The rapid development of spectral-imaging methods in scanning probe,electron,and optical microscopy in the last decade have given rise for large multidimensional datasets.In many cases,the reduction of hyperspectral d...The rapid development of spectral-imaging methods in scanning probe,electron,and optical microscopy in the last decade have given rise for large multidimensional datasets.In many cases,the reduction of hyperspectral data to the lower-dimension materialsspecific parameters is based on functional fitting,where an approximate form of the fitting function is known,but the parameters of the function need to be determined.However,functional fits of noisy data realized via iterative methods,such as least-square gradient descent,often yield spurious results and are very sensitive to initial guesses.Here,we demonstrate an approach for the reduction of the hyperspectral data using a deep neural network approach.A combined deep neural network/least-square approach is shown to improve the effective signal-to-noise ratio of band-excitation piezoresponse force microscopy by more than an order of magnitude,allowing characterization when very small driving signals are used or when a material’s response is weak.展开更多
We investigate the ability to reconstruct and derive spatial structure from sparsely sampled 3D piezoresponse force microcopy data,captured using the band-excitation(BE)technique,via Gaussian Process(GP)methods.Even f...We investigate the ability to reconstruct and derive spatial structure from sparsely sampled 3D piezoresponse force microcopy data,captured using the band-excitation(BE)technique,via Gaussian Process(GP)methods.Even for weakly informative priors,GP methods allow unambiguous determination of the characteristic length scales of the imaging process both in spatial and frequency domains.We further show that BE data set tends to be oversampled in the spatial domains,with~30% of original data set sufficient for high-quality reconstruction,potentially enabling faster BE imaging.At the same time,reliable reconstruction along the frequency domain requires the resonance peak to be within the measured band.This behavior suggests the optimal strategy for the BE imaging on unknown samples.Finally,we discuss how GP can be used for automated experimentation in SPM,by combining GP regression with non-rectangular scans.展开更多
基金This research used resources of the Compute and Data Environment for Science(CADES)at the Oak Ridge National Laboratory,which is supported by the Office of Science of the U.S.Department of Energy under Contract No.DE-AC05-00OR22725。
文摘The rapid development of spectral-imaging methods in scanning probe,electron,and optical microscopy in the last decade have given rise for large multidimensional datasets.In many cases,the reduction of hyperspectral data to the lower-dimension materialsspecific parameters is based on functional fitting,where an approximate form of the fitting function is known,but the parameters of the function need to be determined.However,functional fits of noisy data realized via iterative methods,such as least-square gradient descent,often yield spurious results and are very sensitive to initial guesses.Here,we demonstrate an approach for the reduction of the hyperspectral data using a deep neural network approach.A combined deep neural network/least-square approach is shown to improve the effective signal-to-noise ratio of band-excitation piezoresponse force microscopy by more than an order of magnitude,allowing characterization when very small driving signals are used or when a material’s response is weak.
基金M.A.and D.K.acknowledge support from CNMS user facility,project #CNMS2019-272.
文摘We investigate the ability to reconstruct and derive spatial structure from sparsely sampled 3D piezoresponse force microcopy data,captured using the band-excitation(BE)technique,via Gaussian Process(GP)methods.Even for weakly informative priors,GP methods allow unambiguous determination of the characteristic length scales of the imaging process both in spatial and frequency domains.We further show that BE data set tends to be oversampled in the spatial domains,with~30% of original data set sufficient for high-quality reconstruction,potentially enabling faster BE imaging.At the same time,reliable reconstruction along the frequency domain requires the resonance peak to be within the measured band.This behavior suggests the optimal strategy for the BE imaging on unknown samples.Finally,we discuss how GP can be used for automated experimentation in SPM,by combining GP regression with non-rectangular scans.
