Ghost imaging (GI) has earned novelty in image processing technique due to properties of non-locality, simple structure setups, high detection efficiency and imaging in scattering media. GI has stepped from basic to real time applications with advance methodologies. To get more accuracy in target recognition, optical arrangements with traditional methods showed some limitations which need resolution. Advance methods with deep learning in GI could be a better solution with fewer limitations. In this review, we summarize the target recognition methodologies based on traditional and advanced and compare them with challenges and future suggestions for more evolving this innovative research direction.<p></p>