基于支持向量回归的光谱反射率重建方法

Spectral Reflectance Estimation by Support Vector Regression

提出了一种基于支持向量回归和小框架核的光谱反射率重建方法。光谱反射率重建是光学研究的一个重要问题,其目的是通过各种成像设备所获取的与设备相关的RGB三色值重建出物体本身固有的与设备和光照都无关的光谱反射率。回归方法已经在这一领域取得了广泛应用,如基于多项式模型的正则化最小二乘方法、基于核的正则化最小二乘方法等。提出了一种新的光谱反射率重建方法,这种方法采用了一种可以减弱样本不规则噪音影响的小框架核函数,并将其用于支持向量回归来重建光谱反射率函数。实验表明,新方法可以提高光谱反射率重建的精度和稳定性。

A spectral reflectance estimation method using support vector regression and framelet kernel was proposed.Spectral reflectance estimation is an important subject in optical research.The aim is to convert device-dependent RGB values to device-and illuminant-independent reflectance spectra.Regression methods are widely used to estimate spectral reflectance of surface colors given their camera responses,such as regularized least squares method with polynomial models,kernel based regularized least squares method,etc.In this paper,we introduced a novel estimating approach based on the support vector regression method.The proposed approach utilizes a framelet based kernel,which has the ability to approximate functions with multiscale structure and can reduce the influence of noise in data.Experimental results show that the technique can improve the recovery accuracy and stability.