On the basis of creating a 3-dimension emotional factor space of fabric image and the analysis of the relationship between the low-level features and emotion description of fabric image, the emotional factor data and low-level features data of the 60 different types of fabric image samples are trained by support vector machine (SVM). After that, the trained model is set up, the emotion description is automatically calculated and the recognition of fabric image is realized. After laboratorial tests, the method achieves to better recognition of fabric images.%在建立3维面料图像情感因子空间、分析面料图像低层特征和情感描述之间关系的基础上,通过支持向量机对60张不同种类面料图像样本的情感因子数据和低层特征数据进行训练,建立了面料图像情感语义训练模型,实现了对任一面料图像情感描述的自动计算、识别。经过实验测试,达到了较好的识别效果。
展开▼