Open Problems for Image Compression

1
Open Problems for Image Compression

• The server store a set of image data (2D)
• The client does remote surface extraction.
• Two Problems
  a) Efficient
  image transmission
  b) Efficient surface transmission.
• Wish
  a) Less
  network traffic
  b) Progressive mode.

2
Proposed Research for Remote Surface Access

• Proposal 1 Build hierarchical data structure at
  server side for input images Do error bounded
  surface extraction and encode the surface and
  transmit it. The client simply deciphers the
  received compressed surface and displays it.
• Proposal 2 The server compresses and encode all
  images via hierarchical data structure and then
  progressively transmit the images. The client
  decodes what it has received and does 3D surface
  reconstruction.

3
3D Geometric Compression

• Proposal 3 Geometric compression and encoding.
  a) Take bread-first search to build
  the vertex spanning tree b) Use productive VQ to
  compress and encode geometric quantities.
• Proposal 4 Connectivity compression and
  encoding. a) Mesh simplification by
  re-tiling
  b) Encode the simplified mesh more efficiently
  by using special encoding scheme (see blow).

4
3D Geometric Compression (cont.)

• Proposal 5 Design special compression and
  encoding algorithm for the special objects --- 3D
  surfaces reconstructed from planar coss sections.
• Proposal 6 Design special compression and
  encoding algoirthm for the special objects --- 3D
  surfaces reconstructed from a set of images
  (volume data).
• Proposal 7 Progressive transmission. I will
  take two approach data value refinement and
  spatial resolution refinement.