WiMi Releases A Highly Reliable Silent Speech Recognition System Based on Multi-featured Signal Sensing Technology

BEIJING, April 25, 2023

Voice recognition technology is the hands-free interface mode for VR applications but it has several drawbacks. For example, it has limited availability in noisy environments or public places and for those who cannot speak loudly and clearly. Therefore, WiMi developed this system to enable silent speech recognition by using limb movements or facial electromyography in a holographic AR environment. First, the system acquires holographic EMG signal data or vocal fold vibration signal data presented by a person's face or limbs and then preprocesses the two data types obtained separately. After feature extraction and fusion, the system uses deep learning to perform recognition sequentially and finally sends the recognized command results to the receiving device or the controlled device.

To improve the system's accuracy for signal classification at a distance, WiMi developed a deep neural network-based classification method using an SSR system with fEMG in a holographic environment. The technique uses similar fEMG data previously collected from other individuals and then transformed by holographic dynamic position distortion. When a person is in a state of speech or silent thinking, the vocalization-related limb muscles are affected by the brain's output nerves, which produce different activity states corresponding to the other contents to be represented by the brain. The system performs voiced or unvoiced speech recognition and acquires brain information by collecting surface EMG signals from limb or vocal muscles, processing and recognizing them. For feature recognition, the system uses independent CNNs to learn the features of each channel signal. The developers designed three structures: a 1D convolutional network, a 2D convolutional network, and a parallel convolutional network, configured the network model parameters, and optimized the network model structure. The developers used Support Vector Machine, Random Forest, k-Nearest Neighbor, and Hidden Markov Model for machine learning. The model parameters are continuously optimized during training and recognition.

WiMi's system combines holographic visual information and facial or limb EMG information to achieve silent speech recognition. WiMi expects that this system will contribute to the further development of speech recognition technology and provide new ideas and methods for speech recognition.

About WIMI Hologram Cloud

WIMI Hologram Cloud, Inc. (NASDAQ:WIMI) is a holographic cloud comprehensive technical solution provider that focuses on professional areas including holographic AR automotive HUD software, 3D holographic pulse LiDAR, head-mounted light field holographic equipment, holographic semiconductor, holographic cloud software, holographic car navigation and others. Its services and holographic AR technologies include holographic AR automotive application, 3D holographic pulse LiDAR technology, holographic vision semiconductor technology, holographic software development, holographic AR advertising technology, holographic AR entertainment technology, holographic ARSDK payment, interactive holographic communication and other holographic AR technologies.

Safe Harbor Statements