Research Overview


A valid assessment of image quality is of interest to many. It is of particular interest to manufacturers and consumers of image related devices like televisions, computer monitors, printers, and cameras. Traditional methods of image quality assessment involve human subjects ranking the quality of images as they perceive them. This method makes a lot of sense since humans are the ones who will be using these products. However, these sorts of quality assessments are costly, subjective, and highly variable.

Wouldn't it be nice if we could have a machine that would be able to assess the quality of an image or imaging device in a similar way to humans? A number of current methods exist to do such a task. Unfortunately, they often do not produce the same quality assessment that a human would make. The goal of my research project is to design a system that produces image quality assessments that more closely match the assessments that people would make. We do this by trying to convince a computer to produce reliable estimates of image quality that correspond well with what a person would see. The approach that we are taking is to try to describe mathematically how people perceive things. We can then use a computer to mathematically evaluate the quality of a given image. The mathematically description that we are developing is based on the biology of the human eye and perceptual experiments that we have done in our lab. I have written the same story over again, but have added many more details.

I also have some preliminary results.

More technical details are also available.


t a y l o r@l i n u x a v e.n e t