PDRs4All XX. Haute Couture: Spectral stitching of JWST MIRI-IFU cubes with matrix completion

TL;DR

Haute Couture addresses the challenge of stitching JWST MIRI-MRS data cubes by formulating it as a matrix completion problem solved via nonnegative matrix factorization, while first applying a pre-processing step to homogenize intensities across the twelve sub-cubes. The method preserves the highest spatial resolution across the full spectral range and leverages overlaps between channels to reconstruct a uniform cube, demonstrated on the PDRs4All Orion Bar dataset. Compared with coarse stitching, Haute Couture delivers continuous spectra and maintains spatial detail, enabling recovery of structural features such as the 203-506 protoplanetary disk. The approach, which includes a closed-form solution for intensity-scaling and a low-rank, nonnegative factorization framework, is broadly applicable to JWST MIRI-MRS data and other IFU datasets, with code to be released upon publication.

Abstract

MIRI is the imager and spectrograph covering wavelengths from $4.9$ to $27.9$ $μ$m onboard the James Webb Space Telescope (JWST). The Medium-Resolution Spectrometer (MRS) consists of four integral field units (IFU), each of which has three sub-channels. The twelve resulting spectral data cubes have different fields of view, spatial, and spectral resolutions. The wavelength range of each cube partially overlaps with the neighboring bands, and the overlap regions typically show flux mismatches which have to be corrected by spectral stitching methods. Stitching methods aim to produce a single data cube incorporating the data of the individual sub-channels, which requires matching the spatial resolution and the flux discrepancies. We present Haute Couture, a novel stitching algorithm which uses non-negative matrix factorization (NMF) to perform a matrix completion, where the available MRS data cubes are treated as twelve sub-matrices of a larger incomplete matrix. Prior to matrix completion, we also introduce a novel pre-processing to homogenize the global intensities of the twelve cubes. Our pre-processing consists in jointly optimizing a set of global scale parameters that maximize the fit between the cubes where spectral overlap occurs. We apply our novel stitching method to JWST data obtained as part of the PDRs4All observing program of the Orion Bar, and produce a uniform cube reconstructed with the best spatial resolution over the full range of wavelengths.

Figures (9)

• Figure 1: Examples of intensity gaps observed in overlapping areas of adjacent cubes, obtained from MIRI-MRS data collected within the PDRs4All program (see Section \ref{['sec:data_preprocessing']} for more details). The spectra displayed correspond to the average spectra over the entire field of view for two overlapping sub-channels. Panel a: channel 2-medium and 2-long. Panel b: channel 2-long and 3-short. Panel c: channel 3-short and 3-medium.
• Figure 2: Illustration of the stitching problem for two cubes. On the left, we observe that the shortest-wavelength cube in blue has a better spatial resolution (as illustrated by the finer grid) and smaller field of view than the longest-wavelength cube in red. The space $\times$ wavelength supports of the cubes partially overlap. The coarse stitching procedure shown in Figure 1.a sacrifices spatial resolution, while Haute Couture shown in Figure 1.b enables the reconstruction of a cube with best spatial resolution over the full range of wavelengths.
• Figure 3: The twelve unfolded cubes provided by the twelve sub-channels can be organized as sub-matrices of a larger matrix $\mathbf{X}$ with an almost block-diagonal structure. The wavelength range of consecutive blocks overlap for a few rows, corresponding to spectral bands observed by consecutive sub-channels. Blue blocks represent channel 1, green ones channel 2, yellow ones channel 3 and red ones channel 4. White and light blue parts correspond to unobserved coefficients. The light blue area corresponds to the area we are specifically interested in reconstructing.
• Figure 4: Colorized image NIRCam of the Orion Bar of the PDRs4All ERS program (berne2024). Filters F140M and F210M are in blue; F277W, F300M, F323N and F335M in green; F405N in orange; and F444W, F480M and F470N in red. The pattern in white is the MIRI-MRS mosaic footprint. The red box corresponds to the pointing used in this article.
• Figure 5: The first row presents the average spectra over the entire field of view of each channel (shown in the second row), with a different color (blue, orange, green) for each sub-channel (short, medium, long). The second row presents the MIRI-MRS images at different wavelengths (the images are aligned). It can be observed that the field of view increases with the channel, but that the spatial resolution deteriorates.