基于无线传感技术与卷积神经网络的静态手语识别方法

Static Sign Language Recognition Method Based on Wireless Sensor Technology and Convolutional Neural Network

基于传感网络的手语识别随着手势复杂程度的提升,识别的难度增大,因此,为了获取更加精准的静态手语识别结果,提出了基于无线传感技术与卷积神经网络的静态手语识别方法。通过无线传感技术对静态手语图像进行滤波处理,采用特征提取方法提取静态手语特征并进行融合处理,将融合后的连续二维特征图像直接转换为一维向量,将其作为卷积神经网络的输入,通过迭代训练完成静态手语识别。实验结果表明,所提方法在不同场景下静态手语识别精度高于94.11%,且识别速度较快,由此验证了所提方法可以快速准确完成静态手语识别。

The complexity of gestures increases the difficulty of sign language recognition.Therefore,in order to obtain more accurate static sign language recognition results,a static sign language recognition method based on wireless sensor technology and convolutional neural network is proposed.The static sign language image is filtered by using wireless sensor technology.The static sign language fea⁃tures are extracted by using feature extraction method and fused.The fused continuous two⁃dimensional feature image is direct trans⁃formed into one⁃dimensional vector,which is used as the input of convolutional neural network,and the static sign language recognition is completed through iterative training.The experimental results show that the accuracy of the proposed method is higher than 94.11%in different scenes,and the recognition speed is fast.Therefore,it is verified that the proposed method can complete the static sign language recognition quickly and accurately.