Optimization-based Image Decolorization Why Grayscale Image? Student Name: XU Zhinan Advisor: Professor Chiew-Lan Tai a)Widely used in pattern recognition.

Optimization-based Image Decolorization

Why Grayscale Image?

Student Name: XU Zhinan Advisor: Professor Chiew-Lan Tai

a) Widely used in pattern recognition & image processingb) Small scale of data & simple to code

c) Monochrome printing

Previous Work

[Neumann et al. 2007]

[Gooch et al. 2005] [Grundland et al. 2007]

CIE Y

Local contrast [Kim et al. 2009]

• Disadvantage:• Global contrast is ignored.

• Advantages:• Global mapping• Local contrast preserving

• Conversion: (Lch to grayscale)

• Objective Function:

Global contrast—Lch & RGB

Input Lch RGB Kim et al. CIE Y

• Global contrast: statistics in CIE Lch

Global contrast—Gaussian Pairing

Input Kim et al. Gaussian

• Displacement vector:

Optimization-based Image Decolorization