Local Image Sharpness Database

The CSIQ image sharpness database is a ground-truth dataset consisting of local perceived sharpness ratings in images. The database is provided in the form of original images and corresponding ground-truth "sharpness maps," in which each value of each map denotes the average subjective rating of sharpness for the corresponding location in the corresponding image. This database is ideal for testing sharpness estimation algorithms.

If you use this database, please cite: C. T. Vu, and T. D. Phan, and D. M. Chandler, "S3 A Spectral and Spatial Measure of Local Perceived Sharpness in Natural Images," IEEE Transactions on Image Processing, 21 (3), 2011.


For more information on this database, including details on the experimental protocol, please see our project page on the S3 sharpness estimation algorithm.

Last modified: Friday, April 29, 2016, 8:55 AM