Galaxy classification systems
Pro Research Analysisby 
Searched over 200M research papers
Traditional Galaxy Classification Systems: Hubble and Its Modifications
The foundation of galaxy classification began with the Hubble "tuning fork" diagram, which organized galaxies into ellipticals, spirals, and lenticulars (S0), based on their visual morphology. This system was later revised and expanded by de Vaucouleurs and Sandage to better describe the diversity of observed galaxies, including disk-shaped and non-disk-shaped types, and to address features like bulges, disks, and bars in spirals and S0 galaxies 17. However, the Hubble system has limitations, especially when classifying galaxies in dense cluster environments or at high redshift, where many objects do not fit neatly into its categories .
New and Parallel Classification Systems
To address the limitations of the Hubble system, alternative classification schemes have been proposed. One such system arranges normal spirals and lenticulars in parallel sequences, distinguishing "early" and "late" types by their disk-to-bulge ratios. This approach also introduces "anemic spirals," which are intermediate between gas-rich spirals and gas-poor S0 galaxies, highlighting the influence of environment on galaxy evolution . Other modern systems use three main physical properties—mass, star formation, and dynamical disturbances—to classify galaxies, reflecting the strong correlations among Hubble type, color, and stellar mass. These properties can be measured using structural parameters like concentration, asymmetry, and clumpiness (the CAS system), providing a more physically meaningful framework for classification .
Automated and Machine Learning-Based Galaxy Classification
With the growth of large astronomical datasets, automated classification methods have become essential. Early approaches used artificial neural networks to classify galaxies based on measured parameters, achieving high agreement with human classifications and enabling objective analysis of large samples . More recent studies employ advanced machine learning techniques, such as deep learning and ensemble methods, to classify galaxies into multiple morphological types. These methods can handle complex features in galactic images and achieve high accuracy, with some models reaching over 90% accuracy in distinguishing between spirals, ellipticals, and other classes 36810.
Machine learning models benefit from using a wide range of features, including structural, photometric, and color information. Principal component analysis and feature selection help improve classification performance, especially for challenging classes like mergers or rare types. Oversampling and balanced datasets further enhance the ability to classify minority classes accurately 5810.
Evolving Understanding and Flexibility in Classification
Research shows that galaxy classification is not static—galaxies can change types over time due to evolutionary processes, interactions, and environmental effects. For example, S0 galaxies may follow different evolutionary tracks, and some galaxies can transition between morphological types, as seen in the history of the Large Magellanic Cloud . The dichotomy between normal and barred spirals may also reflect differences in their evolutionary histories .
Conclusion
Galaxy classification systems have evolved from simple visual schemes to sophisticated, physically motivated, and automated methods. While the Hubble system remains influential, new frameworks and machine learning approaches provide greater flexibility and accuracy, accommodating the diversity and complexity of galaxies observed in the universe. These advances support a deeper understanding of galaxy formation, evolution, and the role of environment in shaping galactic properties 1246+3 MORE.
Sources and full results
Most relevant research papers on this topic
The fundamental properties of galaxies and a new galaxy classification system
This new galaxy classification system focuses on mass, star formation, and interactions/mergers, focusing on three key properties to better understand the physical state of nearby galaxies.
Galaxy classification based on deep learning
Deep learning-based galaxy classification methods, such as convolutional neural networks, enable efficient and accurate data processing and analysis in astronomy.
DOI