Published On: Sun, Jun 26th, 2022

The ArtBench Dataset: Benchmarking Generative Models with Artworks


Deep generative models can synthesize diverse and high-fidelity images. Computational understanding of art attracts more and more attention because of its importance for art history, computational creativity and human-computer interaction. A recent paper on arXiv.org proposes ArtBench-10, the first class-balanced, high-quality, cleanly-annotated, and standardized benchmark for artworks synthesis.

The new research proposes the idea to use art for the purposes of benchmarking generative AI models. Pictured: artwork titled Robosophy Philosophy by Predrag K. Nikolic.

The new research proposes the idea to use art for the purposes of benchmarking generative AI models. Pictured: artwork titled Robosophy Philosophy by Predrag K. Nikolic. Image credit: Ars Electronica via Flickr, CC BY-NC-ND 2.0

The dataset is composed of 60,000 images annotated with 10 artistic styles such as Baroque or Surrealism. The images are of high-quality with clean and balanced labels and can be easily incorporated in commonly used deep learning frameworks.

Dataset statistics analysis demonstrates the advantages of the dataset. It addresses problems in previous datasets such as label imbalance, near-duplicates, noisy labels, and poor image quality.

We introduce ArtBench-10, the first class-balanced, high-quality, cleanly annotated, and standardized dataset for benchmarking artwork generation. It comprises 60,000 images of artwork from 10 distinctive artistic styles, with 5,000 training images and 1,000 testing images per style. ArtBench-10 has several advantages over previous artwork datasets. Firstly, it is class-balanced while most previous artwork datasets suffer from the long tail class distributions. Secondly, the images are of high quality with clean annotations. Thirdly, ArtBench-10 is created with standardized data collection, annotation, filtering, and preprocessing procedures. We provide three versions of the dataset with different resolutions (32×32, 256×256, and original image size), formatted in a way that is easy to be incorporated by popular machine learning frameworks. We also conduct extensive benchmarking experiments using representative image synthesis models with ArtBench-10 and present in-depth analysis. The dataset is available at this https URL under a Fair Use license.

Research article: Liao, P., Li, X., Liu, X., and Keutzer, K., “The ArtBench Dataset: Benchmarking Generative Models with Artworks”, 2022. Link: https://arxiv.org/abs/2206.11404




Source link

Most Popular News

Local Business Directory, Search Engine Submission & SEO Tools