GEOMETRIC ATTRIBUTES VIZUALIZATION IN VOXEL SPACE

V magistrskem delu opišemo postopek upodabljanja ter analize geometrijskih atributov vokselskih podatkov. Upodabljanje realiziramo z algoritmom metanja žarka(angl. ray-casting), ki skozi vokselski prostor pošlje žarek. Končno vrednost posameznega piksla prikaže s pomočjo funkcije prikaza ERT, XRAY, MIP ali LMIP. Ker se najzahtevnejši računski del za več pikslov hkrati izvaja na gračni kartici, dosežemo upodobitev v realnem času. Drugi del magistrskega dela opisuje postopek analize vokselskih podatkov, ki jo izvedemo s profifili atributnih razlik, kjer posamezen profifil določa razlika dveh zaporednih atributnih filtriranj. Izhod iz analize prikažemo z modeloma CSL in ABV, ki za predstavitev uporabljata barvna prostora HSV in CIELab LCH.In this work, we describe a visualization and geometric attribute analysis procedure for voxel data. A ray-casting algorithm is used for visualization where for each pixel, a ray is shot through the volume data. The fifinal image is obtained by using one of many available rendering methods like ERT, XRAY, MIP or LMIP. Because all calculations are performed parallel on the GPU, real-time visualization can be achived. The second part focuses mainly on the analysis procedure, where a morhpological concept known as differential attribute profiles was used. Each profifile is calculated from two successive attibute filterings. The analysis result is then displayed using the CSL and ABV model, which use HSV and CIELab LCH color space