Applied and Computational Harmonic Analysis ( IF 2.5 ) Pub Date : 2022-07-11 , DOI: 10.1016/j.acha.2022.07.003 Charles K. Chui
Three families of super-resolution (SR) wavelets , and , to be called Gaussian SR (GSR), spline SR (SSR) and dual-spline SR (DSSR) wavelets, respectively, are introduced in this paper for resolving the super-resolution problem of recovering any point-mass , with for , , and for all , where and are allowed to be arbitrarily small. Let , or , with or w, respectively, where is suppressed. The SR wavelets are designed to have the n-th order of vanishing moments, with Fourier transform of their complex conjugates to possess the following properties: (i) for all , (ii) , where and are positive constants independent of and m, and (iii) the widths (or standard deviations) of , with center at κ, tends to 0 very fast for large values of α. While the most popular approach to resolve this super-resolution problem is to consider the Fourier transform of as the “data function” for solving the inverse problem of recovering L, and of the point-mass , our proposed approach is to consider the “enhanced data function” , where , to be called the search function in this paper, is obtained by taking the continuous wavelet transform (CWT): of the data function , with as the analysis wavelet, followed by applying wavelet thresholding to “de-noise” the data function , by choosing an appropriate thresholding parameter , with , in order not to remove any of the coefficients , where ; and finally by setting . Hence, the enhanced data function is at least cleaner than the data function .
In our proposed approach, instead of directly recovering as in the published literature, we propose a “divide and conquer” strategy: first by applying “bottom-up thresholding” of the search function , with thresholding parameter close to but not exceeding , to separate the set of the local extrema locations of the function in into disjoint intervals of clusters, with more and smaller intervals and less number of local extrema in each interval for larger values of α; and secondly, by applying “top-down thresholding” to extract, one-by-one, of all local maxima, followed by all local minima (after a sign change), for each and every cluster. A desired leeway and lower bounds of the choice of the width parameter α are derived for the iterative application of top-down thresholding. Extension to for is also studied in this paper. For , we observe that the imagery of the enhanced data function for a single point-mass at where , resembles that of an “Airy disk” with center at in light microscopy and celestial telescopy of point-masses.
中文翻译:
用于恢复具有任意小系数的任意接近点质量的超分辨率小波
三族超分辨率 (SR) 小波,和,分别称为高斯SR(GSR),样条SR(SSR)和双样条SR(DSSR)小波,用于解决恢复任何点质量的超分辨率问题, 和为了,, 和对所有人, 在哪里和允许任意小。让,或者, 和或w,分别在哪里被压制。SR 小波被设计为具有n阶消失矩,并对其复共轭进行傅里叶变换具备以下性质:(i)对所有人, (ii), 在哪里和是独立于的正常数和m,以及 (iii) 的宽度(或标准偏差),以κ为中心,对于较大的α值,非常快地趋于 0 。虽然解决这个超分辨率问题最流行的方法是考虑傅里叶变换的作为解决恢复L的逆问题的“数据函数” ,和质点的,我们提出的方法是考虑“增强数据功能”, 在哪里,在本文中称为搜索函数,通过连续小波变换(CWT)得到:数据函数, 和作为分析小波,然后应用小波阈值对数据函数进行“去噪”,通过选择适当的阈值参数, 和, 为了不删除任何系数, 在哪里; 最后通过设置. 因此,增强的数据功能至少比数据函数更干净.
在我们提出的方法中,不是直接恢复与已发表的文献一样,我们提出了“分而治之”的策略:首先通过应用搜索功能的“自下而上阈值”, 带有阈值参数接近但不超过, 以分离局部极值位置的集合功能的在成不相交的簇间隔,间隔越来越小,局部极值的数量越来越少在每个区间中,对于较大的α值;其次,通过应用“自上而下的阈值”来逐一提取所有局部最大值,然后是所有局部最小值(在符号更改后),对于每个集群。理想的余地为自上而下阈值的迭代应用推导出宽度参数α选择的下界。扩展至为了本文也进行了研究。为了,我们观察到单个点质量的增强数据函数的图像在在哪里, 类似于中心在的“艾里斑”在光学显微镜和点质量的天体望远镜中。