Abstract
Color transfer in image processing usually suffers from misleading color mapping and loss of details. This paper presents a novel directive local color transfer method based on dynamic look-up table (D-DLT) to solve these problems in two steps. First, a directive mapping between the source and the reference image is established based on the salient detection and the color clusters to obtain directive color transfer intention. Then, dynamic look-up tables are created according to the color clusters to preserve the details, which can suppress pseudo contours and avoid detail loss. Subjective and objective assessments are presented to verify the feasibility and the availability of the proposed approach. Experimental results demonstrate that our proposed method has better performance on natural color images than classical color transfer algorithms. Furthermore, the reference image can be extended to color blocks instead of images.
Highlights
Color transfer aims to assign colors from a reference image to a source image
Conventional local methods include Gaussian Mixture Models (GMMs) [3,4], dominant color mapping [5], probabilistic moving least square algorithm [6], etc., while machine learning methods have a rapid development in the last decade such as colorization with SVM algorithm [7] and color style changed with neural network [8]
Dynamic color table from the reference image is created and color transfer based on color mapping is executed
Summary
Color transfer aims to assign colors from a reference image to a source image. Color transfer has various applications, such as image enhancement, image colorization and special effects for movies. Color mapping directs the colors in the reference image that should be transferred to the source image. We propose several improvements on conditional local color transfer methods. For manual interaction as proposed in [10], users can manually operate the color mapping relationship between the reference image and the source image. We propose a novel method for abstracting salient regions based on the improved simple non-iterative clustering (I-SNIC) method. The second step is directive color mapping which constructs the mapping relationships of the color clustering between the source image and the reference image. We study several open image datasets and propose a novel color mapping method based on the human vision system, which is more accurate and reasonable than the existed methods. A novel method for abstracting salient regions is proposed based on the improved simple non-iterative clustering (I-SNIC) method.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
