DRIQ (Digitally Retouched Image Quality) database is a full-reference enhanced-image database. This database consists of 104 images in which 78 images were digitally retouched from 26 original images with different contents. Subjective ratings of these images were obtained from 9 human observers. The ability to quantify the visual quality of an image in a manner that agrees with human subjective rating is an important step in many applications, e.g., image compression, image denoising, or image retouching. Researchers in the past several decades have proposed a variety of computational methods for image quality assessment (IQA). A major portion of those methods focus on full-reference IQA problem which assumes the availability of the original image and thus outputs an index that represents the visual quality of the target (modified) image relative to the original image. It is important to note that almost all current full-reference IQA methods were specifically designed for and tested on degraded images. However, in many image processing applications, the target image is actually enhanced from the original image and thus has better visual quality. It is generally impossible for current IQA algorithms to indicate whether the target image has better or worse quality than the original image.
ZIP archives, no password. Links: Subjective rating: http://vision.okstate.edu/driq/DRIQ_DMOS.xlsx Images: http://vision.okstate.edu/driq/DRIQ.zip
The original link was: http://vision.okstate.edu/index.php?loc=driq
References and Citation
VPC14: Cuong Vu, Thien Phan, and Damon Chandler, Can Current Image Quality Assessment Algorithms Predict Visual Quality of Enhanced Images?, preprint, 2014.