基于深度层级特征的中国传统视觉文化符号识别及应用

中国传统视觉文化符号凝结着中国人的智慧和力量,成为世界解读能够诠释中国的一种方式,它能代表中国、影响世界.本论文主要针对中国传统视觉文化符号做识别和分类等智能化处理,目标是探讨一种关于中国传统视觉文化符号智能识别的解决方案.我们先从简单的分类任务入手,利用浅层学习中多特征融合的方法来解决分类问题,再通过使用深度学习中卷积神经网络来处理这一问题.最后,我们提出了基于深度层级特征和SVM 结合的中国传统视觉文化符号识别的方法,在此基础上,我们在特定的博物馆中对文化展品进行识别,并做成基于线上线下的识别APP.

Chinese Traditional Visual Cultural Symbols Recognition and Application Based on the Depth Hierarchy Features

The traditional visual culture symbol of China condenses the wisdom and strength of the Chinese people and becomes a way for the world to interpret China. It can represent China and affect the world. This dissertation focuses on the intelligent processing of Chinese traditional visual cultural symbols, such as recognition and classification, and aims to explore a solution to the intelligent identification of Chinese traditional visual cultural symbols. We start with a simple classification task, using multi-feature fusion in shallow learning to solve classification problems. Then, we deal with this problem by using convolutional neural networks in deep learning. Finally, we propose a method of Chinese traditional visual cultural symbol recognition based on the combination of deep level features and SVM. Based on this, we identify cultural exhibits in specific museums and make a online and offline APP.