复杂背景下的静态手势识别研究  

Static Gesture Recognition under Complex Background

作　　者：颜超 穆平安[1] YAN-chao;MU Ping-an(School of Optional Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093

出　　处：《软件导刊》2022年第8期171-176,共6页Software Guide

摘　　要：为了解决聋哑人在日常生活和工作中的沟通障碍及传统手势识别方法易受复杂环境影响的问题,提出一种基于YCbCr颜色空间和卷积神经网络相结合的静态手势识别算法。该算法首先基于YCbCr颜色空间对手部区域进行提取,然后对提取后的图片进行降维、灰度化和数据增强预处理,再利用卷积神经网络模型对预处理后的图片进行训练分类。最后经过实验验证,该算法在NUS-II和Marcel两种复杂背景手势数据集上的识别准确率分别达98.32%和98.96%。相比于传统算法和其他基于卷积神经网络的算法,识别率更高、识别效果更好。In order to solve the problems of deaf-mute′s communication barriers in daily life and work and that traditional gesture recognition methods are easily affected by complex environment,a static gesture recognition algorithm based on the combination of YCbCr color space and convolutional neural network was proposed.The algorithm firstly extracted the hand region based on YCbCr color space,Then,the extracted images are preprocessed with dimensionality reduction,gray scale and data enhancement,and the convolutional neural network model is used to train and classify the preprocessed images.Finally,experimental verification shows that the recognition accuracy of this algorithm on NUSII and Marcel complex background gesture data sets reaches 98.32%and 98.96%respectively.Compared with traditional algorithms and other algorithms based on convolutional neural networks,the recognition rate is higher and the recognition effect is better.

关 键 词：图像预处理 复杂背景 图像增强 卷积神经网络 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]