R&D UNIT
Kaohsiung Medical University / Professor I-Mei Lin
National Chung Cheng University / Professor Wai-Chi Fang
National Chiao Tung University / Professor Sung-Nien Yu
Technical Introduction
The technology aims to develop an AI-based emotional detections (anger, sadness, happiness, and neutral) and multimodal physiological signals (ECG, EEG, and PPG) integrated system, and apply to patients with cardiovascular disease. Patients monitor their emotional and physical status and administer bio-neuro-feedback to improve their well-being, track disease progression, and prevent adverse prognosis.
Scientific Breakthrough
1.An improved CNN/RNN architecture was developed to perform the AI online training and inferences for physical and mental monitoring. It was implemented as an AISoC chip with the TSMC-28nm process.
2.An EEG-ECG-PPG multimodal intelligent computing platform was developed to perform a variety of system functions for emotion recognition. This system combines IoMT concept and network security technologies, integrating into the cloud and terminal environment to build a diverse healthcare field effectively.
3.A high-performance emotion recognition AI algorithm was developed. The accuracies of our proposed emotion recognition system with ECG-PPG and EEG signals achieved 87.5% and 77.68%, respectively.
Industrial Application
(1)The AI-based integrated system uses AI-algorithm on affective computing for bio-neuro-feedback intervention. The technology moves forward from hospital to home-based self-monitoring on mental and physical health and has profound potential for the home care and medical industry. (2) With this technology, we would cooperate with MedKing company to develop a validated AISOC chip and multimodal affective computing platform. The product combines IoMT and network security technologies to meet the commercial demands of high-security home-based healthcares and clinical psychotherapy.