VIME - Video and Image Models for consumer content Evaluation
VQEG formed the group VIME (Video and Image Models for consumer content Evaluation) to start a body of research to address the challenging area of blind or no-reference perceptual quality evaluation. One of the first areas of focus is the opportunity space of blind image and video quality assessment for the specific application of consumer images and videos that are captured via consumer devices. With the significant changes over the years in how consumers are capturing, manipulating and sharing images and videos, VQEG felt it was time to develop assessment tools for such scenarios. Building on the years of expertise that VQEG has around subjective and objective methodologies, we embark in this direction.
This group's current focus is on mobile phones and tablets, as well as higher end cameras. The group is dedicated to evaluating the quality of these types of consumer images and videos.
The aim is to design and develop perceptual objective methods to evaluate the quality of consumer multimedia visual signals. The goal is quality assessment on realistic distortions, which are typically multiple (multi-distortion) and complex by nature.
To address objective perceptual algorithm design, the group is also involved in subjective testing for the purpose of algorithm design.
In summary, the project addresses:
1) New approaches to subjective study design for the purpose of addressing emerging quality assessment needs (as market and consumer demands evolve).
2) No-reference models for image quality assessment of consumer content with real distortions (as opposed to simulated artificial distortions), including capture and rendering aspects.
3) No-reference models for video quality assessment of consumer content with real distortions, including capture and rendering aspects.
Download VIQET for no-reference photo quality evaluation!
The VQEG Image Quality Evaluation Tool (VIQET) is an objective, no-reference photo quality evaluation tool. VIQET is a free open source tool designed to evaluate quality of consumer photos. In order to perform photo quality evaluation, VIQET requires a set of photos from the test device. It estimates an overall Mean Opinion Score (MOS) for a device based on the individual image MOS scores in the set. The estimated MOS by VIQET falls in a range of 1 to 5, where 1 corresponds to a low quality rating and 5 corresponds to excellent quality.
- VIQET is an open source project that is available at www.GitHub.com/VIQET.
- The desktop tool installer can be downloaded at: https://github.com/VIQET/VIQET-Desktop/releases
- The source code can be found at: https://github.com/VIQET/VIQET-Desktop
Contribute images to the VIME project
The VIME (Video and Image Models for consumer content Evaluation of) project in VQEG is dedicated to no-reference image quality and is currently collecting images to build a dataset that can be used to conduct our research work.
- Instructions on how to: https://www.linkedin.com/grp/post/5100193-6009891915024121858
- VIME Flickr site: https://www.flickr.com/groups/vime/