List of Codes
(back to main page)
The GPU-Efficient Recursive Filtering and Summed-Area Tables (gpufilter) project is a set of C for CUDA functions to compute recursive filters and summed-area tables in GPUs. This project presents a new algorithmic framework for parallel evaluation. It partitions the image into 2D blocks, with a small band of data buffered along each block perimeter. A remarkable result is that the image data is read only twice and written just once, independent of image size, and thus total memory bandwidth is reduced even compared to the traditional serial algorithm.
The ZSIG library is a compound of C++-source codes to compute vertex signatures based on Zernike moments. Each vertex signature is a set of Zernike coefficients embedding the corresponding vertex in feature space. This space has the interesting property of grouping similar vertices, that is similarity-based processing is reduced to searching nearest neighbours in this space.
The a48 library is a template-based C++ library to support general mesh representations mainly in the Computer Graphics research field. The library provides three principal components: efficient polygonal mesh representation; mesh manipulation via stellar operations; and adaptive multi-resolution structure with simplification and refinement strategies.
Original and Improved version of the Projected Tetrahedra with Partial Pre-Integration algorithm, the main theme of my M.Sc. dissertation and part of my D.Sc. thesis. This package contains two programs for Scientific Visualization, aiming at generating images from volumetric datasets. The research area of Volume Rendering is mainly used for visualization of geological data, fluid simulation and inspection of medical images.
Regular datasets version of Projected Tetrahedra with Partial Pre-Integration, developed in the beginning of my D.Sc. thesis. This application is the same basic idea of PTINT but specialized for regular datasets. This type of volumetric dataset appears normally in medical images.
The Hardware-Assisted Projected Tetrahedra is the last adaptation of the Projected Tetrahedra algorithm to the GPU in my D.Sc. thesis, following the previous PTINT versions. HAPT provides a flexible framework capable of extracting iso-surfaces and handling time-varying data on-the-fly.
The Visible-Faces driven Ray-Casting is a memory-efficient GPU-based ray-casting method. The main idea is to explore ray coherence, improving caching performance by keeping in memory only the face data of the traversals of a set of nearby rays.