IPD / README.md
jigarsiddhpura's picture
dataset uploaded by roboflow2huggingface package
4aec081 verified
metadata
task_categories:
  - object-detection
tags:
  - roboflow
  - roboflow2huggingface
jigarsiddhpura/IPD

Dataset Labels

['dry-person', 'object', 'wet-swimmer']

Number of Images

{'test': 77, 'valid': 153, 'train': 1608}

How to Use

pip install datasets
  • Load the dataset:
from datasets import load_dataset

ds = load_dataset("jigarsiddhpura/IPD", name="full")
example = ds['train'][0]

Roboflow Dataset Page

https://universe.roboflow.com/resq/tiny-people-detection-rpi/dataset/1

Citation

@misc{ tiny-people-detection-rpi_dataset,
    title = { Tiny people detection RPI Dataset },
    type = { Open Source Dataset },
    author = { ResQ },
    howpublished = { \\url{ https://universe.roboflow.com/resq/tiny-people-detection-rpi } },
    url = { https://universe.roboflow.com/resq/tiny-people-detection-rpi },
    journal = { Roboflow Universe },
    publisher = { Roboflow },
    year = { 2023 },
    month = { sep },
    note = { visited on 2024-02-11 },
}

License

CC BY 4.0

Dataset Summary

This dataset was exported via roboflow.com on February 10, 2024 at 7:28 AM GMT

Roboflow is an end-to-end computer vision platform that helps you

  • collaborate with your team on computer vision projects
  • collect & organize images
  • understand and search unstructured image data
  • annotate, and create datasets
  • export, train, and deploy computer vision models
  • use active learning to improve your dataset over time

For state of the art Computer Vision training notebooks you can use with this dataset, visit https://github.com/roboflow/notebooks

To find over 100k other datasets and pre-trained models, visit https://universe.roboflow.com

The dataset includes 1838 images. People are annotated in COCO format.

The following pre-processing was applied to each image:

  • Auto-orientation of pixel data (with EXIF-orientation stripping)
  • Resize to 640x640 (Stretch)

The following augmentation was applied to create 3 versions of each source image:

  • Randomly crop between 0 and 67 percent of the image
  • Salt and pepper noise was applied to 4 percent of pixels

The following transformations were applied to the bounding boxes of each image:

  • Random shear of between -5° to +5° horizontally and -5° to +5° vertically