Depth Image Object Recognition using Moment Invariants

Abstract

Moment invariants have been widely used in various applications in image pattern recognition research field due to its invariant features of translation, scaling and rotation. In this paper, a computer vision based ambient assisted living technique for drinking activity assistance by using moment invariant to recognize the depth image object was developed. The extracted depth image objects dataset contains different type of mugs with different views. We employed the invariant features on the extracted depth image objects to recognize and classify the objects into different categories. Experimental result shows that higher accuracy can be achieve with the proposed technique compared to existing methods on depth image objects recognition.