WP XD

Powerful fluorescence and noise reduction

Home Products Raman Raman Software WP XD

WP XD

Record Raman spectra even for highly fluorescent samples

In many Raman applications, the sample’s autofluorescence can overwhelm the generally weak Raman signal. While several common approaches, like baseline subtraction, can remove the fluorescence background, the Raman signal often still remains hidden under the fluorescence’s counting noise. Wasatch’s WP XD algorithm uses a U-Net AI network, trained to remove both fluorescence and noise, which extracts a clean Raman signal even for challenging autofluorescent samples.

High-Quality Raman Spectra for Autofluorescent Samples

Sample fluorescence has limited the more widespread adoption of Raman spectroscopy. The counting noise associated with the strong fluorescence signal can exceed the size of the Raman peaks, and traditional approaches, such as baseline correction or shifted excitation struggle. Wasatch’s WP XD post-processing reduces both the fluorescence and its associated noise, yielding surprisingly clean Raman spectra with accurate signal strength. This post-processing approach opens new application areas, especially for natural and biological samples, which tend to show strong autofluorescence.

WP XD Recognizes General Raman Peaks not Specific Samples

WP XD employs a U-Net neural network structure, which is more typically applied to 2D image segmentation, adapted here for the clean-up of 1D Raman spectra. The spectra are interpolated to a fixed wavenumber spacing to allow the intermediate CNN (convolutional neural network) layers of the network to distinguish cleanly between slowly varying fluorescence background, rapidly varying noise and Raman peaks with intermediate frequency signal variation. The network discards noise and fluorescence but keeps all Raman-peaks in their accurate signal strength and is trained on actual Raman spectra recorded with Wasatch spectrometers. WP XD is agnostic to the sample identity: it recognizes typical Raman-looking peaks, not specific Raman spectra.

WP XD Built Into Free ENLIGHTEN^TM Software

WP XD was successfully tested with a wide variety of Wasatch Raman systems. It is fully integrated into the standard, free, open source software packages, ENLIGHTEN^TM and ENLIGHTEN^TM Mobile, and can be added to any data collection with the push of a button. The AI network cleans up spectra almost instantaneously, allowing a live display, even on regular, modest, computer hardware, including smart phones, it has even been tested to run on a Raspberry Pi.