视觉影像分析算法在美术场景中的应用

Visual Image Analysis Algorithm in Art Scenes

利用视觉算法对绘画进行解读。首先利用符号学对绘画进行分析,将绘画中的特征与符号相结合。然后,基于卷积神经网络建立视觉影像分析算法,利用卷积神经网络提取图像特征。最后结合符号学的相关研究对绘画图像中的图像特征进行识别和分类。实验结果显示,视觉影像分析算法对绘画的分类准确率为99%以上,可以准确地对绘画风格进行识别和分类。利用计算机对绘画视觉艺术的解读,对研究视觉算法对绘画特征的识别具有参考价值。

The purpose is to interpret paintings using visual algorithms.First,paintings are analyzed by semiotics,combining the features and symbols in the paintings.Second,visual image analysis algorithms are established based on convolutional neural networks(CNNs),and image features are extracted using CNNs.Finally,image features in painting images are identified and classified in conjunction with related research in semiotics.The experimental results show that the visual image analysis algorithm’s accuracy for painting classification is over 99%,identifying and classifying painting accurately,thus realizing painting interpretation by computers and providing a reference for painting feature identification by visual algorithms.