Deep learning based data-driven approaches have been successfully applied in various image under- standing applications ranging from object recognition, semantic segmentation to visual question an- swering. However, the lack of knowledge integration as well as higher-level reasoning capabilities with the methods still pose a hindrance. In this work, we present a brief survey of a few represen- tative reasoning mechanisms, knowledge integration methods and their corresponding image under- standing applications developed by various groups of researchers, approaching the problem from a va- riety of angles. Furthermore, we discuss upon key efforts on integrating external knowledge with neu- ral networks. Taking cues from these efforts, we conclude by discussing potential pathways to im- prove reasoning capabilities.