Dataset Preview
Full Screen
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
[Errno 13] Permission denied: '/tmp/hf-datasets-cache/medium/datasets/97467667511723-config-parquet-and-info-Voxel51-DUTS-0afaa33d/downloads/46ab68e45bc7e5041a184c36d9a6b012f28c96465c4281b0e69a08d574738ad8.incomplete'
Error code:   UnexpectedError

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

image
image
End of preview.

Dataset Card for DUTS

image/png

This is a FiftyOne dataset with 15572 samples.

Installation

If you haven't already, install FiftyOne:

pip install -U fiftyone

Usage

import fiftyone as fo
import fiftyone.utils.huggingface as fouh

# Load the dataset
# Note: other available arguments include 'max_samples', etc
dataset = fouh.load_from_hub("Voxel51/DUTS")

# Launch the App
session = fo.launch_app(dataset)

Dataset Details

Dataset Description

DUTS is a saliency detection dataset containing 10,553 training images and 5,019 test images. All training images are collected from the ImageNet DET training/val sets, while test images are collected from the ImageNet DET test set and the SUN data set. Both the training and test set contain very challenging scenarios for saliency detection. Accurate pixel-level ground truths are manually annotated by 50 subjects.

  • Curated by: Lijun Wang, Huchuan Lu, Yifan Wang, Mengyang Feng, Dong Wang, Baocai Yin, and Xiang Ruan
  • Language(s) (NLP): en
  • License: unknown

Dataset Structure

Name:        DUTS
Media type:  image
Num samples: 15572
Persistent:  False
Tags:        []
Sample fields:
    id:           fiftyone.core.fields.ObjectIdField
    filepath:     fiftyone.core.fields.StringField
    tags:         fiftyone.core.fields.ListField(fiftyone.core.fields.StringField)
    metadata:     fiftyone.core.fields.EmbeddedDocumentField(fiftyone.core.metadata.ImageMetadata)
    ground_truth: fiftyone.core.fields.EmbeddedDocumentField(fiftyone.core.labels.Segmentation)

The dataset has 2 splits: "train" and "test". Samples are tagged with their split.

Dataset Creation

Introduced by Wang et al. in Learning to Detect Salient Objects With Image-Level Supervision

Citation

BibTeX:

@inproceedings{wang2017,
title={Learning to Detect Salient Objects with Image-level Supervision},
author={Wang, Lijun and Lu, Huchuan and Wang, Yifan and Feng, Mengyang 
and Wang, Dong, and Yin, Baocai and Ruan, Xiang}, 
booktitle={CVPR},
year={2017}
}

Dataset Card Authors

Dataset conversion and data card contributed by Rohith Raj Srinivasan.

Dataset Card Contact

Rohith Raj Srinivasan

Downloads last month
234