Computing WHERE-WHAT Classification Through FLIKM and Deep Learning Algorithms

Object recognition has gained substantial importance in the artificial intelligence technology to visualize, understand and to take a decision. Many complex algorithms achieve AI utilities. The application of these object recognition methods extends from the medical field to wide variety of other sectors, especially industries. As per the literature survey, it defines that many systems or algorithms are available to process the recognition, but accuracy toward invariant 2D data with scaling, translation, rotation and partial properties is not up to the mark. In this paper, an efficient, systematic approach is developed to understand the place or a situation of a particular scene. Optimized deep learning in association with a clustering algorithm (FLIKM) is used in this approach to build an intelligent self-surveillance system. The experimental results of the state-of-the-art system evaluated with various benchmark datasets show an acceptable accuracy rate of 94 approximately.