One of the new standard extensions is called informally JP3D and deals with the coding of 3D image datasets coming from medical data (MRI's, CAT scan's, etc), scientific measurements or 3D numerical simulations. Formally, this is called JPEG 2000 Part 10while what is commonly known as JPEG 2000 is really JPEG 2000 Part 1 (the baseline system).
For the most part, JP3D is just a straightforward extension of the JPEG 2000 Part 1 coding principles from two to three dimensions. The major innovations introduced by JP3D are the coding of floating point data (as opposed to the typical coding of 24 bit RGB data) and the coding of variable resolution grids featuring adaptive refinement of the grid cells. Both these innovations are motivated by the wealth of existing scientific measurement and simulation 3D datasets exhibiting these characteristics.
This paper proposes a technology that extends the JPEG 2000 coding pipeline to accommodate adaptive refinement grids. It features a first pass where the structure of the adaptive grid is entropy coded in a lossless manner and then a second pass where the cascade of wavelet transforms is adapted to work over irregular grids. The paper was presented at the Visual Communications and Image Processing 2003 Conference, held in Lugano, Switzerland. The final version is in the SPIE Digital Library and the draft version is below:
For the most part, JP3D is just a straightforward extension of the JPEG 2000 Part 1 coding principles from two to three dimensions. The major innovations introduced by JP3D are the coding of floating point data (as opposed to the typical coding of 24 bit RGB data) and the coding of variable resolution grids featuring adaptive refinement of the grid cells. Both these innovations are motivated by the wealth of existing scientific measurement and simulation 3D datasets exhibiting these characteristics.
This paper proposes a technology that extends the JPEG 2000 coding pipeline to accommodate adaptive refinement grids. It features a first pass where the structure of the adaptive grid is entropy coded in a lossless manner and then a second pass where the cascade of wavelet transforms is adapted to work over irregular grids. The paper was presented at the Visual Communications and Image Processing 2003 Conference, held in Lugano, Switzerland. The final version is in the SPIE Digital Library and the draft version is below:
No comments:
Post a Comment