WiMi Hologram Cloud Inc., a leading global Hologram Augmented Reality Technology provider, announced that it developed explainable artificial intelligence (XAI)-based fNIRS neuroimage classification, bringing a breakthrough in the development of BCI technology. By combining the latest AI technology and BCI parsing, this system is expected to bring advances in BCI technology.
WiMi’s XAI-based fNIRS neuroimage classification system consists of several key modules that work together to process, analyze, and interpret data for accurate brain activity classification and interpretation. The system architecture is designed to improve classification accuracy and interpretability, and to ensure the accuracy and utility of the system. The system includes a data preprocessing module for filtering, denoising and normalizing the raw fNIRS data to improve the accuracy of subsequent data analysis.
WiMi’s XAI-based fNIRSfNIRS neuroimage classification system employs two key classification modules, i.e., a one-dimensional sliding-window-based convolutional neural network (CNN) and a long short-term memory (LSTM) neural network. These two modules are used to classify different types of brain activity patterns respectively, thus improving the applicability and generalization ability of the system. To address the need for interpretation of model outputs, the system introduces an interpretability module, which employs the machine learning interpretability tool SHapley Additive exPlanations (SHAP) for interpreting the outputs of CNN models. By interpreting the model input variables, the system is able to identify the features that contribute the most to the classification of a particular brain activity, helping researchers to gain insight into the association between brain activity patterns and external device control.
Also Read: Skan launches ProcessGPT™, the first suite of generative AI capabilities for process mining and…
Through these methods and techniques, the system is able to efficiently transform fNIRS data into interpretable classification results. The preprocessing of the data, the application of CNN and LSTM models, and the SHAP interpretation module together form the core of the system, enabling it to improve the accuracy of brain activity classification and provide researchers with interpretable results.
WiMi’s XAI-based fNIRSfNIRS neuroimage classification system shows good application prospects and potential. In real brain-controlled robots, prosthesis control and virtual reality scenarios, the system’s high-precision classification results provide reliable support for device control and offer new possibilities for the application of BCI technology in medical rehabilitation and virtual reality.
The research and application of WiMi‘s XAI-based fNIRSfNIRS neuroimage classification system brings new insights to the field of brain science. By parsing brain activity patterns through the interpretation module, the system reveals for researchers the association and mechanism of action between functional regions of the brain, and promotes the development of the entire field of brain science. These important results show that the XAI-based fNIRSfNIRS neuroimage classification system not only improves the classification accuracy of brain activities, but also brings new perspectives to the development and application of BCI. It is foreseeable that it will promote the development and popularization of BCI in the future, and bring a revolutionary change to the interaction between humans and machines.
SOURCE: PRNewswire
