dataset_info:
features:
- name: image
dtype: image
- name: label
dtype:
class_label:
names:
'0': no_duo
'1': duo
splits:
- name: original
num_bytes: 4440451
num_examples: 40
- name: augmented
num_bytes: 25195032
num_examples: 300
download_size: 28523872
dataset_size: 29635483
configs:
- config_name: default
data_files:
- split: original
path: data/original-*
- split: augmented
path: data/augmented-*
pretty_name: duo or no duo
Dataset Card for Duo or no Duo
Dataset Details
Photos in this dataset were taken by Scotty McGee on an iPhone 15 and converted from HEIC to RGB and resized to 224 × 224 pixels for consistency
Dataset Description
This dataset consists of images paired with a binary label indicating whether “Duo” is present (duo) or absent (no_duo), making it a straightforward two-class image classification task. All images converted from HEIC to RGB and resized to 224 × 224 pixels for consistency. To expand the dataset, over 300 synthetic images were created by applying label-preserving transformations such as random cropping and resizing, horizontal flips, small rotations, color jitter, perspective distortion, and slight blurring to the original images.
Dataset Use
This dataset is intended to test if you can identify if an image has Duo or not.
- Curated by: Scotty McGee