Understanding Raman Spectral Based Classifications with Convolutional Neural Networks Using Practical Examples of Fungal Spores and Carotenoid-Pigmented Microorganisms
Numerous publications showing that robust prediction models for microorganisms based on Raman micro-spectroscopy in combination with chemometric methods are feasible, often with very precise predictions. Advances in machine learning and easier accessibility to software make it increasingly easy for...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-01-01
|
| Series: | AI |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2673-2688/4/1/6 |
| _version_ | 1797614033398923264 |
|---|---|
| author | Thomas J. Tewes Michael C. Welle Bernd T. Hetjens Kevin Saruni Tipatet Svyatoslav Pavlov Frank Platte Dirk P. Bockmühl |
| author_facet | Thomas J. Tewes Michael C. Welle Bernd T. Hetjens Kevin Saruni Tipatet Svyatoslav Pavlov Frank Platte Dirk P. Bockmühl |
| author_sort | Thomas J. Tewes |
| collection | DOAJ |
| description | Numerous publications showing that robust prediction models for microorganisms based on Raman micro-spectroscopy in combination with chemometric methods are feasible, often with very precise predictions. Advances in machine learning and easier accessibility to software make it increasingly easy for users to generate predictive models from complex data. However, the question regarding why those predictions are so accurate receives much less attention. In our work, we use Raman spectroscopic data of fungal spores and carotenoid-containing microorganisms to show that it is often not the position of the peaks or the subtle differences in the band ratios of the spectra, due to small differences in the chemical composition of the organisms, that allow accurate classification. Rather, it can be characteristic effects on the baselines of Raman spectra in biochemically similar microorganisms that can be enhanced by certain data pretreatment methods or even neutral-looking spectral regions can be of great importance for a convolutional neural network. Using a method called Gradient-weighted Class Activation Mapping, we attempt to peer into the black box of convolutional neural networks in microbiological applications and show which Raman spectral regions are responsible for accurate classification. |
| first_indexed | 2024-03-11T07:02:59Z |
| format | Article |
| id | doaj.art-e0e39297472041478259b3615171eefe |
| institution | Directory Open Access Journal |
| issn | 2673-2688 |
| language | English |
| last_indexed | 2024-03-11T07:02:59Z |
| publishDate | 2023-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | AI |
| spelling | doaj.art-e0e39297472041478259b3615171eefe2023-11-17T09:08:41ZengMDPI AGAI2673-26882023-01-014111412710.3390/ai4010006Understanding Raman Spectral Based Classifications with Convolutional Neural Networks Using Practical Examples of Fungal Spores and Carotenoid-Pigmented MicroorganismsThomas J. Tewes0Michael C. Welle1Bernd T. Hetjens2Kevin Saruni Tipatet3Svyatoslav Pavlov4Frank Platte5Dirk P. Bockmühl6Faculty of Life Sciences, Rhine-Waal University of Applied Sciences, Marie-Curie-Straße 1, 47533 Kleve, GermanyDepartment of Robotics, Perception, and Learning, KTH Royal Institute of Technology, 10044 Stockholm, SwedenFaculty of Life Sciences, Rhine-Waal University of Applied Sciences, Marie-Curie-Straße 1, 47533 Kleve, GermanyFaculty of Life Sciences, Rhine-Waal University of Applied Sciences, Marie-Curie-Straße 1, 47533 Kleve, GermanyFaculty of Life Sciences, Rhine-Waal University of Applied Sciences, Marie-Curie-Straße 1, 47533 Kleve, GermanyFaculty of Life Sciences, Rhine-Waal University of Applied Sciences, Marie-Curie-Straße 1, 47533 Kleve, GermanyFaculty of Life Sciences, Rhine-Waal University of Applied Sciences, Marie-Curie-Straße 1, 47533 Kleve, GermanyNumerous publications showing that robust prediction models for microorganisms based on Raman micro-spectroscopy in combination with chemometric methods are feasible, often with very precise predictions. Advances in