Fine-tuned Vision Transformers applied to 2D spectral plots from real SDSS and LAMOST data achieve higher classification accuracy than SVMs and Random Forests while matching AstroCLIP performance on redshift estimation across diverse objects.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
astro-ph.IM 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Applying Vision Transformers on Spectral Analysis of Astronomical Objects
Fine-tuned Vision Transformers applied to 2D spectral plots from real SDSS and LAMOST data achieve higher classification accuracy than SVMs and Random Forests while matching AstroCLIP performance on redshift estimation across diverse objects.