Technical Name An integrated system of AI affective computingmultimodal physiological signal in patients with high-risk of cardiovascular disorder,the development of a home-based biofeedback intervention.
Project Operator Kaohsiung Medical University
Project Host 方偉騏、余松年
ECG, PPGEEG are adopted in the system. Combining with the physiological parametersthe mental status extracted by AI recognition model, the information can be used on biofeedback treatment. Finally, a platform with multimodal signal measuringaffective computing engine using deep learning ASIC chip are applied for portablereal-time mental treatment.
Scientific Breakthrough
(1)Adding Baseline NormalizationCNN accelerated chip to recognize different emotions between subjects. (2) Using 8 EEG channels  of frontaltemporal  to choose feature selection 
method with small operation load which are suitable for real hardware system. (3) Combined  with 28 nm  prospective cell-based processes to produce fast, low power consumptionsmall deep learning arithmetic 
Industrial Applicability
The system aims to develop an integrated system of affective computingmultimodal physiological signal measuring to detect the emotionphysiological response of high-risk patients with cardiovascular disease for disease prevention purpose. By monitoring personal mentalphysical health status at homein the hospital, user can perform the mental relieving exercise with notifying mecha
Keyword Cardiovascular disease emotion recognition biofeedbcak electroencephalography electrocardiography photoplethysmography multiple physiological signals artificial intelligence affective computing CNN model
  • Contact