深度学习与人脸识别方法研究

A Study of Deep Learning and Human Face Recognition Methods

传统的机器学习方法需要对每个领域都获取大量的训练数据,这样就会在研究中耗费大量的人力物力。深度学习可以更好地通过增加数据集的规模来改善学习结果。深度学习更适合于未标记数据,而这超出了自然语言处理的范畴,后者更多限于实体识别。基于深度学习的优点,这篇文章利用深度学习方法来进行人脸识别,提出了构建深度学习网络的方法,它能够识别训练集中没有身份的表情信息。

Traditional machine learning method needs to obtain a large amount of training data for each field, so it will cost a lot of manpower and material resources in the research. Deep learning can better improve learning results by increasing the size of data sets. Deep learning is more suitable for unmarked data, which goes beyond natural language processing, and the latter is more limited to entity recognition. Based on the advantages of deep learning, this paper applies the deep learning method to face recogni-tion, and proposes a method of constructing deep learning network, which can recognize information of facial expression without identi-ty in the training set.