基于TensorFlow实现手写数字识别

Handwritten Digit Recognition Based on TensorFlow

研究TensorFlow深度学习环境的体系框架、数据模型、运行模型与模型评价机制,以手写数字图片识别的深度学习为例,对手写数字图片的训练、测试与验证进行应用实践,得出手写数字识别模型,获得百分之九十以上的识别成功率。

This paper studies the system framework,data model,operation model and model evaluation mechanism of the TensorFlow deep leaming environment.A case study of deep learning is practiced using handwritten digital image recognition example.The training,testing and verification process of handwritten digital pictures has been implemented. The hand-written digital recognition success rate is more than 90%.