A Polynomial Regression Algorithm for RGBXYZ Conversion

摘要：

It is known that the conversion from monitordependent RGB value to device-independent CIE XYZ value is a necessary technology in ICC color management. The conversion is usually based on three non-interdependent per-channel tone reproduction curves to convert between non-linear and linear RGB values and a 3×3 matrix to convert linear RGB values into relative XYZ values. To improve the transformation precision and to take the interdependent relation between channels into consideration, some researchers have developed polynomials to substitute for tone reproduction curves or 3×3 matrix. In this paper, a new third-order polynomial regression algorithm is proposed to convert non-linear RGB values to XYZ values directly without transformation to linear RGB values. To test the validity and the precision of the proposed polynomials, RGB color samples were chosen randomly. The errors between the samples’ XYZ values computed with the proposed polynomials and the values measured were computed. The result showed that average error was small, which proved the feasibility of the proposed method.

关键词： RGB XYZ Polynomial Regression Error Analysis

作者: Xia Zhang Guangwei Liu

作者单位: Printing and Packaging School, Wuhan University 129, Luoyu Rd., Wuhan 430079, Hubei, P.R.China Printing and Packaging School, Wuhan University 129, Luoyu Rd., Wuhan 430079, Hubei, P.R.China

会议类型: 国际会议

会议名称: The 31st International Congress on Imaging Science(第31届国际影像科学大会 ICIS2010)

会议地点: 北京

会议语种:英文

页码: 189-192

在线出版日期: 2010-05-12（万方平台首次上网日期，不代表论文的发表时间）