Table 1.
Method Index. Alphabetic index of methods treated in this review and the IMS application areas in which they have been demonstrated
| Methods | Section | Demonstrated applications in IMS |
|---|---|---|
| Adaptive edge‐preserving denoising | Section III.E | Image segmentation incorporating spatial information |
| AMASS | Section III.F | Soft image segmentation, probability‐based model, built‐in feature selection |
| Artificial neural networks | Section IV | Nonlinear dimensionality reduction, image segmentation, visualization of high‐dimensional IMS data |
| Autoscaling | Section II.5 | Data preprocessing |
| Autoencoders | Section IV | Nonlinear dimensionality reduction, image segmentation, visualization of high‐dimensional IMS data |
| Bisecting k‐means | Section III.C | Image segmentation, interactive exploration of clustering tree |
| Compressive sensing | Section II.B | Dimensionality reduction, increase spatial resolution |
| CX/CUR matrix decomposition | Section II.4 | Non‐negative pattern extraction and unmixing (in context of IMS), data size and dimensionality reduction |
| DWT | Section II.B | Dimensionality reduction, feature extraction |
| FCM | Section III.F | Soft image segmentation |
| Filter scaling | Section II.5 | Data preprocessing |
| GMM clustering | Section III.D | Image segmentation (hard and soft) |
| GSOM | Section IV.B | Nonlinear dimensionality reduction with built‐in dimensionality selection, image segmentation, visualization of high‐dimensional IMS data |
| HC | Section III.B | Image segmentation, interactive exploration of clustering tree |
| HDDC | Section III.D | Image segmentation (hard and soft), built‐in dimensionality reduction |
| ICA | Section II.C | Pattern extraction and unmixing, dimensionality reduction |
| k‐means clustering | Section III.C | Image segmentation, grouping of similar ion images |
| Kohonen map | Section IV.B | Nonlinear dimensionality reduction, image segmentation, visualization of high‐dimensional IMS data |
| Latent Dirichlet allocation | Section III.F | Soft image segmentation, probability‐based and generative model |
| MAF | Section II.D | Pattern extraction and unmixing incorporating spatial information, dimensionality reduction |
| MCR | Section II.1 | Non‐negative pattern extraction and unmixing, dimensionality reduction |
| MCR‐ALS | Section II.1 | Non‐negative pattern extraction and unmixing, dimensionality reduction |
| MNF transform | Section II.D | Pattern extraction and unmixing incorporating spatial information, dimensionality reduction |
| MOLDL | Section II.5 | Non‐negative pattern extraction and unmixing (in context of IMS) using prior information, dimensionality reduction |
| MRF | Section III.E | Image segmentation incorporating spatial information |
| NMF | Section II.2 | Non‐negative pattern extraction and unmixing, dimensionality reduction |
| NN‐PARAFAC | Section II.E.3 | Non‐negative pattern extraction and unmixing, pattern extraction, dimensionality reduction |
| PCA | Section II.A | Pattern extraction and unmixing, data size and dimensionality reduction |
| pLSA | Section II.3 | Non‐negative pattern extraction and unmixing, dimensionality reduction, generative and statistical mixture model |
| PMF | Section II.2 | Non‐negative pattern extraction and unmixing, dimensionality reduction |
| Poisson scaling | Section II.5 | Data preprocessing |
| Random projections | Section II.B | Dimensionality reduction |
| Shift‐variance scaling | Section II.5 | Data preprocessing |
| SMCR | Section II.5 | Non‐negative pattern extraction and unmixing, dimensionality reduction |
| SOM | Section IV.B | Nonlinear dimensionality reduction, image segmentation, visualization of high‐dimensional IMS data, image registration |
| Spatial shrunken centroids | Section III.F | Image segmentation (hard and soft), built‐in feature selection |
| Spatially aware clustering | Section III.E | Image segmentation incorporating spatial information |
| SVD | Section II.A | Pattern extraction and unmixing, dimensionality reduction |
| t‐SNE | Section IV.A | Nonlinear dimensionality reduction, image segmentation, visualization of high‐dimensional IMS data |
| Varimax | Section II.A.6 | Improve interpretability of matrix decomposition |
AMASS, algorithm for MSI analysis by semisupervised segmentation; DWT, discrete wavelet transform; FCM, fuzzy c‐means clustering; GMM, Gaussian mixture model; GSOM, growing self‐organizing map; HC, hierarchical clustering; HDDC, high dimensional data clustering; ICA, independent component analysis; IMS, imaging mass spectrometry; MAF, maximum autocorrelation factorization; MCR, multivariate curve resolution; MCR‐ALS, multivariate curve resolution by alternating least squares; MNF, minimum noise fraction; MOLDL, MOLecular Dictionary Learning; MRF, Markov random fields; NMF, non‐negative matrix factorization; NN‐PARAFAC, non‐negativity constrained parallel factor analysis; PCA, principal component analysis; pLSA, probabilistic latent semantic analysis; PMF, positive matrix factorization; SMCR, self modeling curve resolution; SOM, self‐organizing map; SVD, singular value decomposition; t‐SNE, t‐distributed stochastic neighborhood embedding.