Equal-Average Hyperplane Partitioning Method for Vector Quantization of Image Data.pptVIP

下载本文档

3
0
约3.32千字
约 14页
2017-04-01 发布于江苏
举报
版权申诉

Equal-Average Hyperplane Partitioning Method for Vector Quantization of Image Data.ppt

1、原创力文档（book118）网站文档一经付费（服务费），不意味着购买了该文档的版权，仅供个人/单位学习、研究之用，不得用于商业用途，未经授权，严禁复制、发行、汇编、翻译或者网络传播等，侵权必究。。
2、本站所有内容均由合作方或网友上传，本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺！文档内容仅供研究参考，付费前请自行鉴别。如您付费，意味着您自己接受本站规则且自行承担风险，本站不退款、不进行额外附加服务；查看《如何避免下载的几个坑》。如果您已付费下载过本站文档，您可以点击这里二次下载。
3、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等，请点击“版权申诉”（推荐），也可以打举报电话：400-050-0827(电话支持时间：9:00-18:30)。
4、该文档为VIP文档，如果想要下载，成为VIP会员后，下载免费。
5、成为VIP后，下载本文档将扣除1次下载权益。下载后，不支持退款、换文档。如有疑问请联系我们。
6、成为VIP后，您将拥有八大权益，权益包括：VIP文档下载权益、阅读免打扰、文档格式转换、高级专利检索、专属身份标志、高级客服、多端互通、版权登记。
7、VIP文档为合作方或网友上传，每下载1次，网站将根据用户上传文档的质量评分、类型等，对文档贡献者给予高额补贴、流量扶持。如果你也想贡献VIP文档。上传文档

Equal-Average Hyperplane Partitioning Method for Vector Quantization of Image Data

Equal-Average Hyperplane Partitioning Method for Vector Quantization of Image Data L. Guan and M. Kamel Speaker: 高名揚 Adviser: 張智星 Date: 12/26/2000 Outline Introduction Motivation Equal-average hyperplane Method Algorithm Complexity analysis Experimental results Conclusion Motivation Speed up vector quantization Image process Speech process Expect that each block or vector be assigned to the codeword which is closest to it The codeword assignment problem in vector quantization is a Nearest Neighbor Searching problem Observation The x and y axes represent values of two nearby pixels, and the z is the population of such paris Observation (Continue) The colors of blocks of an image are similar Let a block be a vector of dimension k Each codeword vector (x1, x2, ……, xk) is close to the central line passing through points (0, 0, ……, 0) and (G, G, ……, G), where G is the largest gray value in the image Equal-Average Hyperplane Equal-average hyperplane is a plane which is normal to the central line Equal-Average Hyperplane Let a=(a1, a2, ……, ak) ≠ 0 be a point on central line L ai=aj=s, i,j=1,2,……,k The hyperplane H normal to L which intersects L at a is Equal-Average Hyperplane For example, k = 5, s = 2, hyperplane The mean of coordinate values of each point on H equals 2, such as (2, 2, 2, 2, 2), (4, 0, 1, 3, 2), (5, 3, -2, 5, -1), etc. Use parallel equal-average hyperplanes to partition the prototypes (codewords) and leads to a computationally efficient nearest neighbor searching Method The search can be conducted among prototypes whose s values are located between smin and smax Algorithm Let L be the list of all the prototypes, ci, sorted on their coordinate average values, sci, i=1,2,…,N, and q be the query point Algorithm Equal_Average_NNS(q,L) begin 1. Initialize distance r to be a large value and c to be nil 2. Calculate the coordinate average of q, sq 3. repeat until L is empty 3.1. Search L for ci with sci closest to sq 3.2. Calcul