Overlapping Histograms for 3D Data Classification

3D object scanning and recognition is becoming more widely available with the rapid advancement of technology. In this paper, we developed a 3 Point Histogram Comparison algorithm for improving 3D image classification accuracy. In our research, we used the Structure Sensor by Occipital, an inexpensive alternative to state-of-the-art scanners. We also built and refined an extensive 3D object reference library. Results show that we improved the accuracy in about 10-15% in comparison to previous methods, as well as shorten the computation time significantly.