ꟻLIP: A Difference Evaluator for Alternating Images

Image quality measures are becoming increasingly important in the field of computer graphics. There is, for example, currently a major focus on generating photo-realistic images in real time by combining path tracing with denoising, for which such quality assessment is needed. We present ꟻLIP, which is an image difference technique with a particular focus on rendered images. Our algorithm produces an error map that approximates the difference perceived by humans when alternating between two images. ꟻLIP is a combination of modified existing building blocks, and the net result is surprisingly powerful. We have compared our work against a wide range of existing image difference algorithms and we have visually inspected over a thousand image pairs retrieved from image databases or that were generated in-house. We also present results from a user study which indicate that our method performs substantially better than the others on average. We provide source code in C++, MATLAB, Python, and GPU shader code for our technique.

Paper: coming (undergoing a revision right now)
Link to images and error maps: http://www.240hz.org/image-metrics/
Code: C++, Matlab, shader, pytorch, coming.

Pontus Andersson (NVIDIA)
Magnus Oskarsson (Lund University)
Kalle Åström (Lund University)
Mark D. Fairchild (Rochester Institute of Technologu)
Publication Date: 
Monday, July 13, 2020
Research Area: