Project Update - Paschal
Progress and Timeline:
COMPLETE - Week of March 9: Pre-production planning, shoot test video
COMPLETE - March 17: Individual parts researched and partially implemented
IN PROGRESS - Week of March 24: Finish individual parts, shoot footage
March 31-April 6: Integration
April 10: Finished effect
Pashcal is on schedule to finish independent code for collaboration and integration this week. We have completed a second round of test video which delayed shooting feature footage. Additional tests on our "canvas" checker board are necesary before live footage can be taken. We are also continuing to tweak our storyline while we have time.
GTTiP Update:
We're extremely close to finishing up GTTiP. Unfamiliarity with openGL has delayed completion from what was hoped for, but it is still within the range of our deadline.
Foreground elements will be done by hand, making the gui to create and place the elements would require too much time.
Also, we're looking into outputting texture coordinates instead of pre-rendered textures. There are several advantages to this, not the least of which is the savings in time.
The program so far only loads square bitmaps (bmp) up to 1024 pixels. Non-square scenes can be edited to extend the canvas, so this isn't a large problem. A more complicated problem, and one that touring into the picture doesn't run into, is the framing of the scene. With tour into the picture, portions of the walls which have no visual data from the scene will never be seen. Our project is designed to view the painting from the outside, mesh layouts rarely have corners matching those of the box.
Accomplished:
- create a mesh to design a simple 3d object on top of a background
Still to go:
Screenshots:
Code snapshot:
Written and compiled on OS 10.2.4. Binary and sample images located in build directory. Source code and binaries provided.
GTTiP03252003.zip
Motion Tracking Update:
Motion Tracking is going well. The software can recognize the squares on the grid. However, tests with the tracking reveal that we want to have three colors on our checker board instead of two. The third color will replace the squares on the four corners of the board. We will composite all three colors out of the video. First test frames of a 3 color board are successful. The complicated part is converting 2D tracking to the 3D world. The output from the motion tracking will be a sequence of rotation matrices.
The four corner pts tracked by software.
MEL Script Generator Update:
The MEL script can create a cube and move it according to a rotation matrix for each frame. The output from the motion capture will be a sequence of rotation matrices. Maya may have to slice the image texture to be applied to the side of the cube depending upon OpenGL complications with our TIP code. So far no foreground elements have been added to the cube. We are on schedule to have all of the MEL Script Generator functionality solid by the end of the week.
Links to this Page
- DVFX2003: Groups last edited on 1 May 2003 at 4:53 pm by davinci.cc.gatech.edu
- Paschal last edited on 25 March 2003 at 1:52 pm by adsl-158-58-162.asm.bellsouth.net