We extend the self-organizing mapping algorithm to the problem of visualizing data on Grassmann manifolds. In this setting, a collection of k points in n-dimensions is represented by a k-dimensional subspace, e.g., via the singular value or QR-decompositions. Data assembled in this way is challenging to visualize given abstract points on the Grassmannian do not reside in Euclidean space. The extension of the SOM algorithm to this geometric setting only requires that distances between two points can be measured and that any given point can be moved towards a presented pattern. The similarity between two points on the Grassmannian is measured in terms of the principal angles between subspaces, e.g., the chordal distance. Further, we employ a formula for moving one subspace towards another along the shortest path, i.e., the geodesic between two points on the Grassmannian. This enables a faithful implementation of the SOM approach for visualizing data consisting of k-dimensional subspaces of n-dimensional Euclidean space. We illustrate the resulting algorithm on a hyperspectral imaging application.
more »
« less
Endmember Extraction on the Grassmannian
Endmember extraction plays a prominent role in a variety of data analysis problems as endmembers often correspond to data representing the purest or best representative of some feature. Identifying endmembers then can be useful for further identification and classification tasks. In settings with high-dimensional data, such as hyperspectral imagery, it can be useful to consider endmembers that are subspaces as they are capable of capturing a wider range of variations of a signature. The endmember extraction problem in this setting thus translates to finding the vertices of the convex hull of a set of points on a Grassmannian. In the presence of noise, it can be less clear whether a point should be considered a vertex. In this paper, we propose an algorithm to extract endmembers on a Grassmannian, identify subspaces of interest that lie near the boundary of a convex hull, and demonstrate the use of the algorithm on a synthetic example and on the 220 spectral band AVIRIS Indian Pines hyperspectral image.
more »
« less
- Award ID(s):
- 1633830
- NSF-PAR ID:
- 10064957
- Date Published:
- Journal Name:
- 2018 IEEE Data Science Workshop (DSW 2018)
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Finding prototypes (e.g., mean and median) for a dataset is central to a number of common machine learning algorithms. Subspaces have been shown to provide useful, robust representations for datasets of images, videos and more. Since subspaces correspond to points on a Grassmann manifold, one is led to consider the idea of a subspace prototype for a Grassmann-valued dataset. While a number of different subspace prototypes have been described, the calculation of some of these prototypes has proven to be computationally expensive while other prototypes are affected by outliers and produce highly imperfect clustering on noisy data. This work proposes a new subspace prototype, the flag median, and introduces the FlagIRLS algorithm for its calculation. We provide evidence that the flag median is robust to outliers and can be used effectively in algorithms like Linde-Buzo-Grey (LBG) to produce improved clusterings on Grassmannians. Numerical experiments include a synthetic dataset, the MNIST handwritten digits dataset, the Mind's Eye video dataset and the UCF YouTube action dataset. The flag median is compared the other leading algorithms for computing prototypes on the Grassmannian, namely, the l_2-median and to the flag mean. We find that using FlagIRLS to compute the flag median converges in 4 iterations on a synthetic dataset. We also see that Grassmannian LBG with a codebook size of 20 and using the flag median produces at least a 10% improvement in cluster purity over Grassmannian LBG using the flag mean or l_2-median on the Mind's Eye dataset.more » « less
-
Hyperspectral super-resolution refers to the task of fusing a hyperspectral image (HSI) and a multispectral image (MSI) in order to produce a super-resolution image (SRI) that has high spatial and spectral resolution. Popular methods leverage matrix factorization that models each spectral pixel as a convex combination of spectral signatures belonging to a few endmembers. These methods are considered state-of-the-art, but several challenges remain. First, multiband images are naturally three dimensional (3-d) signals, while matrix methods usually ignore the 3-d structure, which is prone to information losses. Second, these methods do not provide identifiability guarantees under which the reconstruction task is feasible. Third, a tacit assumption is that the degradation operators from SRI to MSI and HSI are known - which is hardly the case in practice. Recently [1], [2] proposed a coupled tensor factorization approach to handle these issues. In this work we propose a hybrid model that combines the benefits of tensor and matrix factorization approaches. We also develop a new algorithm that is mathematically simple, enjoys identifiability under relaxed conditions and is completely agnostic of the spatial degradation operator. Experimental results with real hyperspectral data showcase the effectiveness of the proposed approach.more » « less
-
Abstract Lattice structures exhibit unique properties including a large surface area and a highly distributed load-path. This makes them very effective in engineering applications where weight reduction, thermal dissipation, and energy absorption are critical. Furthermore, with the advent of additive manufacturing (AM), lattice structures are now easier to fabricate. However, due to inherent surface complexity, their geometric construction can pose significant challenges. A classic strategy for constructing lattice structures exploits analytic surface–surface intersection; this, however, lacks robustness and scalability. An alternate strategy is voxel mesh-based isosurface extraction. While this is robust and scalable, the surface quality is mesh-dependent, and the triangulation will require significant postdecimation. A third strategy relies on explicit geometric stitching where tessellated open cylinders are stitched together through a series of geometric operations. This was demonstrated to be efficient and scalable, requiring no postprocessing. However, it was limited to lattice structures with uniform beam radii. Furthermore, existing algorithms rely on explicit convex-hull construction which is known to be numerically unstable. In this paper, a combinatorial stitching strategy is proposed where tessellated open cylinders of arbitrary radii are stitched together using topological operations. The convex hull construction is handled through a simple and robust projection method, avoiding expensive exact-arithmetic calculations and improving the computational efficiency. This is demonstrated through several examples involving millions of triangles. On a typical eight-core desktop, the proposed algorithm can construct approximately up to a million cylinders per second.more » « less
-
more » « less
Abstract Spectral imaging approaches provide new possibilities for measuring and discriminating fluorescent molecules in living cells and tissues. These approaches often employ tunable filters and robust image processing algorithms to identify many fluorescent labels in a single image set. Here, we present results from a novel spectral imaging technology that scans the fluorescence excitation spectrum, demonstrating that excitation‐scanning hyperspectral image data can discriminate among tissue types and estimate the molecular composition of tissues. This approach allows fast, accurate quantification of many fluorescent species from multivariate image data without the need of exogenous labels or dyes. We evaluated the ability of the excitation‐scanning approach to identify endogenous fluorescence signatures in multiple unlabeled tissue types. Signatures were screened using multi‐pass principal component analysis. Endmember extraction techniques revealed conserved autofluorescent signatures across multiple tissue types. We further examined the ability to detect known molecular signatures by constructing spectral libraries of common endogenous fluorophores and applying multiple spectral analysis techniques on test images from lung, liver and kidney. Spectral deconvolution revealed structure‐specific morphologic contrast generated from pure molecule signatures. These results demonstrate that excitation‐scanning spectral imaging, coupled with spectral imaging processing techniques, provides an approach for discriminating among tissue types and assessing the molecular composition of tissues. Additionally, excitation scanning offers the ability to rapidly screen molecular markers across a range of tissues without using fluorescent labels. This approach lays the groundwork for translation of excitation‐scanning technologies to clinical imaging platforms.
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
