【摘要】 从信息论讲，审美就是对审美信息进行获取、存储和加工的过程，审美信息测量是计算美学的重要研究问题。图像作为审美对象，其颜色和纹理特征是其质量的低层特征，也是审美认知的基础因素。为了实现审美信息的测量，一是通过图像熵建立图像的颜色与审美信息测量之间的关系，二是通过图像梯度建立图像的纹理与审美信息测量的关系。利用图像灰度级别计算了图像的信息熵，实验测试证明，图像熵与图像颜色相关；利用Tenengrad数学模型计算了图像的梯度，并利用Sobel卷积因子进行滤波，实验测试证明，图像梯度与图像纹理相关；根据向量矩阵级联的方法，将图像熵与图像梯度的乘积作为融合图像颜色和纹理等特征的综合审美信息测量指标，实验测试证明，该综合审美测量指标值与人的审美判断评价具有一致性。更多还原

【Abstract】 From the perspective of information theory, aesthetics is the process of obtaining, storing and processing aesthetic information. The measurement of aesthetic information is an important research problem in Computational aesthetics. Images as aesthetic objects, color and texture are low-level features of its quality, and it is the basic factor of aesthetic cognition. In order to achieve the calculations of aesthetic information, we have firstly established the relationship between the color of the image and the measurement of aesthetic information through image entropy, and then established the relationship between image texture and aesthetic information measurement by image gradient. Specifically, it is to calculate the information entropy of the image by using the image grayscale. The experimental tests prove that image entropy is related to image color. The gradient of the image is calculated by using the Tenengrad mathematical model. The Sobel convolution factor is used to be filter. Experimental tests prove that image gradients are related to image textures. According to the method of cascading vector matrices, the multiplication of image entropy and image gradient is used as a synthetic aesthetic information measurement indicator, which integrated image color and texture. Experiments prove that the value of the synthetic aesthetic measurement indicator is consistent with the evaluation of human aesthetic judgment.更多还原