jennifee/HW1-image-dataset

Dataset Card for Book Cover Images

Dataset of fiction and nonfiction book covers. In this case, we aim to judge a book by its cover.

Dataset Details

Dataset Description

For a selection of books on my bookshelf, I collected some image data. I selected 15 fiction and 15 nonfiction books. I then took a picture of the cover of each book. These variables were collected for the 30 books that make up my original split, and then they were augmented to create 330 additional examples.

• Curated by: Jennifer Evans
• Language(s) (NLP): English
• License: MIT

Uses

Direct Use

The aim of this dataset is to evaluate if you can identify if a book is fiction or nonfiction based on the cover art.

Out-of-Scope Use

This dataset could be used for other evaluations, like what book cover details or art styles are popular.

Dataset Structure

• dataset_info:

• features:

name: CoverImage

dtype: image

name: Fiction?

dtype: string

• splits:

name: train

num_bytes: 2881963.0

num_examples: 30

name: augmented

num_bytes: 26808163.0

num_examples: 330

download_size: 29691254

dataset_size: 29690126.0

• configs:

config_name: default

• data_files:

split: train

path: data/train

plit: augmented

path: data/augmented

Dataset Creation

Curation Rationale

The motivation for developing this dataset is to assess if it is possible to identify if a book is fiction or nonfiction based on the cover. In particular, the selected fonts, iconography, and colors may indicate what type of book it is.

Source Data

30 books were selected from my bookshelf to be photographed and labeled as fiction or nonfiction.

Data Collection and Processing

The selected books were photographed with my iPhone 15 Pro. They were then cropped into a centered square. These 30 covers were then used to create the augmented dataset.

Who are the source data producers?

Jennifer Evans photographed the book covers. Each book cover was developed with the book authors and publication houses.