methods is that they are costly, and difficult to registration, analysis and interpretation. Results: According to the results of various experiments on polygraph data, Photo Plethysmography signals and skin electrical resistance are mostly depend on mental states, especially stress. The Improving results is achieved by having a more complete statistical population, which eliminates the effect of interventional factors such as excessive stress. Conclusions: Due to previous experiences, the use of cognitive systems with fuzzy logic and using of dynamic neural networks is recommended to improve the accuracy of diagnosis. Also, the variations of these signals have the most weight in the general conclusion of the parameters changes. In this paper, lie detection data recorded at the Intelligent Signal Processing Research Institute were used to show that the ELMAN classifier has a proper percentage of accuracy.

کلیدواژه ها:

Lie detector systems ، Normalization ، Plethysmography signal ، Skin electrical resistance signal ، linear separator