Cifar10Mnist: Synthetic Images with MNIST Digits Overlaid on CIFAR-10

Cifar10Mnist is a synthetic image dataset created by overlaying MNIST digit images on top of CIFAR-10 images. Each example contains a 32x32 RGB image and a paired label tuple: the original CIFAR-10 class name plus the MNIST digit label. The dataset was created by author 'shotegni' and was last updated on 2026-06-03.

Use Cases

• Multi-label image classification based on the paired CIFAR-10 and MNIST labels.
• Transfer learning research based on the synthetic combination of two distinct image domains.
• Synthetic data research based on the described method of overlaying images.
• Computer vision model robustness testing based on the composite image structure.

Strengths

• Each image is paired with a label tuple containing both a CIFAR-10 class and an MNIST digit.
• Images are a synthetic composite of two well-known benchmark datasets.
• Dataset was last updated on 2026-06-03.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

huggingface

Collection Method

Synthetically created by overlaying MNIST digit images on CIFAR-10 images.

Freshness

Last updated 2026-06-03 22:42:34; freshness should be verified.