Dataset · Voxel51
Indoor scene recognition images across 67 categories
Voxel51/IndoorSceneRecognition
A collection of 15,620 interior photos annotated across 67 categories in FiftyOne format.
What's inside
It contains 15,620 JPG photos distributed across 67 different indoor categories. Each class features at least 100 examples, though the total number of images varies per category. The dataset is prepared for easy loading and use through the FiftyOne library.
What you can do with it
- Training scene classification and indoor type recognition models
- Testing computer vision algorithms for home robot navigation
- Fine-tuning visual encoders for real estate analysis tasks
- Tasks
- Image classification
- Languages
- English
- Size
- up to 1K rows
- Downloads this month
- 21.2K
- Likes
- 4
How to load it
from datasets import load_dataset
ds = load_dataset("Voxel51/IndoorSceneRecognition")