Search for the edge-on galaxies using an artificial neural network

S. S. Savchenko,D. I. Makarov,A. V. Antipova,I.S. Tikhonenko

DOI: https://doi.org/10.1016/j.ascom.2023.100771

2023-12-05

Abstract:We present an application of an artificial neural network methodology to a modern wide-field sky survey Pan-STARRS1 in order to build a high-quality sample of disk galaxies visible in edge-on orientation. Such galaxies play an important role in the study of the vertical distribution of stars, gas and dust, which is usually not available to study in other galaxies outside the Milky Way. We give a detailed description of the network architecture and the learning process. The method demonstrates good effectiveness with detection rate about 97\% and it works equally well for galaxies over a wide range of brightnesses and sizes, which resulted in a creation of a catalogue of edge-on galaxies with $10^5$ of objects. The catalogue is published on-line with an open access.

Instrumentation and Methods for Astrophysics,Astrophysics of Galaxies

What problem does this paper attempt to address?

### What problems does this paper attempt to solve?

This paper aims to screen out high - quality edge - on disk - galaxy samples from the modern wide - field survey Pan - STARRS1 by using the Artificial Neural Network (ANN) method. Specifically, this research mainly solves the following problems:

1. **Observation limitations**:
- Edge - on galaxies are very important for studying the vertical distribution of stars, gas, and dust, because these features are difficult to directly observe in other non - edge - on galaxies.
- Traditional manual classification methods (such as those based on visual recognition) have the problems of subjectivity and low efficiency.

2. **Deficiencies in existing catalogs**:
- The currently existing edge - on galaxy catalogs (such as RFGC and EGIS) are limited in number and may have inaccurate classification problems.
- Although the crowd - sourced classification method (such as Galaxy Zoo 1) has a large amount of data, it has significant deviation and misclassification problems, especially the strict definition of "edge - on" is not consistent enough.

3. **Challenges in automatic detection**:
- Edge - on galaxies have diverse morphologies, including red Sa - type galaxies dominated by bulges to blue Sd - type galaxies with almost no bulges.
- The axis ratio of a galaxy cannot uniquely determine its inclination, so relying solely on this parameter may lead to sample contamination or bias towards certain types of galaxies.
- Other factors such as foreground object interference, false splitting caused by dust bands, and image artifacts also increase the difficulty of automatic detection.

To solve these problems, the authors propose a deep - learning - based Convolutional Neural Network (CNN) method for efficiently and accurately identifying edge - on galaxies. Through this method, they created an edge - on galaxy catalog containing more than 105 objects and made it publicly accessible online. This method shows a 97% detection rate and is applicable to galaxies of different brightness and size.

### Main contributions of the paper

- **Improve classification accuracy**: Through the deep - learning model, the identification accuracy of edge - on galaxies is significantly improved.
- **Reduce human - induced bias**: Avoid the subjective bias in traditional visual classification methods.
- **Expand sample size**: Create the largest edge - on galaxy sample library to date.
- **Wide application**: This method is applicable to galaxies of different brightness and size and has wide application potential.

### Conclusion

By introducing advanced machine - learning techniques, especially the Convolutional Neural Network, this research successfully overcomes the limitations of traditional methods and provides astronomers with higher - quality edge - on galaxy samples, which is helpful for in - depth study of the structure and evolution of galaxies.