Research On Converting Color Image Into Gray Image

With the development of science and technology,the pictures become the basic elements of the digital media and deep into all aspects of our lives.Because of rich visual effect and more colors,the color image occupy dominant position in the information transmission.But the gray image have fewer data and contain the main information of color image,so the gray image also have an important application in practical,such as digital printing,modern medical research and treatment,digital image processing,pattern recognition and other scientific fields.So far there are many conversion methods of color image to gray image.At the same time,the existing methods for few images of the conversion loss a lot of information.From the point of the space,the conversion is a reduced dimension of the mapping process,which is a transformation of lots of dimensions to one dimension.Obviously multidimensional color image can contain more information and colors,and the conversion must lead to information loss.So the question,which is how to retain the characteristics,brightness and structure of color image in the conversion,is the main problem.A good conversion can make the gray image meet the demand on the aspect of visual and application.This is the core problem of conversion from color image to gray image.Based on the study of several popular gray conversion algorithms,we proposed a novel framework for optimal RGB to grayscale image conversion method.The main idea is that: firstly,image entropy has such a characteristic: the image entropy is bigger,then the image intensity distribution is more uniform,the gray level range is wider;the entropy is smaller then the image gray level distribution range is narrower.We propose an image entropy-based optimization framework to choose the optimal line direction so that all the pixel color vectors in an image have the most spread-out projections,thus increasing the grayscale image contrast;secondly,we make use of histogram specification on all the projection points to further increase the image contrast.Experimental results show that our method achieves a good visual effect in human perception,keeps the structure of the color image and local details.At the end of our essay we compare a few existing classical algorithms with our method on the two aspects of subjective andobjective quality assessment,our algorithm have better visual effect on subjective quality assessment;objective quality assessment include : the CAF model for no-reference image quality evaluation,the C2G_SSIM and E_score model for full-reference image quality evaluation.it proved that our method is much more practical and superior.