Disclosed subject matter relates to image processing that includes a method for marking content on a surface of an object using laser. A transformation identification computing device receives an input image of the object and extracts data corresponding to region of interest in the input image. The extracted data is compared with extracted template data corresponding to regions of interest present in template images of one of one or more template objects. A transformation in position of the object with respect to the position of the one of the one or more template objects is determined, based on the comparison. Finally, an inverse of the transformation is applied to content data that is to be marked at a desired location within the region of interest of the object. The present disclosure analyses the images at a sub-pixel level based on machine learning approach to determine transformation.