Imagenet-1K-VL-Enriched

An enriched version of the ImageNet-1K Dataset with image caption, bounding boxes, and label issues! With this additional information, the ImageNet-1K dataset can be extended to various tasks such as image retrieval or visual question answering.

The label issues helps to curate a cleaner and leaner dataset.

Description

The dataset consists of 6 columns:

• image_id: The original filename of the image from ImageNet-1K dataset.
• image: Image data in the form of PIL Image.
• label: Label for the image, provided by the authors of the ImageNet-1K dataset.
• label_bbox_enriched: Enriched labels for the image. Consists of bounding box coordinates, confidence scores, and labels for the bounding box generated using object detection models.
• caption_enriched: Enriched captions for the image. Generated by BLIP2 captioning model.
• issues: Quality issues found such as duplicate, mislabeled, dark, blurry, bright, and outlier images.

Usage

This dataset can be used with the Hugging Face Datasets library.:

import datasets
ds = datasets.load_dataset("visual-layer/imagenet-1k-vl-enriched")

More in this notebook.

Interactive Visualization

Visual Layer provides a platform to interactively visualize a dataset and highlight quality issues such as duplicates, mislabels, outliers, etc. Check it out here. No sign-up required.

License & Disclaimer

We provide no warranty on the dataset, and the user takes full responsibility for the usage of the dataset. By using the dataset, you agree to the terms of the ImageNet-1K dataset license.

About Visual Layer

Copyright © 2024 Visual Layer. All rights reserved.