Milestones
Milestones
- Stage 0 - Data preparation
- Convert video streams into image sequences
- Label sequences according to sequence runs
- Each distinct viewpoint gets a unique camera name. all adjacent frames from the viewpoint follow in sequence
- For example: a0000.jpg a0001.jpg a0002.jpg b0000.jpg
- Prep the images by pushing them through an outlining / edge-detection filter
- Stage 1 - Data Segmentation and Flattening
- Segment the images into self-similar patches
- Edge the images to approximate edges of patches
- Crawler?
- Use outlined and original image as a guide
- Describe patch by a generalized histogram - ex: find mean and std.dev for each channel
- Match patches within image sequence - assume that the patches do not move "greatly" from frame to frame
- Edge the images to approximate edges of patches
- Segment the images into self-similar patches
- Stage 2 - Stitching patchwork and testing coherency
- Stitch the patches into adjacent patchworks
- Use "feedback loops" to test temporal / spatial coherence
- If texture A and B are adjacent in frame f but not in frame g, either they're not adjacent on the object's surface or there is something occluding the joint between A and B.
- "Flatten" surface texture - Find highest quality, average level set of pixels that best describe the texture
- Stitch the patches into adjacent patchworks
- Stage 3 - 3D Reconstruction
- SIFT and SFM ? to create sparse 3D reconstruction
- Describe how all permutations change the texture's representation (shade, highlight, etc.)
page revision: 7, last edited: 18 Dec 2008 20:09