Building on the recent advances in computer vision with convolutional neural networks, we have built SpectralNet, a spectroscopy-optimized convolutional neural network architecture capable of classifying spectra despite large temporal (i.e. translational, chemical, calibration) shifts. Present methods of measuring the local chemical environment of atoms at the nano-scale involve manual feature extraction and dimensionality reduction of the original signal such as: using the peak onset, the ratio of peaks, or the full-width half maximum of peaks. Convolutional neural networks like SpectralNet are able to automatically find parts of the spectra (i.e. features) of the spectra which maximally discriminate between the classes without requiring manual feature extraction. The advantage of such a process is to remove bias and qualitative interpretation in spectroscopy analysis which occurs during manual feature extraction. Because of this automated feature extraction process, this method of spectroscopy analysis is also immune to instrument calibration differences since it performs classification based on the shape of the spectra. Convolutional neural networks are an ideal statistical classifier for spectroscopy data (i.e. time-series data) due to its shared weighting scheme in neural network weights which is ideal for identifying local correlations between adjacent dimensions of the time-series data.
Over 2000 electron energy loss spectra were collected using a scanning transmission electron microscope of three oxidation states of Mn. SpectralNet was trained to learn the differences between them. We prove generalizability by training SpectralNet on electron energy loss spectroscopy data from one instrument, and test it on a variety of reference spectra found in the literature with perfect accuracy. We also test SpectralNet against a wide variety of high noise samples which a trained human spectroscopist would find incomprehensible. We also compare other neural network architectures used in the literature and determine that SpectralNet, a dense-layer free neural network, is immune to calibration differences whereas other styles of network are not. / Thesis / Master of Applied Science (MASc) / Spectroscopy is the study of the interaction between photons or electrons and a material to determine what that material is made of. One advanced way to make accurate measurements down to the atomic scale is to use high energy electrons in a transmission electron microscope. Using this instrument, a special type of photograph can be taken of the material (a spectrograph or spectrum) which is detailed enough to identify which kinds of atoms are in the material. The spectrographs are very complicated to interpret and the human eye struggles to find patterns in noisy and low resolution data. Depending on which instrument that the spectrographs are taken on, the resulting spectrograph will also change which adds extra difficulty. In this study, advanced algorithms are used to identify which types of atoms can be identified in the noisy signal from the spectrograph regardless of which instrument is used. These algorithms (convolutional neural networks) are also used in self-driving cars for a similar task of identifying objects whereas in this study we use it for identifying atoms.
| Identifer | oai:union.ndltd.org:mcmaster.ca/oai:macsphere.mcmaster.ca:11375/22832 |
| Date | 06 1900 |
| Creators | Chatzidakis, Michael |
| Contributors | Botton, Gianluigi, Materials Science and Engineering |
| Source Sets | McMaster University |
| Language | English |
| Detected Language | English |
| Type | Thesis |
Page generated in 0.0021 seconds