基于深度学习的手势识别算法实现.doc

内容摘要 基于深度学习的手势识别处理计算机视觉领域中的关键技术之一， 深度学习技术在近20年得到了快速的发展，其中在2014年就提出了VGG和GoogleNet这两大高精度高准确率的网络模型。 手势识别是在图像中捕抓到手部关键信息再通过手势给机器传达信息。手势识别大多数情况下都是在实时视频作为应用环境，这就对模型有着实时处理和准确性这两个基本要求。 随着科技的发展和5G时代的到来，身处于深度学习领域的手势识别技术越来越完善，在技术层面上相对现在一定会有了很大的进步，但同时也面临更加深层的困难，例如各个种群的人类肤色、背景复杂度、光照条件、多人情况下的手部重叠这类打击性很大的干预因素，是每个开发手势识别技术人员都会面临的难题。 本次论文将以近年来较为流行的深度学习领域中的卷积神经网络等等神经算法展开，将获取到的手部信息关键点给予逻辑处理得到准确无误的信息。 关键词：手势识别 深度学习 卷积神经网络 手部关键信息点 Abstract One of the key technologies in the field of computer vision in gesture recognition processing based on deep learning, Deep learning technology has developed rapidly in the past 20 years. Among them, the two high-precision and high-accuracy network models of VGG and GoogleNet were proposed in 2014. Gesture recognition is to capture the key information of the hand in the image and then convey the information to the machine through the gesture. Gesture recognition is in most cases using real-time video as the application environment, which has two basic requirements for real-time processing and accuracy of the model. With the development of technology and the advent of the 5G era, the gesture recognition technology in the field of deep learning is becoming more and more perfect. There must be a lot of progress on the technical level compared with the present, but it also faces deeper difficulties, such as The striking intervention factors such as human skin color, background complexity, lighting conditions, and hand overlap in the case of multiple people are challenges faced by everyone who develops gesture recognition technology. This paper will be developed with neural algorithms such as convolutional neural networks, which are more popular in the field of deep learning in recent years, and the key points of the hand information obtained will be logically processed to obtain accurate information. ?Key words: gesture recognition deep learning convolutional neural network key information points of hand 目 录 TOC \o "1-3" \h \z \u 第一章 绪论 1 1.1 技术