Most interest point detection algorithms are highly sensitive to illumination changes. We have developed a method to find interest points robustly even under large photometric changes. The method, which we call contrast invariant feature transform (CIFT), determines salient interest points in an image by calculating and analyzing contrast signatures. A contrast signature shows the response of an interest point detector with respect to a set of contrast stretching functions. The CIFT is generic and can be used with most interest point detectors.

Related Publications:

Reliable interest point detection under large illumination variations, Murat Gevrekci and Bahadir K. Gunturk, to appear in ICIP 2008. [pdf]

On geometric and photometric registration of images, Murat Gevrekci,  Bahadir K. Gunturk, IEEE Int. Conf. on Acoustics, Speech, Signal Processing (ICASSP), 2007. [pdf]

Sample Results:   

The following example compares the repeatability rates of several interest point detectors.

Figure 1. Image in (d) is chosen as reference. Repeatability rates between this image and the other images are computed. The images in (a),(b),(c),(e),(f), and (g) are labeled as 1 to 6 in the following figure.

Figure 2. Repeatability rates for several algorithms are plotted. The CIFT technique improves the Harris detector to outperform the other algorithms.