An integrated circuit for epilepsy detection based on sparse extreme learning machine algorithm

Abstract

The invention discloses an epilepsy detection integrated circuit based on a sparse extreme learning machine algorithm, which uses a window containing 256 points to divide the signal in the EEG data, and the data type is a 16-bit fixed-point number, including 8 bits of the integer part and decimals The part is 8 bits, the EEG data of the known category is generated through the wavelet transform circuit module to generate a four-dimensional signal, and then the generated signal is input into the feature extraction circuit module to obtain an eight-dimensional feature vector, and then the eight-dimensional feature vector is input into the classifier circuit module after training Classification output: the EEG signals of unknown categories are sequentially passed through the wavelet transform circuit module and the feature extraction circuit module to obtain an eight-dimensional feature vector, and the eight-dimensional feature vector is input into the classifier circuit module for classification and output. The invention uses the lifting wavelet transform to process the EEG signal, and can obtain the characteristics of different frequency bands in the time domain and frequency domain of the signal, and the result is superior to the traditional filter and Fourier transform.

Description

technical field [0001] The invention belongs to the technical field of integrated circuits, and in particular relates to an epilepsy detection integrated circuit based on a sparse extreme learning machine algorithm. Background technique [0002] Epilepsy is caused by abnormal discharge of brain neurons, so epileptic seizures can have a great impact on people's EEG signals. Due to its non-invasive and convenient features, EEG examination plays an important role in the diagnosis of epileptic seizures, and it also helps to classify the types of epileptic seizures. All cases of clinically suspected epilepsy are diagnosed by EEG examination. diagnostic. At present, epilepsy detection through EEG still mainly relies on doctors to manually analyze the patient's EEG signals. This method is not only time-consuming, but also has no unified basis for judgment. its accuracy. [0003] Due to the many shortcomings of manual analysis of signals, coupled with the continuous development o...

AI Technical Summary

Problems solved by technology

[0004] After searching the existing literature, it is found that there is no report on the implementation of integrated circuits using lifting wavelet transform to process EEG signals, extract features, and then use sparse extreme learning machines for training and classification.

Images