Abstract:

Abstract—With the rapid advancement of deep learning, convolutional neural networks (CNNs) have been widely applied to electromyography (EMG)-based motion recognition tasks. However, conventional CNN models often suffer from high computational complexity and large memory overhead, making them unsuitable for direct deployment on resource-constrained wearable devices or edge computing platforms. To address this challenge, this paper proposes a lightweight CNN model based on mixed-precision quantization. By assigning different bit-widths (2/4/8 bits) to each network layer according to its sensitivity to precision, the proposed method significantly reduces both model size and inference complexity while maintaining classification performance. Specifically, we design a CNN model targeting five-class EMG gesture recognition based on the Ninapro DB1 dataset. The original model achieves an accuracy of 92.68% under FP32 precision, with a model size of 252.4 KB. By applying a layer-wise mixed-precision quantization strategy combining post-training quantization (PTQ) and quantization-aware training (QAT), the model is compressed to 36.58 KB (6.9× compression) with only a 3.1% accuracy drop. Experimental results demonstrate that the proposed method achieves an excellent trade-off between accuracy and efficiency, showing strong potential for embedded deployment in low-power, low-storage real-time applications.