| 摘要: |
| 借助于二值图像的可重叠矩形区域编码的思想,通过使用可重叠矩形非对称逆布局的模式表示模型(RNAM)和扩展的Gouraud 阴影法,给出了可重叠同类块逆布局的4 个准则,提出了一种基于可重叠RNAM 的灰度图像表示算法,简称为ORNAMC 表示算法.在ORNAMC 表示算法中,通过使用3 个用于标识顶点类型的水平矩阵H、垂直矩阵V 和单点矩阵I 代替混合矩阵R,解决了灰度图像的可重叠RNAM 表示中矩阵R 的不可解码性问题;同时,通过将顶点类型及码字进行重新定义,提出了一种对矩阵H,V 和I 中所有非零元素坐标进行编码的坐标数据压缩算法.以图像处理领域里惯用的标准灰度图像等作为典型测试对象,实验结果表明,与已提出的非重叠RNAMC和流行的STC,SDCT 等灰度图像表示方法相比,在保持图像质量的前提下,ORNAMC 表示方法具有更高的压缩比和更少的块数,因而是灰度图像表示的一种更好的方法. |
| 关键词: 图像表示 矩形非对称逆布局的模式表示模型(RNAM) 可重叠 灰度图像 扩展的Gouraud 阴影法 SDCT STC 坐标数据压缩 同类块 |
| DOI：10.3724/SP.J.1001.2012.04236 |
| 分类号: |
| 基金项目:国家自然科学基金(60973085); 国家高技术研究发展计划(863)(2006AA04Z211); 广东省自然科学基金(S2011040005815); 广东高校优秀青年创新人才培养计划(LYM11015); 教育部博士点基金(20120172120036); 中央高校基本科研业务费专项资金(2011ZM0074); 国家级大学生创新训练计划(111056154) |
|
| Gray Image Representation Algorithm Based on Overlapping RNAM |
|
ZHENG Yun-Ping1, CHEN Chuan-Bo2, LI Zu-Jia1
|
|
1.School of Computer Science and Engineering, South China University of Technology, Guangzhou 510006, China;2.School of Software, Huazhong University of Science and Technology, Wuhan 430074, China
|
| Abstract: |
| The idea of an overlapping rectangular region coding of binary images inspired the overlapping rectangular non-symmetry, the anti-packing pattern representation model (RNAM), and the extended Gouraud shading approach. A novel lossy gray image representation algorithm based on the overlapping RNAM, which is called ORNAMC representation algorithm, is proposed. Also, the four principles for anti-packing the overlapping homogenous blocks are presented in this paper. In the proposed ORNAMC representation algorithm, the wrong decoding problem of the matrix R for the overlapping RNAM representation of gray images is solved separately by using the horizontal, vertical, and isolated matrices, i.e.,H, V and I. These are used to identify the vertex types instead of using a single hybrid matrix, i.e., R. In addition, by redefining the codeword set for the three vertices symbols, this paper proposed a new coordinate data compression algorithm for coding the coordinates of all non-zone elements in the three matrices H, V and I. By taking some idiomatic standard gray images in the field of image processing as typical test objects, and by comparing the proposed ORNAMC representation algorithm with the latest non-overlapping RNAMC, the experimental results show that the former has a higher compression ratio and a fewer number of blocks and yet, maintains image quality. Therefore, a better method is to represent the gray image. |
| Key words: image representation rectangular non-symmetry and anti-packing pattern representation model (RNAM) overlapping gray image extended Gouraud shading approach SDCT STC coordinate data compression homogenous block |
