Using spectral or syntactic domain pattern recognition algorithms, objects are recognised in digital photographs. In order to recognise items that would otherwise go undetected and latent, quicker, more accurate, and smarter pattern recognition techniques must be developed due to the constantly growing volume of data gathered by digital picture collecting systems. Using cryptic learning neural networks for object recognition is one such endeavour. Knowledge that was gathered from the outlines of different objects that were present in a digital picture served as the input to this system. In order to recognise things using a neural network and extract information about the delineations of different items and constituentsof a digital picture, this study proposes a unique technique.