Szu-Chi Huang

Several motor imagery classification methods have been developed and achieve higher accuracy. Machine learning (ML) based algorithms utilizing manually designed features often encounter robustness issues, leading to diminished accuracy. While deep learning (DL) based algorithms exhibit promising accuracy, their extensive computational requirements present challenges in implementing them on portable devices, thereby restricting their practical applications. In this paper, we improve the ML-based algorithm’s feature robustness problems by combining common spatial patterns with Riemannian tangent space mapping, enhancing the algorithm’s feature quality. Furthermore, we introduce a method that utilizes the distance between data points and the SVM hyperplane to compute category scores, thereby enhancing classifier performance. Our experiment uses the BCI Competition IV 2A, BCI Competition III 3A, and a self-recorded dataset for subject-specific experiments to validate the algorithm’s classification performance. Experimental results show that the proposed algorithm achieves the best classification performance, with an accuracy of 78.55%, 83.33%, and 57.44% for BCI Competition IV 2A, BCI Competition III 3A, and the self-recorded dataset. Additionally, to assess the practicality of a real-time portable application, we implemented the proposed algorithm on Raspberry Pi and Jetson Nano, measuring their computation time and peak memory usage. The results demonstrate that our algorithm necessitates only 0.08 to 0.3 seconds of computation time and employs a mere 15MB of memory.

Many deep-learning-based seizure detection algorithms have achieved good classification, which usually outperformed traditional machine-learning-based algorithms. However, the hand-engineered features increase the computational complexity and potentially have an ineffectiveness problem for the category. Therefore, this paper proposes a novel end-to-end deep-learning model comprising an inception module and a residual module to analyze the multi-scales of original EEG signals and realize seizure detection without feature extraction. Experiments were conducted and evaluated on the Bonn dataset and the CHB-MIT dataset. In the subject-dependent experiments, our model achieved an average F1-score of 69.34% on the CHB-MIT dataset. In subject-independent experiments, our method achieved an average accuracy of 99.04% on the Bonn dataset and an average F1-score of 37.31% on the CHB-MIT dataset. A series of analyses confirmed that our proposed model has better classification performance and lower computational complexity than existing end-to-end seizure detection models.

In recent years, many studies have proposed epilepsy detection algorithms, but most of them require high computing resources and a large amount of memory, which are difficult to implement in wearable devices. This paper proposes an epilepsy detection algorithm that uses a small number of features to reduce the memory requirements of the algorithm. During feature extraction, this paper proposes an entropy estimation method that uses bitwise operations instead of logarithmic operations to reduce the algorithm’s demand for computing resources. The experimental results show that the computing time can be reduced by about 81.58%. The epilepsy detection algorithm in this paper is implemented in an ultra-low power embedded system and performs 7 classification tasks in the Bonn data set to verify the performance of the algorithm. The average classification performance is: Accuracy (97.13%), Specificity (97.57%) and Sensitivity (98.42%). Compared with previous studies, the algorithm of this paper has comparable classification performance, but the proposed algorithm only needs 0.23 seconds from feature extraction to classification result, to the best of our knowledge, which is the epilepsy detection algorithm with the least computing time currently applied to wearable devices.