Histogram based similarity control with exposurenormalization of 3D objects

Recent technological achievements and studies show that extracting and interpreting the properties of 3D objects become important and be handled effectively and efficiently. In medicine, engineering, industry, and even archaeology, it has become a necessity to analyse the encountered 3D objects, even in embedded platforms, hence effective representations and efficient similarity computation techniques need to be developed. Most of the existing techniques for defining the shape content of 3D objects have their own limitations and short comings. Here, we presents simple yet low-complexity histogram-based similarity model, where the employed histograms were shown to be effective in 2D. The extension of these histograms to 3D with appropriate normalization steps is presented along with the radial serialization scheme to alleviate the integrated use. These histograms are invariant under rotation, scaling, and translation. We created a dataset with challenging modifications of each set of objects to verify the effectiveness of the proposed technique. Moreover, the performance experiments on Princeton Shape Benchmark show that the findings are reasonable when compared to the literature.