3D Image Processing using Machine Learning based Input Processing for Man-Machine Interaction

摘要

In various real time applications, several assisted services are provided by the human-robot interaction (HRI). The concept of convergence of a three-dimensional (3D) image into a plane-based projection is used for object identification via digital visualization in robotic systems. Recognition errors occur as the projections in various planes are misidentified during the convergence process. These misidentifications in recognition of objects can be reduced by input processing scheme dependent on the projection technique. The conjoining indices are identified by projecting the input image in all possible dimensions and visualizing it. Machine learning algorithm is used for improving the processing speed and accuracy of recognition. Labeled analysis is used for segregation of the intersection without conjoined indices. Errors are prevented by identifying the non-correlating indices in the projections of possible dimension. The inputs are correlated with related inputs that are stored with labels thereby preventing matching of the indices and deviations in the planes. Error, complexity, time and recognition ratio metrics are verified for the proposed model.