基于内容的情感图像获取模型

Content-Based Emotion Image Retrieval Model

随着信息技术的迅猛发展,情感信息处理已成为21世纪人工智能领域所面临的重要挑战之一。其中,情感图像获取的研究属于多学科交叉研究,本文以形容词作为处理对象,借鉴认知心理学、绘画艺术的研完成果,在图像内容的基础上,提出了一种包括公共情感和个性化情感的情感图像获取模型。往该模型中,借鉴心理学中的"维量"思想,建立情感空间;同时,抽取图像的主色调、不变矩、颜色和灰度分布,这些较容易引起情感变化的特征作为图像的视觉特征,建立图像的特征空间;采用支持向量机的方法建立图像的低层特征空间到用户的高层情感空间之间的映射,自动注释用户未曾评估的图像,实现了图像情感注释,在情感空间进行公共情感检索,快速获得用户情感信息,在此基础上,采用可视化交互式遗传算法实现因人而异的个性化情感检索,该模型应用于风景图像的情感检索,取得了较好的实验结果。

With the rapid development and popularity of the Internet and multimedia,emotion information processin ghas become a great challenge faced by artificial intelligence today. The research of emotion image retrieval is an intersectional research. Inspired from the research products of psychology and painting,a content-based emotion image retrieval model which includes common emotion and individual emotion has been proposed in this paper. First, based on the idea of 'dimension' from psychology,an emotion space is constructed. Second,dominant colors,moment invariants,color and gray compositions,which can stimulate the emotion of human very easily,have been extracted from images to construct the feature space. Then,support vector machines are used to map images from the low level feature space to the high level emotion space,and automatically annotate unevaluated images based on user's common emotion. After that, the common emotion image retrieval has been indexed in the common emotion space and quickly grasps user's subjectivity inherent in vision,while an interactive individual emotion image retrieval using visualized interactive genetic algorithm is presented to adapt to individual variation and improve accuracy of the retrieval results. Based on image content ,an emotion scenery image retrieval system has been realized. The experimental results demonstrate the effectiveness of our approach.