metadata

license: apache-2.0
task_categories:
  - image-classification
language:
  - en
tags:
  - Scene-Detection
  - buildings
  - forest
  - glacier
  - mountain
  - sea
  - street
  - climate
size_categories:
  - 10K<n<100K

OpenScene-Classification Dataset

A high-quality image classification dataset curated for scene detection tasks, particularly useful in training and evaluating models for recognizing various natural and man-made environments.

Dataset Summary

The OpenScene-Classification dataset contains labeled images categorized into six distinct scene types:

buildings
forest
glacier
mountain
sea
street

This dataset is structured for supervised image classification, suitable for deep learning models aiming to identify and classify real-world scenes.

Dataset Structure

Split: train (currently only one split)
Format: parquet
Modality: Image
Labels Type: Integer class indices with corresponding string names
License: Apache-2.0

Each entry in the dataset includes:

image: the image of the scene
label: the class index (e.g., 0 for buildings)
label_name: the class name (e.g., "buildings")

Note: The dataset viewer on Hugging Face may take a moment to load all samples.

Label Mapping

Class Index	Label
0	buildings
1	forest
2	glacier
3	mountain
4	sea
5	street

Dataset Stats

Size: 10K - 100K images
Language: English (tags, metadata)
Tags: Scene-Detection, buildings, forest, glacier, mountain, sea, street

Intended Use

This dataset is ideal for:

Scene classification model training
Benchmarking computer vision algorithms
Educational purposes in machine learning and AI