Application of mydriasis identification methods in parental control systems
Olesia Barkovska, Natalia Axak, Dmytro Rosinskiy, Serhii Liashenko · 2018 IEEE 9th International Conference on Dependable Systems, Services and Technologies (DESSERT) · 2018
The article is devoted to solving the problem of secure Internet surfing. The authors proposed the control model of data stream transmitted over a network based on network driver interface specification. The emphasis was made on the development of a complex parental control system structure, which includes several function blocks for the in-depth analysis of text and graphic site content, as well as special block for mydriasis identification. This approach enables to track the impact of irritators on the child and send a signal with the up-to-the-minute desktop screenshot to the parent device. The tests show accurate eye detection (83%) and accurate mydriasis detection (85%) out of the total amount of accurately detected eyes. Pupillometry speed analysis showed that CUDA applications for massively parallel processing enabled to obtain the result approximated to the real-time mode - in 0.033 sec.