Introduction to 3d Content-Based Search on 3d Krawtchouk Moments Seminar Topic:
This white paper discussed about the methods of “3D Content-based Search” tool and pose normalization method. Also mentioned how 3D content based search tool developed based on research work done by Krawtchouk in image processing. 3D analysis which was introduced by weighted 3D Krawtchouk moments was more suitable for content based search applications.
Extraction of Krawtchouk Descriptors:
Weighted 3D Krawtchouk Moments are introduced in the form of Weighted Krawtchouk polynomials. When a 3D object is our input, the weighted 3D Krawtchouk moments are calculated and used as a descriptor vector. A description of a 3D object in terms of a significant discriminative descriptor vector is obtained. The descriptor extraction is too fast and the matching process can be completed in seconds. For each query 3D model, it requires a position normalization step and preprocessing pose.
Weighted 3D Krawtchouk moments can be presented as a function defined in a discrete space and can used as 3D object descriptor. The proposed method was verified for its performance on 3D content-based search application with the help of Princeton Shape Benchmark Database. To find the ability of the proposed method in order to discriminate between objects classes, every 3D model was explicitly used as a query object. Weighted 3D Krawtchouk Moments efficiency in order to capture the object edges can be represented using Princeton Shape Benchmark. Relatively more spatial frequency components are acquired by weighted 3D Krawtchouk moments and Krawtchouk polynomials of low order. So, weighted 3D Krawtchouk moments can able to capture sharp shape object changes.
Conclusions:
Weighted Krawtchouk polynomials are defined over discrete field, weighted 3D Krawtchouk
Moments are based on these polynomials. Error cannot be inserted when moment computation takes place because of discretization. The proposed method in this paper is depends on the Weighted 3D Krawtchouk moments, which can form a highly compact and discriminative descriptor vector, due to their ability to capture sharp changes with low order moments.
Download 3d Content-Based Search on 3d Krawtchouk Moments .