Abstract:We present a generalized Non-negative factorization (NMF)-based data reduction pipeline for circumstellar disk and exoplanet detection. By using an adaptable pre-processing routine that applies algorithmic masks and corrections to improper data, we are able to easily offload the computationally-intensive NMF algorithm to a graphics processing unit (GPU), significantly increasing computational efficiency. NMF has been shown to better preserve disk structural features compared to other post-processing approaches and has demonstrated improvements in the analysis of archival data. The adaptive pre-processing routine of this pipeline, which automatically aligns and applies image corrections to the raw data, is shown to significantly improve chromatic halo suppression. Utilizing HST-STIS and JWST-MIRI coronagraphic datasets, we demonstrate a factor of five increase in real-time computational efficiency by using GPUs to perform NMF compared to using CPUs. Additionally, we demonstrate the usefulness of higher numbers of NMF components with SNR and contrast improvements, which necessitates the use of a more computationally efficient approach for data reduction.

What problem does this paper attempt to address?

This paper aims to solve several key problems in the detection of circumstellar disks and exoplanets. Specifically, it attempts to improve the detection efficiency and quality of these celestial structures by improving the data - processing flow. The following are the main problems addressed in this paper: 1. **Improve computational efficiency**: - The traditional Non - negative Matrix Factorization (NMF) algorithm is very time - consuming when processing large - scale data. To this end, the paper proposes a GPU - accelerated data - processing pipeline, which significantly improves the computational efficiency of the NMF algorithm. By offloading computationally intensive tasks to the GPU, a higher real - time computational efficiency than that of the CPU is achieved. 2. **Improve image - processing effects**: - The paper proposes an adaptive pre - processing program that can automatically align and correct various errors in the original image, thereby effectively suppressing the chromatic halo. This helps to better preserve the structural features of the dust disk and improve the signal - to - noise ratio (SNR) and contrast. 3. **Optimize the number of NMF components**: - Research shows that using more NMF components can better capture quasi - static noise characteristics, thereby improving the detection sensitivity. The paper shows that using more NMF components can significantly improve the SNR and contrast, which is crucial for high - quality astronomical image processing. 4. **Reduce forward - modeling dependence**: - Most existing high - contrast imaging algorithms rely on forward - modeling to accurately extract signals, but this method requires prior assumptions about the disk structure. The NMF algorithm does not require such assumptions and can therefore more flexibly detect faint disk structures. ### Specific methods and techniques - **NMF algorithm**: The paper details the basic principles of NMF and its application in astronomical image processing. NMF constructs a model point - spread function (PSF) by decomposing a matrix into two non - negative matrices and then removes these models from the scientific image. \[ X = WH \] where \(X\) is the initial matrix, and \(W\) and \(H\) are the target matrices to be decomposed. To minimize the cost function, the vectorized update rules provided by Zhu are adopted: \[ H \leftarrow H\odot\frac{W^T (V\odot X)}{W^T [V\odot(WH)]} \] \[ H \leftarrow H\odot\frac{(V\odot X)H^T}{[V\odot(WH)]H^T} \] - **GPU acceleration**: Utilize the powerful parallel computing capabilities of the GPU to significantly accelerate computationally intensive tasks such as matrix multiplication. Especially when dealing with a large number of components, the advantages of the GPU are more obvious. - **Adaptive pre - processing**: It includes steps such as automatic alignment and image correction to ensure the quality and consistency of the input data. ### Results and impacts The paper shows the processing results of multiple dust disks and exoplanets, proving the effectiveness of the new method. In particular, the processing of HIP 65426b shows its potential in planet detection. In addition, by comparing the processing effects under different numbers of components, the influence of the number of NMF components on image quality and detection accuracy is further verified. In conclusion, this paper significantly improves the efficiency and quality of dust disk and exoplanet detection by introducing GPU - acceleration and an optimized NMF algorithm, providing strong support for future astronomical observations and research.

NMF-based GPU accelerated coronagraphy pipeline

Parallel Adaptive Sparsity-Constrained NMF Algorithm for Hyperspectral Unmixing.

Coronagraphic Data Post-processing Using Projections on Instrumental Modes

The GPU Phase Folding and Deep Learning Method for Detecting Exoplanet Transits

Quasi-real-time analysis of dynamic near field scattering data using a graphics processing unit

Distributed Out-of-Memory NMF on CPU/GPU Architectures

cuFFS: A GPU-accelerated code for Fast Faraday Rotation Measure Synthesis

Hybrid propagation physics for the design and modeling of astronomical observatories: a coronagraphic example

Redundant Apodized Pupils (RAP) for high-contrast imagers robust to segmentation-due aberrations and island effects

Using Non-Negative Matrix Factorization to Improve Calibration of the Keck OSIRIS Integral Field Spectrograph

GPU Accelerated Anomaly Detection of Large Scale Light Curves

Disentangling Multiple Emitting Components in Molecular Observations with Non-negative Matrix Factorization

Sparse Non-negative Matrix Factorization on GPUs for Hyperspectral Unmixing

Minimization of non common path aberrations at the Palomar telescope using a self-coherent camera

Machine learning for exoplanet detection in high-contrast spectroscopy: Combining cross-correlation maps and deep learning on medium-resolution integral-field spectra

Direct Exoplanet Detection Using Deep Convolutional Image Reconstruction (ConStruct): A New Algorithm for Post-Processing High-Contrast Images

Disentangling Multiple Emitting Components in Molecular Observations with Nonnegative Matrix Factorization

Direct Exoplanet Detection using Convolutional Image Reconstruction (ConStruct): A New Algorithm for Post-processing High-contrast Images

NFD: Toward Real-time Mining of Short-timescale Gravitational Microlensing Events