摘要
虽然Gabor变换在语音与图象处理、雷达、声纳、振动信号的处理与理解等很多领域被认为是非常有用的方法,然而实时应用却因其很高的计算复杂性而受到限制。为了有效地和快速地计算离散Gabor变换,论文提出了在临界抽样条件下和在过抽样条件下,一维离散Gabor变换系数求解的块时间递归算法以及由变换系数重建原信号的块时间递归算法,研究了两算法使用并行格型结构的实现方法,并讨论和比较了算法的计算复杂性和优越性。
Gabor transforms are considered as an effective time-frequency analysis technique in diverse areas such as speech and image processing,radar,sonar and seismic data processing and interpretation;however,its real time applications have been limited due to its high computational complexity.In this paper,firstly,block time-recursive algorithms for the efficient and fast computation of the 1-D DGT coefficients and for the fast reconstruction of the original signal from the DGT coefficients will be developed in both the critical sampling case and the oversampling case;secondly,unified parallel lattice structures for the implementation of the algorithms will be studied;and finally,the computational complexity and the advantages of the proposed algorithms will be discussed and compared.
出处
《计算机工程与应用》
CSCD
北大核心
2004年第30期57-60,共4页
Computer Engineering and Applications
基金
教育部优秀青年教师资助计划项目(教人司眼[2002]演40号)
安徽省自然科学基金项目(编号:01042210)
安徽省教育厅自然科学重点研究项目(编号:2001kj020zd)
安徽大学人才队伍建设经费资助
关键词
离散
GABOR
变换
并行格型结构
块时间递归算法
discrete Gabor transforms,parallel lattice structure,block time-recursive algorithm
