Transformation Matrix for 3D computer Graphics Based on FPGA(English)

Abstract

The real time of the computer graphics system performance is one of the fast many computing applications. The 3D (three-dimensional) geometric transformations are one of the most important principles of interactive computer graphics, which are essential for modeling, viewing and animation. This paper tends to construct a general form of a single matrix representation for multiple geometric transformations for threedimensional objects. This way, a speed up factor of 1 to 5 can be gained. Architecture is designed, and implemented as a hardware unit, and then tested for the single matrix transformation, using Field Programmable Gate Array (FPGA).

The acceleration of 3D graphics transformations based on CUDA

“…The transformation operations are highly computationally intensive, as they involve matrix multiplication of trigonometric functions that is applied for each vertex individually. So, the speed of these transformations is the challenge of producing realistic vision of animation scenes (Ali, 2012). This realistic vision requires fast execution of addition, multiplication and trigonometric functions which attracted and still do many research work and different architectures as presented by the following reviews:…”

“…Then they modified this algorithm in Mondal et al (2016) to be able to calculate AT of four voxel locations by performing a single transformation, but these algorithms assumed symmetry about four-pixel locations for the transformed object. Ali (2012) and Kamal and Salim (2013) implemented a design using same FPGA type for object transformation with maximum 10 M and 138 M vertices/s respectively while Sahin (2010) reached to 72 M vertices/s in his design.…”