Full-Reference High Dynamic Range Image Quality Assessment Software Release

The Full-Reference High Dynamic Range Image Quality Assessment 1.0 software release provides several automated visual quality evaluation methods for high-dynamic range (HDR) images. HDR images represent each color value at each pixel as a floating-point number so as to provide a much wider dynamic range of color intensities.

The software release includes the Graph-Based Visual Saliency package ('gbvs'). Please see the readme.txt file in the 'gbvs' folder for details about installation and usage. The present implementation is modified to implement the saliency measure in [1].

This software release supports [2] and provides the following MATLAB files:

1. SHDR_TMQI.m
2. TMQI_NSS_Ent.m
3. TMQI_NSS_Sigma.m
4. run_TMQI_NSS_Sigma.m

Two images have been provided with the release so that the following example code in run_TMQI_NSS_Sigma.m will run:

clc; close all; clear all;
imHDR = hdrread('1.hdr');
imSDR = imread('1.jpg');
Q = TMQI_NSS_Sigma(imHDR, imSDR)

More images are available in the Tone Mapped Quality Index Database [3].

This software release has been tested on MATLAB 2012a and 2013a versions on both OSX (64-bit) and Linux(32-bit) platforms.

References

1. J. Petit, R. Bremond, and J.-P. Tarel, "Saliency maps of high-dynamic range images", Proc. ACM Symp. Appl. Perception in Graphics and Visualization, 2009, pp. 134–134.
2. D. Kundu and B. L. Evans, "Visual Attention Guided Quality Assessment of Tone-Mapped images using Scene Statistics", Proc. IEEE Int. Conf. on Image Processing, Sep. 25-28, 2016.
3. "Tone-Mapped Quality Index Database", The University of Waterloo, Waterloo, Ontario, Canada.