Skip to content
/ MPD Public

Image Quality Assessment: From Human to Machine Preference

Notifications You must be signed in to change notification settings

lcysyzxdxc/MPD

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 

Repository files navigation

Image Quality Assessment: From Human to Machine Preference

Evaluate the subjective preference of machine visual system

Chunyi Li1,2, Yuan Tian1,2, Xiaoyue Ling1,
1Shanghai Jiaotong University, 2Shanghai AI Lab, 3Nanyang Technological University

What images do machines prefer? This remains an open question. Over the past two decades, IQA has witnessed significant development for humans. Conversely, machine preference development is much slower, regardless of an end-to-end IQA algorithm, there is even no database that annotates images with an overall preference score of machines. This hinders the comprehensive evaluation of image processing metrics designed for machine vision.

Therefore, we establish the first large-scale, fine-grained, multi-dimensional quality database for machines, namely MPD. As the visual data consumption volume of machines has already surpassed humans, we hope that MPD can inspire effective IQA metrics to serve the upcoming machine-centric era.

The code and data will be released soon.

About

Image Quality Assessment: From Human to Machine Preference

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published