Adaptive three-dimensional wavelet analysis for denoising TOF-SIMS images: Toward digital staining of pathological specimens

In this report, we propose a three-dimensional adaptive wavelet analysis algorithm for rapidly denoising data acquired by time-of-flight secondary ion mass spectrometry (TOF-SIMS). In the computation, the TOF-SIMS data are stored in three-dimensional space, where the xy-plane corresponds to the measured area and the z-axis corresponds to the m/z mass spectrum. Reconstructed images for specific m/z peaks along the z-axis indicate the spatial distributions of components such as protein and lipid. In this algorithm, two different basis functions are applied to the xy-plane and the z-axis, and three-dimensional wavelet shrinkage and reconstruction is performed. One basis function is suitable for continuous data such as the distribution of components, and the other basis function is suitable for discrete data such as a mass spectra. We apply this algorithm to denoising of simulated and experimental image data. The results show that the noise is markedly reduced without loss of the original signals and that the signal-to-noise ratio is substantially improved by the three-dimensional noise reduction algorithm.