Multi-source intrinsic colourisation

A novel scheme combined with example-based colourisation and scribble-based colourisation is presented according to differences of illumination eliminated between the target image and multiple reference images. As the target image and reference images are in different illumination conditions, all the images are classified as illumination images and reflectance images to remove the affect of illumination. After modelling the regional colour distribution, appropriate colours are selected from given reference reflection images to colourise each target reflectance region. The recoloured target reflection image and the target illumination image are combined to obtain the final result. Experimental results demonstrate that our method generates more powerful performance than previous ones.

[1]  Liang Lin,et al.  Painterly animation using video semantics and feature correspondence , 2010, NPAR.

[2]  Song Wang,et al.  Image inpainting based on scene transform and color transfer , 2010, Pattern Recognit. Lett..

[3]  Frederick A.A. Kingdom,et al.  Chromatic properties of the colour-shading effect , 2005, Vision Research.

[4]  Erik Reinhard,et al.  Color Transfer between Images , 2001, IEEE Computer Graphics and Applications.

[5]  Nam Ik Cho,et al.  Colorization based on soft segmentation , 2011, J. Electronic Imaging.

[6]  Enhua Wu,et al.  Colorization Using the Rotation-Invariant Feature Space , 2011, IEEE Computer Graphics and Applications.

[7]  Klaus Mueller,et al.  Transferring color to greyscale images , 2002, ACM Trans. Graph..

[8]  Sung-Hsien Hsieh,et al.  Weighted Map for Reflectance and Shading Separation Using a Single Image , 2009, ACCV.

[9]  H. Barrow,et al.  RECOVERING INTRINSIC SCENE CHARACTERISTICS FROM IMAGES , 1978 .

[10]  Yair Weiss,et al.  Deriving intrinsic images from image sequences , 2001, Proceedings Eighth IEEE International Conference on Computer Vision. ICCV 2001.

[11]  Hong Li,et al.  Selective color transfer with multi-source images , 2009, Pattern Recognit. Lett..

[12]  Dani Lischinski,et al.  Colorization using optimization , 2004, SIGGRAPH 2004.

[13]  Stefano Messelodi,et al.  Linking the von Kries model to Wien's law for the estimation of an illuminant invariant image , 2011, Pattern Recognit. Lett..