Image Classification of Ten Challenging Dog Breeds

Imagewoof is a dataset for image classification containing a subset of 10 challenging dog breed classes from ImageNet. The breeds include Australian terrier, Border terrier, Samoyed, Beagle, Shih-Tzu, English foxhound, Rhodesian ridgeback, Dingo, Golden retriever, and Old English sheepdog. It was created by author frgfm and last updated in December 2022.

Use Cases

• Train a convolutional neural network to classify images into the 10 challenging dog breed categories.
• Benchmark model performance on fine-grained visual categorization tasks using the specific breed labels.
• Use the dataset for transfer learning, fine-tuning a pre-trained model on the 10 dog breed classes.
• Analyze model confusion between visually similar breeds like the Australian terrier and Border terrier.

Strengths

• Contains 10 challenging dog breed classes selected for difficult classification.
• Derived from the authoritative ImageNet dataset, ensuring a foundational source.
• Last updated in December 2022, indicating recent maintenance.

Limitations

• The specific number of images (rows) is unknown, making scale assessment difficult.
• Class distribution and potential label imbalances from crowdsourced annotations are unspecified.
• Geographic origin is indicated as US ('Regionus'), which may introduce a sampling bias.

Provenance

Source

Subset of ImageNet.

Collection Method

Crowdsourced annotations.

Freshness

2022-12-11

Geography

United States (inferred from 'Regionus' tag).