This data release supports an AI-ready pipeline for Euclid VIS galaxy analysis by providing standardized galaxy image cutouts, curated morphology annotations, and precomputed representation vectors derived from the Galaxy Zoo Euclid (Q1) morphology catalog. The dataset includes 365,513 galaxy-centered FITS images, standardized to 224 × 224 pixels. The image cutouts are generated using an automated coordinate-based extraction pipeline applied to Euclid VIS imaging data. Minimal preprocessing is performed to preserve the intrinsic morphological structures of galaxies. Due to the large number of files, the FITS cutouts are distributed as multiple compressed archives for efficient download and storage. A manifest and checksum file are provided to ensure consistent mapping between images, annotations, and embeddings, as well as data integrity.

In addition to the image data, the dataset includes a curated morphology catalog in both FITS and CSV formats, containing high-confidence annotations derived from probabilistic labels. To facilitate direct use in machine learning workflows, we provide DINOv2-based embedding representations stored as a NumPy array of shape (365513, 384), where each row corresponds to a galaxy image.

 The original Galaxy Zoo Euclid (Q1) morphology catalog is an external data product. This release provides derived image cutouts, curated annotations, and learned representations for downstream analysis. This data product is released to ensure reproducibility of the proposed AI-ready pipeline and to facilitate its reuse in large-scale Euclid imaging analysis within modern machine learning frameworks, including morphology analysis, representation learning, anomaly detection, and similarity-based retrieval.