Image processing system configured to generate a colorized image
AssigneeNEC Corporation
InventorTakahiro Toizumi
Life of this patent
Estimating the term…
Abstract
An image processing system according to an example aspect of the present disclosure includes: an image acquisition unit configured to acquire a monochrome image; a pixel value correction unit configured to correct a pixel value of the monochrome image based on information related to a pixel value; and a colorization generation unit configured to generate a colorized image corresponding to the monochrome image from the corrected monochrome image by using a colorization prediction model trained by machine learning. With the present disclosure, it is possible to improve the reproduction accuracy of a color in colorization of a monochrome image.
Claims7
An image processing system comprising:
- at least one memory storing instructions;
- and
- at least one processor configured to execute the instructions to:
- acquire a monochrome image;
- decompose a plurality of pixels in the monochrome image into superpixels based on pixel values and pixel positions of the plurality of pixels;
- cluster the superpixels based on a representative pixel value and a representative pixel position of the superpixels to generate a cluster;
- determine an amount of correction of a pixel value of a pixel of the monochrome image based on a representative value of the pixel values of the cluster including one or more pixels associated with a color hint indicating a color of a target part in the monochrome image and a hint pixel value corresponding to the color hint;
- correct the pixel value of the pixel of the monochrome image based on information related to the pixel value to generate a corrected monochrome image, the information related to the pixel value comprising the color hint;
- and
- generate a colorized image corresponding to the monochrome image from the corrected monochrome image by using a colorization prediction model trained by machine learning.
An image processing method comprising:
- acquiring a monochrome image;
- decomposing a plurality of pixels in the monochrome image into superpixels based on pixel values and pixel positions of the plurality of pixels;
- clustering the superpixels based on a representative pixel value and a representative pixel position of the superpixels to generate a cluster;
- determining an amount of correction of a pixel value of a pixel of the monochrome image based on a representative value of the pixel values of the cluster including one or more pixels associated with a color hint indicating a color of a target part in the monochrome image and a hint pixel value corresponding to the color hint;
- correcting the pixel value of the pixel of the monochrome image based on information related to the pixel value to generate a corrected monochrome image, the information related to the pixel value comprising the color hint;
- and
- generating a colorized image corresponding to the monochrome image from the corrected monochrome image by using a colorization prediction model trained by machine learning.
A non-transitory computer readable medium storing an image processing program for causing a computer to:
- acquire a monochrome image;
- decompose a plurality of pixels in the monochrome image into superpixels based on pixel values and pixel positions of the plurality of pixels;
- cluster the superpixels based on a representative pixel value and a representative pixel position of the superpixels to generate a cluster;
- determine an amount of correction of a pixel value of a pixel of the monochrome image based on a representative value of the pixel values of the cluster including one or more pixels associated with a color hint indicating a color of a target part in the monochrome image and a hint pixel value corresponding to the color hint;
- correct the pixel value of the pixel of the monochrome image based on information related to the pixel value to generate a corrected monochrome image, the information related to the pixel value comprising the color hint;
- and
- generate a colorized image corresponding to the monochrome image from the corrected monochrome image by using a colorization prediction model trained by machine learning.
Dependencies are inferred from the claim text.