摘要
将结构相似度引入到稀疏编码模型中,提出基于结构相似度的稀疏编码模型.基于该模型提取出图像的稀疏编码特征.实验结果表明,改进后的稀疏编码模型更好地保持了结构信息,更加符合人眼视觉系统特性.将文中提出的模型应用到特征提取中,可获得结构信息保持得更好的图像特征.
The structural similarity is introduced into sparse coding model, and a sparse coding model based on structural similarity is proposed. Then, the model is employed to extract the image sparse coding feature. The experimental results show that the improved sparse coding model is consistent with human visual system for its capacity of structural information preservation. Furthermore, compared with the standard sparse coding model, the proposed model attains the reconstructed image which preserves better structural information of the original image.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2010年第1期17-22,共6页
Pattern Recognition and Artificial Intelligence
基金
国家973计划项目(No.2007CB311004)
国家863计划项目(No.2007AA01Z132)
国家科技支撑计划项目(No.2006BAC08B06)
国家自然科学基金项目(No.60805041)资助
关键词
特征提取
稀疏编码
结构相似度
基函数
图像重构
Feature Extraction, Sparse Coding, Structural Similarity, Basis Function, Image Reconstruction
