基于卷积神经网络的毕加索绘画艺术风格分类研究

Research on the Classification of Picasso’s Painting Art Style Based on Convolutional Neural Network

毕加索一生共有6万多幅作品,作品风格多样,依靠人工对毕加索不同风格的作品进行分类,将会是非常耗时的工作。为此论文研究并提出一种深度学习分类算法,实现毕加索绘画的艺术风格分类。首先论文使用RGB(Red,Green,Blue)和HSV(Hue,Saturation,Value)两种不同色彩模型对绘画作品进行数据处理,接着通过卷积神经网络模型实现对绘画样本风格特征的提取,最后使用支持向量机对分类结果进行融合与分类。结果表明在绘画艺术风格识别的准确度上使用HSV色彩模型,其表现优于RGB色彩模型,表明HSV色彩模型包含更多特征信息,能有助于提高分类准确率。

Picasso has more than 60,000 works in his lifetime,with various styles.It would be a very time-consuming task to manually classify Picasso’s works.Therefore,in this paper,a deep learning classification algorithm to classify Picasso paintings by artistic style is studied and proposed.Firstly,two different color models of RGB(Red,Green,Blue)model and HSV(Hue,Saturation,Value)model are used to present images.And then the style characteristics of the paintings are described through the convolutional neural network model.Finally,a SVM unit is adopted to achieve the final fusion of classifications.The results show that the HSV color model is better than the RGB color model in the accuracy of painting art style recognition.It shows that the HSV color model contains more feature information,which can effectively improve the classification accuracy.