An OLED lighting interactive device that can use to perform cognitive training is proposed in this work. User needs to memorize lighting sequence of OLEDs, and repeat the sequence by touching OLED panels. After that, the system records information of training and delivers it to cloud database. Then, several classifiers are used to predict the risk of dementia to assist doctors to diagnosis user

The proposed cognitive training is based on the trial purposes of mini-mental state examination, which are attention, registration, recall and repetition. The proposed system can obtain the probability of elder with dementia by using several classifiers to predict risk of dementia and voting method. The results can be reference for doctor to diagnosis user and take suitable treatment.

The proposed system can be used in hospitals as early dementia estimation assisting tool. Due to the simple training process and easy understanding smart phone interface, family members of patient can perform cognitive training instead of medical staff which can significantly reduce the workload of staffs. Moreover, the patient can use device to take long-term training to relieve dementia.