Abstract
The application of Raman and infrared (IR) microspectroscopy is leading to hyperspectral data containing complementary information concerning the molecular composition of a sample. The classification of hyperspectral data from the individual spectroscopic approaches is already state-of-the-art in several fields of research. However, more complex structured samples and difficult measuring conditions might affect the accuracy of classification results negatively and could make a successful classification of the sample components challenging. This contribution presents a comprehensive comparison in supervised pixel classification of hyperspectral microscopic images, proving that a combined approach of Raman and IR microspectroscopy has a high potential to improve classification rates by a meaningful extension of the feature space. It shows that the complementary information in spatially co-registered hyperspectral images of polymer samples can be accessed using different feature extraction methods and, once fused on the feature-level, is in general more accurately classifiable in a pattern recognition task than the corresponding classification results for data derived from the individual spectroscopic approaches.
Highlights
The comprehensive term vibrational spectroscopy summarizes a number of optical measuring concepts that are applied to analyze the molecular structure and composition of a sample [18]
It could be shown that a feature-level fusion of information, extracted by statistic-based dimensionality reduction (DR) techniques, led to a numerical improvement for a clear majority of different classification setups, varying in features, training data and classification models
The results indicate that features derived from both spectroscopic approaches complement the feature space for classification tasks in a meaningful way
Summary
The comprehensive term vibrational spectroscopy summarizes a number of optical measuring concepts that are applied to analyze the molecular structure and composition of a sample [18]. Material-specific spectral fingerprints can be derived, providing information that enables the identification and the qualitative or quantitative analysis of samples [18,19] Both methods share an underlying principle, they differ in terms of the demands concerning their physical processes, leading to the rule of mutual exclusion [26], which defines that vibrations and rotations that are active for Raman spectroscopy are non-active for IR spectroscopy, and vice versa. The polymer sample only contains 2 different components, leading to a 3-class classification problem of pure material targets In this contribution, we build on the results from Gowen and Dorrepal in Ref.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
