Abstract
Automatic object detection from a large-scale database has become an established discipline in the field of pattern recognition. It becomes difficult to retrieve the relevant image when the images deformed by some geometric deformation. Moments and their invariants are very useful for invariant feature extraction due to their rotation, scale, and translation invariance properties. In the present study, we utilize Hu's moments to address the issues of geometric transformation. Three experiments are conducted i.e. scale invariant, rotation invariant, and combined rotation and scale invariance