Development and Validation of an EEG-Based Real-Time Emotion Recognition System Using Edge AI Computing Platform With Convolutional Neural Network System-on-Chip Design

Fang, Wai-Chi; Wang, Kai-Yen; Fahier, Nicolas; Ho, Yun-Lung; Huang, Yuan Ming

doi:10.1109/jetcas.2019.2951232

Cited by 61 publications

(49 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In Ref. [18], CNN is applied to recognize emotion. The feature extraction engine formats the six collected raw EEG data from specified channels into one EEG image.…”

Section: Feature Extractionmentioning

confidence: 99%

“…In Ref. [18], the tailored CNN chip was implemented in TSMC 28-nm CMOS technology and was tested using the Agilent 93000 SoC platform. In the paper, the authors have proposed "a multiphase CNN execution method to accommodate hardware resource constraints.…”

Section: Emotion Recognition System Chip Using Cnnmentioning

confidence: 99%

“…Figure 19 in Ref. [18] shows the die photo of the chip, and the Table II in Ref. [18] gives the related summary.…”

Section: Emotion Recognition System Chip Using Cnnmentioning

confidence: 99%

“…4 in Ref. [18]. The results of the SE unit were used to calculate two groups of weight values to carry out optimized channel reconstruction weighting based on whether the average SE value of the frontal channels was less than that of prefrontal channels.…”

Section: Feature Extractionmentioning

confidence: 99%

“…Table III in Ref. [18] shows the comparison. The average accuracy of the binary classification and the quaternary classification derived from the multidimensional model in Ref.…”

Section: Emotion Recognition System Chip Using Cnnmentioning

confidence: 99%

See 4 more Smart Citations

An introduction and review on innovative silicon implementations of implantable/scalp EEG chips for data acquisition, seizure/behavior detection, and brain stimulation

Shi

Zhang

et al. 2020

Brain Science Advances

View full text Add to dashboard Cite

Technological advances in the semiconductor industry and the increasing demand and development of wearable medical systems have enabled the development of dedicated chips for complex electroencephalogram (EEG) signal processing with smart functions and artificial intelligence‐based detections/classifications. Around 10 million transistors are integrated into a 1 mm2 silicon wafer surface in the dedicated chip, making wearable EEG systems a powerful dedicated processor instead of a wireless raw data transceiver. The reduction of amplifiers and analog‐digital converters on the silicon surface makes it possible to place the analog front‐end circuits within a tiny packaged chip; therefore, enabling high‐count EEG acquisition channels. This article introduces and reviews the state‐of‐the‐art dedicated chip designs for EEG processing, particularly for wearable systems. Furthermore, the analog circuits and digital platforms are included, and the technical details of circuit topology and logic architecture are presented in detail.

show abstract

“…In Ref. [18], CNN is applied to recognize emotion. The feature extraction engine formats the six collected raw EEG data from specified channels into one EEG image.…”

Section: Feature Extractionmentioning

confidence: 99%

Section: Emotion Recognition System Chip Using Cnnmentioning

confidence: 99%

“…Figure 19 in Ref. [18] shows the die photo of the chip, and the Table II in Ref. [18] gives the related summary.…”

Section: Emotion Recognition System Chip Using Cnnmentioning

confidence: 99%

Section: Feature Extractionmentioning

confidence: 99%

“…Table III in Ref. [18] shows the comparison. The average accuracy of the binary classification and the quaternary classification derived from the multidimensional model in Ref.…”

Section: Emotion Recognition System Chip Using Cnnmentioning

confidence: 99%

See 3 more Smart Citations