Skip to Main Content
Moments can uniquely describe an image. Moment invariants with stability to geometric transformation can be used as features to recognize images despite geometric transformations. However, moment invariants lose stability when gray-levels are adjusted. Usually image processing may cause adjusting of gray-level, which limits application of moment invariants. This paper analyses the relation between invariant moments and gray-level adjustment, and concludes that gray-level histogram equalization can enhance the stability of invariant moments effectively. The conclusion is well proved by experiments.