DATASET
Open Source Community
CBSD68-dataset
The CBSD68 dataset is a color version of the BSD68 benchmark for image denoising, containing original .jpg files converted to lossless .png format and augmented with different levels of Gaussian white noise.
Updated 5/23/2024
github
Description
CBSD68‑dataset Overview
Dataset Description
- Name: CBSD68‑dataset
- Purpose: Image denoising benchmark
- Contents: Original
.jpgfiles converted to lossless.pngformat, with added Gaussian white noise at various levels
Source
- Affiliation: Part of The Berkeley Segmentation Dataset and Benchmark
- Link: https://www2.eecs.berkeley.edu/Research/Projects/CS/vision/bsds/
Authors
- Authors: D. Martin, C. Fowlkes, D. Tal, J. Malik
- Paper Title: A Database of Human Segmented Natural Images and its Application to Evaluating Segmentation Algorithms and Measuring Ecological Statistics
- Publication:
- Conference: Proc. 8th Intl Conf. Computer Vision
- Year: 2001
- Month: July
- Volume: 2
- Pages: 416–423
AI studio
Generate PPTs instantly with Nano Banana Pro.
Generate PPT NowAccess Dataset
Login to Access
Please login to view download links and access full dataset details.
Topics
Image Denoising
Benchmarking
Source
Organization: github
Created: 7/26/2018
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.