基于心理测试图像模板的图像情感识别

Affective-Cognitive of Images Based on Models of Psychology Pictures

以著名心理学家弗海姆的形状-颜色测试表中的9幅不同特征的图像为模板,提出了一种新的图像情感识别的方法。在HSV空间中,用区间量化方法将颜色特征量化为48柄的矢量。再采取矢量结合Zernike矩的方法,用二次距衡量图像之间的相似性。最终结合测试者的个性特征和测试图像与特征模板的比较结果,映射出测试者对测试图像的喜恶指数。

we propose a new method of affective cognitive of images. We use 9 pictures that constitute the famous psychologist Furnham＇s Shape and Color Test as standard templates. In the HSV color space, we get a vector which has 48 dimensions by using part-division method on color character. And then we compute Quadratic Distance of Vector and Zernike moment to compute comparahilities of images. Combined the personality of Conner with the results of comparability of test pictures and template pictures, we get the like-unlike exponent of Conner towards test pictures.