基于特征融合的亚洲食物图片分类  

Feature fusion based Asian food image classification

作　　者：武若琪 徐冰 曲志坚[1] WU Ruo-qi;XU Bing;QU Zhi-jian(School of Computer Science and Technology,Shandong University of Technology,Zibo 255000,China)

机构地区：[1]山东理工大学计算机科学与技术学院,山东淄博255000

出　　处：《计算机工程与设计》2020年第12期3464-3471,共8页Computer Engineering and Design

基　　金：国家自然科学基金项目(61473179);山东省高等学校青年创新团队发展计划基金项目(2019KJN048);山东省自然科学基金项目(ZR2017LF004);山东省高等学校科技计划基金项目(J16LN20)。

摘　　要：为对亚洲食物图片进行高效分类,提出一种结合图片分割与特征融合的图片分类模型SLGC。使用图片分割技术对图片进行预处理,提取食物主体,有效缩减特征提取时间。通过对亚洲食物的特有结构进行研究,在局部与全局两个维度分别将图片的颜色特征融合到表示向量中,进一步提升特征提取效果。实验结果表明,SLGC模型能够更加全面地提取图片特征,提升分类准确率。To classify Asian food images efficiently,an image classification model SLGC that combining image segmentation and feature fusion was proposed.Image segmentation was used to preprocess the image and the food subject of the image was extracted,which effectively shortened the time for feature extraction.By studying the unique structure of Asian foods,the color features of the images were merged into the representation vectors in the local and global dimensions,respectively.The effect of feature extraction was further enhanced.Experimental results show that the SLGC model can express the intrinsic characteristics of Asian food images more comprehensively and improve the classification accuracy.

关 键 词：亚洲食物 图片分类 图片分割 特征融合 词袋模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]