Abstract:Radio galaxies exhibit a rich diversity of characteristics and emit radio emissions through a variety of radiation mechanisms, making their classification into distinct types based on morphology a complex challenge. To address this challenge effectively, we introduce an innovative approach for radio galaxy classification using COSFIRE filters. These filters possess the ability to adapt to both the shape and orientation of prototype patterns within images. The COSFIRE approach is explainable, learning-free, rotation-tolerant, efficient, and does not require a huge training set. To assess the efficacy of our method, we conducted experiments on a benchmark radio galaxy data set comprising of 1180 training samples and 404 test samples. Notably, our approach achieved an average accuracy rate of 93.36\%. This achievement outperforms contemporary deep learning models, and it is the best result ever achieved on this data set. Additionally, COSFIRE filters offer better computational performance, $\sim$20$\times$ fewer operations than the DenseNet-based competing method (when comparing at the same accuracy). Our findings underscore the effectiveness of the COSFIRE filter-based approach in addressing the complexities associated with radio galaxy classification. This research contributes to advancing the field by offering a robust solution that transcends the orientation challenges intrinsic to radio galaxy observations. Our method is versatile in that it is applicable to various image classification approaches.

What problem does this paper attempt to address?

### What problem does this paper attempt to solve? This paper aims to solve the complex challenge of classifying radio galaxies according to their morphological characteristics. Specifically, radio galaxies exhibit rich diversity and emit radio waves through multiple radiation mechanisms, which makes it very difficult to classify them into different types based on morphology. To effectively address this challenge, the author introduced an innovative method, namely using trainable COSFIRE filters (Combination of Shifted Filter Responses filters) to classify radio galaxies. #### Main problems and challenges 1. **Morphological diversity**: Radio galaxies have complex morphological features, including but not limited to Fanaroff - Riley I (FRI), Fanaroff - Riley II (FRII), compact type (Compact) and bent type (Bent). 2. **Directional change**: Radio galaxies have different directions in the data set, increasing the difficulty of classification. 3. **Limitations of existing methods**: - Deep - learning models such as convolutional neural networks (CNNs) perform well, but they require a large amount of training data and are prone to over - fitting for small data sets. - CNNs have an inherent vulnerability in dealing with rotational changes and require additional data augmentation or pre - processing steps to solve the rotational invariance problem. - High computational cost, especially when applied in resource - limited environments. #### Proposed solutions The author proposed a lightweight paradigm based on COSFIRE filters, with the following characteristics: - **Strong adaptability**: COSFIRE filters can adapt to the shape and direction of prototype patterns in the image. - **No need for a large amount of training data**: The COSFIRE method is unsupervised learning and does not require a huge training set. - **Rotational invariance**: By configuring multiple rotated versions of COSFIRE filters, tolerance to rotational changes is achieved. - **High efficiency**: The computational performance is better than that of existing deep - learning models (such as DenseNet), reducing the number of operations by about 20 times. #### Experimental results The author conducted experiments on a standard radio galaxy data set containing 1,180 training samples and 404 test samples, achieving an average accuracy of 93.36%, which is the best result obtained on this data set so far. In addition, COSFIRE filters are also significantly more computationally efficient than existing deep - learning models. In conclusion, this research provides a robust and efficient solution by introducing COSFIRE filters, solving complex problems such as morphological diversity and directional change in radio galaxy classification.

Classification of Radio Galaxies with trainable COSFIRE filters

Classification of radio galaxies with trainable COSFIRE filters

Content-Based Image Retrieval Using COSFIRE Descriptors with application to Radio Astronomy

Feature Guided Training and Rotational Standardisation for the Morphological Classification of Radio Galaxies

A multi-band AGN-SFG classifier for extragalactic radio surveys using machine learning

Classification of compact radio sources in the Galactic plane with supervised machine learning

Morphological classification of compact and extended radio galaxies using convolutional neural networks and data augmentation techniques

Morphological classification of Radio Galaxies using Semi-Supervised Group Equivariant CNNs

Radio Galaxy Zoo: Towards building the first multi-purpose foundation model for radio astronomy with self-supervised learning

Radio Galaxy Zoo: Giant Radio Galaxy Classification using Multi-Domain Deep Learning

Transfer learning for radio galaxy classification

Transfer learning for galaxy feature detection: Finding Giant Star-forming Clumps in low redshift galaxies using Faster R-CNN

Radio Galaxy Zoo: Knowledge Transfer Using Rotationally Invariant Self-Organising Maps

AutoSourceID-Classifier. Star-Galaxy Classification using a Convolutional Neural Network with Spatial Information

Quantifying Non-parametric Structure of High-redshift Galaxies with Deep Learning

An Observationally Driven Multifield Approach for Probing the Circum-Galactic Medium with Convolutional Neural Networks

Advances on the classification of radio image cubes

E(2)-Equivariant Features in Machine Learning for Morphological Classification of Radio Galaxies

RadioGalaxyNET: Dataset and Novel Computer Vision Algorithms for the Detection of Extended Radio Galaxies and Infrared Hosts

Simulating images of radio galaxies with diffusion models