An Adaptive Codebook Vector Quantization Algorithm for large scale volume data

摘要：

To reduce encoding time of large scale volume data with Vector Quantization method greatly, this paper proposed a GPU-based Adaptive Codebook Vector Quantization (ACVQ).The algorithm first extracted the density distribution of the original volume data, then selected an appropriate initial codebook generation algorithm.Next, raw data was divided into several groups and loaded into GPU to carry out parallel computation.Each time a group of data was loaded in, a new codebook was produced by using the characteristics of this group to expand the initial codebook and then this group was encoded with the new codebook.Experimental results demonstrated that ACVQ improved the encoding speed and ensured the quality of vector data reconstruction.

关键词： vector quantization GPU codebook volume data

作者: Degui Xiao Yang Zhang Lei Li

作者单位: College oflnformation Science and Engineering Hunan University Changsha,China 410082

会议类型: 国际会议

会议名称: 2011 3rd International Conference on Education Technology and Computer(2011第三届教育技术与计算机国际会议 ICETC2011)

会议地点: 成都

会议语种:英文

页码: 117-121

在线出版日期: 2011-07-15（万方平台首次上网日期，不代表论文的发表时间）