Eye movement identification method based on Transform and LSTM hybrid network
The invention provides an eye movement recognition method based on a Transform and LSTM hybrid network, and the method comprises the steps: obtaining eye movement data, carrying out the preprocessing of the eye movement data, and obtaining the preprocessed eye movement data; constructing a hybrid ne...
Saved in:
Main Authors | , , , , , , |
---|---|
Format | Patent |
Language | Chinese English |
Published |
21.06.2024
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | The invention provides an eye movement recognition method based on a Transform and LSTM hybrid network, and the method comprises the steps: obtaining eye movement data, carrying out the preprocessing of the eye movement data, and obtaining the preprocessed eye movement data; constructing a hybrid network eye movement recognition model, training the hybrid network eye movement recognition model by using the preprocessed eye movement data, and optimizing the hybrid network eye movement recognition model according to a training result; and according to the preprocessed eye movement data, using the optimized hybrid network eye movement recognition model to obtain an eye movement recognition result. According to the method, the eye movement characteristics of the time domain and the frequency domain can be extracted, when data in different time periods are used for testing on multiple databases, the performance is obviously improved, and compared with the most advanced algorithm at present, the EER is obviously re |
---|---|
Bibliography: | Application Number: CN202410147131 |