Development of A Novel Image Coding Method Using Color Correlations
| Project/Area Number |
09650398
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
情報通信工学
|
|---|
| Research Institution | Chiba University |
|---|
Principal Investigator |
KOTERA Hiroaki faculty of Engineering, Chiba University, Prefessor, 工学部, 教授 (70282449)
|
|---|
| Project Period (FY) |
1997 – 1999
|
| Project Status |
Completed (Fiscal Year 1999)
|
| Budget Amount *help |
¥3,100,000 (Direct Cost: ¥3,100,000)
Fiscal Year 1999: ¥900,000 (Direct Cost: ¥900,000)
Fiscal Year 1998: ¥800,000 (Direct Cost: ¥800,000)
Fiscal Year 1997: ¥1,400,000 (Direct Cost: ¥1,400,000)
|
|---|
| Keywords | Color image / Image compression / Color correlation / Luminance-chrominance space / Segmentation / PCA / Block coding / Projection / カラー画像 / 画像符号化 / 画像情報圧縮 / 輝度・色差分離 / 関数近似 / 色領域分割 / カラークラスタリング |
|---|
| Research Abstract |
This research has been addressed to a novel color image coding method during 1997 to 1999 and the following results have been obtained.
[1] Color image coding using color correlations between luminanec and chrominance signals.
Luminance-chrominance coding is popularly used in the present standard image compression technology, but chrominance is coded independent of luminance. Here, the strong correlatins between LィイD1*ィエD1 and (aィイD1*ィエD1,bィイD1*ィエD1) have been still observed after transform from RGB to LAB and chrominance (aィイD1*ィエD1,bィイD1*ィエD1) could be compressed by a polynomial expansion of luminance signal LィイD1*ィエD1 in sub-divided blocks.
[2] Co-oeration with conventional transform coding method.
The above luminance-chrominance correlation coding method has been extended to co-operate with conventional JPEG of Wavelet coding and the coding efficiency has been improved.
[3] Proposal of a novel color image coding method by PCA in image segmentation.
Color image has been segmented into different object areas with clustered color distributions. After the segmentation, each object area is characterized by Principal Component Analysis (PCA). The chrominance aィイD1*ィエD1 and bィイD1*ィエD1 values in CIELAB space are estimated by the projection of LィイD1*ィエD1 on to chromatic plane using the eigen vector and the mean vector. The segmented colored areas are indexed by the class number, which has been greatly compressed to 1/20 by conventional loss-less coding. The coded class number is transmitted with LィイD1*ィエD1 image and reduced PCA parameters by one set for each object. The (aィイD1*ィエD1,bィイD1*ィエD1) value of each pixel are approximately restored from LィイD1*ィエD1. LィイD1*ィエD1 image can be compressed by the conventional coding method such as JPEG or Wavelet. Finally, the full color image has been reproduced by combining the luminance LィイD1*ィエD1 with (aィイD1*ィエD1,bィイD1*ィエD1) with the color difference of ΔEィイD1*ィエD1ab=6〜7.
|
Report
(4 results)
Research Products
(31 results)
