Scalable Galaxy Morphology Classification through Hybrid Handcrafted and Deep Learning Approaches
Resumo
The rapid growth of astronomical surveys has transformed galaxy morphology classification into a large-scale data analysis problem that requires automated and scalable machine learning solutions. In this work, we evaluate handcrafted and deep learning approaches using a balanced subset of the Galaxy Zoo 2 dataset containing seven morphological classes. We first investigate the impact of a multi-step preprocessing pipeline and extract handcrafted descriptors capturing textural, morphological, intensity, and structural properties of galaxies. These descriptors are evaluated using classical machine learning models, including SVM, MLP, Random Forest, and KNN. We then assess convolutional neural networks (AlexNet, ResNet50, MobileNetV2, and EfficientNetB0) using transfer learning on both preprocessed and raw centrally cropped images. Finally, we explore a hybrid ensemble combining handcrafted and deep representations. The best configuration, based on a weighted soft-voting scheme, achieved an F1-score of 0.929. These results indicate that handcrafted descriptors and deep representations capture complementary information, and that their integration can improve scalable scientific image analysis in data-intensive research environments.
Referências
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., and Fei-Fei, L. (2009). Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255. Ieee.
Fortson, L., Masters, K., Nichol, R., Borne, K., Edmondson, E., Lintott, C., Raddick, J., Schawinski, K., and Wallin, J. (2011). Galaxy zoo: Morphological classification and citizen science.
He, K., Zhang, X., Ren, S., and Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770–778.
Hubble, E. P. (1926). Extragalactic nebulae. Astrophysical Journal, 64, 321-369 (1926), 64.
Kalvankar, S., Pandit, H., and Parwate, P. (2020). Galaxy morphology classification using efficientnet architectures. arXiv preprint arXiv:2008.13611.
Krizhevsky, A., Sutskever, I., and Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems, 25.
Ntwaetsile, K. and Geach, J. E. (2021). Rapid sorting of radio galaxy morphology using haralick features. Monthly Notices of the Royal Astronomical Society, 502(3):3417–3425.
Ojansivu, V. and Heikkilä, J. (2008). Blur insensitive texture classification using local phase quantization. In International conference on image and signal processing, pages 236–243. Springer.
Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Prettenhofer, P., Weiss, R., Dubourg, V., Vanderplas, J., Passos, A., Cournapeau, D., Brucher, M., Perrot, M., and Duchesnay, E. (2011). Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, 12:2825–2830.
Sandler, M., Howard, A., Zhu, M., Zhmoginov, A., and Chen, L.-C. (2018). Mobilenetv2: Inverted residuals and linear bottlenecks. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 4510–4520.
Shamir, L. (2009). Automatic morphological classification of galaxy images. Monthly Notices of the Royal Astronomical Society, 399(3):1367–1372.
Tan, M. and Le, Q. (2019). Efficientnet: Rethinking model scaling for convolutional neural networks. In International conference on machine learning, pages 6105–6114. PMLR.
Trickz, J. (2020). Galaxy Zoo 2: Images. [link].
Willett, K. W., Lintott, C. J., Bamford, S. P., Masters, K. L., Simmons, B. D., Casteels, K. R. V., Edmondson, E. M., Fortson, L. F., Kaviraj, S., Keel, W. C., Melvin, T., Nichol, R. C., Raddick, M. J., Schawinski, K., Simpson, R. J., Skibba, R. A., Smith, A. M., and Thomas, D. (2013). Galaxy zoo 2: detailed morphological classifications for 304 122 galaxies from the sloan digital sky survey. Monthly Notices of the Royal Astronomical Society, 435(4):2835–2860.
Zhu, X.-P., Dai, J.-M., Bian, C.-J., Chen, Y., Chen, S., and Hu, C. (2019). Galaxy morphology classification with deep convolutional neural networks. Astrophysics and Space Science, 364(4).