machine learning and easier accessibility to software make it increasingly easy for users to generate predictive models from complex data. However, the question regarding why those predictions are so accurate receives much less attention. In our work, we use Raman spectroscopic data of fungal spores and carotenoid-containing microorganisms to show that it is often not the position of the peaks or the subtle differences in the band ratios of the spectra, due to small differences in the chemical composition of the organisms, that allow accurate classification. Rather, it can be characteristic effects on the baselines of Raman spectra in biochemically similar microorganisms that can be enhanced by certain data pretreatment methods or even neutral-looking spectral regions can be of great importance for a convolutional neural network. Using a method called Gradient-weighted Class Activation Mapping, we attempt to peer into the black box of convolutional neural networks in microbiological applications and show which Raman spectral regions are responsible for accurate classification.https://www.mdpi.com/2673-2688/4/1/6Ramanconvolutional neural networkGrad-CAMmicroorganismscarotenoidsconidia |
| spellingShingle | Thomas J. Tewes Michael C. Welle Bernd T. Hetjens Kevin Saruni Tipatet Svyatoslav Pavlov Frank Platte Dirk P. Bockmühl Understanding Raman Spectral Based Classifications with Convolutional Neural Networks Using Practical Examples of Fungal Spores and Carotenoid-Pigmented Microorganisms AI Raman convolutional neural network Grad-CAM microorganisms carotenoids conidia |
| title | Understanding Raman Spectral Based Classifications with Convolutional Neural Networks Using Practical Examples of Fungal Spores and Carotenoid-Pigmented Microorganisms |
| title_full | Understanding Raman Spectral Based Classifications with Convolutional Neural Networks Using Practical Examples of Fungal Spores and Carotenoid-Pigmented Microorganisms |
| title_fullStr | Understanding Raman Spectral Based Classifications with Convolutional Neural Networks Using Practical Examples of Fungal Spores and Carotenoid-Pigmented Microorganisms |
| title_full_unstemmed | Understanding Raman Spectral Based Classifications with Convolutional Neural Networks Using Practical Examples of Fungal Spores and Carotenoid-Pigmented Microorganisms |
| title_short | Understanding Raman Spectral Based Classifications with Convolutional Neural Networks Using Practical Examples of Fungal Spores and Carotenoid-Pigmented Microorganisms |
| title_sort | understanding raman spectral based classifications with convolutional neural networks using practical examples of fungal spores and carotenoid pigmented microorganisms |
| topic | Raman convolutional neural network Grad-CAM microorganisms carotenoids conidia |
| url | https://www.mdpi.com/2673-2688/4/1/6 |
| work_keys_str_mv | AT thomasjtewes understandingramanspectralbasedclassificationswithconvolutionalneuralnetworksusingpracticalexamplesoffungalsporesandcarotenoidpigmentedmicroorganisms AT michaelcwelle understandingramanspectralbasedclassificationswithconvolutionalneuralnetworksusingpracticalexamplesoffungalsporesandcarotenoidpigmentedmicroorganisms AT berndthetjens understandingramanspectralbasedclassificationswithconvolutionalneuralnetworksusingpracticalexamplesoffungalsporesandcarotenoidpigmentedmicroorganisms AT kevinsarunitipatet understandingramanspectralbasedclassificationswithconvolutionalneuralnetworksusingpracticalexamplesoffungalsporesandcarotenoidpigmentedmicroorganisms AT svyatoslavpavlov understandingramanspectralbasedclassificationswithconvolutionalneuralnetworksusingpracticalexamplesoffungalsporesandcarotenoidpigmentedmicroorganisms AT frankplatte understandingramanspectralbasedclassificationswithconvolutionalneuralnetworksusingpracticalexamplesoffungalsporesandcarotenoidpigmentedmicroorganisms AT dirkpbockmuhl understandingramanspectralbasedclassificationswithconvolutionalneuralnetworksusingpracticalexamplesoffungalsporesandcarotenoidpigmentedmicroorganisms |