Sketchy Content Image for Style Transfer

Citation Author(s):
yinqi
chen
Submitted by:
Yinqi Chen
Last updated:
Mon, 05/06/2024 - 02:08
DOI:
10.21227/t92m-y870
Data Format:
License:
0
0 ratings - Please login to submit your rating.

Abstract 

The Sketchy images refer to hand-drawn drawings, while SCIST are those with unclear or weak semantic information, represent a distinctive cases from natural scenes.The primary objective of this dataset is to facilitate the style transfer, whether originating from manual sketches or digital renderings, into enriched and artistically embellished counterparts through the utilization of software.

Instructions: 

To foster increased interest in exploring sketchy images, we introduce the Sketchy Content Image for Style Transfer dataset (SCIST). This dataset comprises 2,051 distinct content images with a resolution of  sourced from various outlets, including children's hand-drawn creations and AI-generated images. The majority of these images consist of graffiti without clear semantic information. The primary objective of this dataset is to facilitate the style transfer, whether originating from manual sketches or digital renderings, into enriched and artistically embellished counterparts through the utilization of software.