The 'human-robot interface' or HRI provides multiple assistance services in a variety of real-time applications. In robotic devices, the notion of object classification by digital visualisation is predicated on the converging of a 'three-dimensional (3D)' image into a plane-based representation. During the convergence process, the projectors in multiple planes are mistaken, resulting in identification mistakes. The information processing approach, which is based on the result showed, can lessen these misidentifications in object identification. The combining indexes are found by seeing and extending the input image in all conceivable directions. A machine learning technique is utilised to ways that improve processing precision and agility. This research paper aims to evaluate the detailed analysis of 3D image processing by using machine learning for man machine interaction. In this context, secondary analysis is to be done by taking information from different journals and articles. Google scholar, ProQuest databases are used to get relevant and proper journals related to topic.