JUHE API Marketplace
DATASET
Open Source Community

ccHarmony

The ccHarmony dataset is a color‑checker‑based image harmonization dataset designed to better reflect natural illumination variations while maintaining a distribution similar to the Hday2night dataset, but at a lower collection cost. It comprises 350 real images and 426 segmented foregrounds; each foreground is paired with 10 synthetic composite images, resulting in a total of 4,260 synthetic‑real image pairs.

Updated 5/12/2024
github

Description

Dataset Overview

Dataset Name

ccHarmony

Dataset Characteristics

  • Image harmonization dataset based on a color checker (cc).
  • By transferring foregrounds from real images to a standard illumination condition and then to another illumination condition, synthetic composite images are generated.
  • Contains 4,260 synthetic‑real image pairs, divided into 3,080 training pairs and 1,180 testing pairs.

Dataset Content

  • 350 real images.
  • 426 segmented foregrounds, each associated with 10 synthetic composite images.

Dataset Download

  • Baidu Cloud: link (access code: bulf)
  • OneDrive: link

Experimental Results

  • Multiple image harmonization methods were evaluated, including DoveNet, RainNet, IIH, etc.
  • Results show that our proposed GiftNet achieves the best performance on MSE, fMSE, PSNR, and fSSIM metrics.

Dataset Examples

  • Several real images and their corresponding synthetic composites are displayed.

Dataset Applications

  • Suitable for research and testing of image harmonization methods.
  • Supports dataset augmentation via SycoNet to generate high‑quality composite images.

AI studio

Generate PPTs instantly with Nano Banana Pro.

Generate PPT Now

Access Dataset

Login to Access

Please login to view download links and access full dataset details.

Topics

Image Processing
Illumination Variation

Source

Organization: github

Created: 5/31/2022

Power Your Data Analysis with Premium AI Models

Supporting GPT-5, Claude-4, DeepSeek v3, Gemini and more.

Enjoy a free trial and save 20%+ compared to official pricing.