Wyniki wyszukiwania - BazTech

Ograniczanie wyników

Znaleziono wyników: 2

Liczba wyników na stronie

Wyniki wyszukiwania

Sortuj według:

Ogranicz wyniki do:

Detection of potentially anomalous cosmic particle tracks acquired with CMOS sensors: Validation of rough k-means clustering with PCA feature extraction

Hachaj Tomasz, Piekarczyk Marcin, Wąs Jarosław

International Journal of Applied Mathematics and Computer Science

2025

Vol. 35, no. 1

7--18

We present a method capable of detecting potentially anomalous cosmic particle tracks acquired with complementary metal-oxide-semiconductor (CMOS) sensors. We apply a principal components analysis-based feature extraction method and rough k-means clustering for outlier detection. We evaluated our approach on more than 104 images acquired by the Cosmic Ray Extremely Distributed Observatory (CREDO). The method presented in this work proved to be an effective solution. The analysis of the behavior of the rough k-means clustering-based algorithm presented here and the method of selecting its parameters showed that the algorithm performs as expected and demonstrates efficiency, stability, and repeatability of results for the test data set. The results included in this work are very relevant to the international CREDO project and the broader problem of anomaly analysis in image data sets. We plan to deploy the presented methodology in the image processing pipeline of the large data set we are working on in the CREDO project. The results can be reproduced using our source code, which is published in an open repository.

Determination of spherical coordinates of sampled cosmic ray flux distribution using principal components analysis and deep encoder-decoder network

Hachaj Tomasz, Piekarczyk Marcin, Bibrzycki Łukasz, Wąs Jarosław

Machine Graphics and Vision

2024

Vol. 33, No. 2

29--45

In this paper we propose a novel algorithm based on the use of Principal Components Analysis for the determination of spherical coordinates of sampled cosmic ray flux distribution. We have also applied a deep neural network with encoder-decoder (E-D) architecture in order to filter-off variance noises introduced by sampling. We conducted a series of experiments testing the effectiveness of our estimations. The training set consisted of 92 250 images and validation set of 37 800 images. We have calculated mean absolute error (MAE) between real values and estimations. When E-D is applied, the number of cases (estimations) where MAE < 10 increases from 48% to 79% for θ and from 62% to 65% for ϕ, MAE < 5 increases from 24% to 45% for θ and from 47% to 52% for ϕ, MAE < 1 increases from 6% to 9% for θ and from 12% to 16% for ϕ, where θ is the zenith angle, and ϕ is the azimuthal angle. This is a significant change and it demonstrates the high utility of the E-D network use and shows the accuracy of the PCA-based algorithm. We also publish the source code used in our research in order to make it reproducible.