Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Method and System for Detecting Monochromatic Images

IP.com Disclosure Number: IPCOM000241908D
Publication Date: 2015-Jun-08
Document File: 3 page(s) / 291K

Publishing Venue

The IP.com Prior Art Database

Related People

Nikhil Rasiwasia: INVENTOR [+3]

Abstract

A method and system is disclosed for detecting monochromatic images wherein the monochromatic images are of one color or are shaded in one color. The method and system includes detecting three different low-level features of an image to detect monochromatic images.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 53% of the total text.

Method and System for Detecting Monochromatic Images

Abstract

A method and system is disclosed for detecting monochromatic images wherein the monochromatic images are of one color or are shaded in one color.  The method and system includes detecting three different low-level features of an image to detect monochromatic images.

Description

Disclosed is a method and system for detecting monochromatic images by using three different low-level features of an image.  The monochromatic images may include black-and-white, sepia tone, cyanotype (“blue print”) and other split-toned images.  The different types of monochromatic images are illustrated in the Figure.

Figure

As illustrated in the figure, ‘a’ represents a color image, ‘b’ represents a black-and white image of ‘a’, ‘c’ represents a sepia tone of the image, ‘d’ represents the cyanotype or blue print of the image and ‘e’ represents the split-tone of the image.

The above mentioned monochromaticity of an image is detected by identifying three different low-level features of the image.  It is assumed that color values for an image at pixel/position (x, y) is denoted as (R(x, y), G(x, y), B(x, y)) and the total number of pixels in the image is #pixels.

A first low-level feature of the image that is detected is a chroma score of the image.  The chroma score is used to measure saturation of the image.  The saturation of the image is defined based on the chroma score obtained as a result of colors in the image.  For example, an image with intense colors results in high chroma score and another image with washed out colors results in low chroma score.  The chroma score ‘C’ at pixel (x, y) is computed as follows:

Let, M(x, y) = max(R(x, y), G(x, y), B(x, y)); and

m(x ,y) = min(R(x, y), G(x, y), B(x, y))

Then, chroma C(x, y) = M(x, y) - m(x, y).

An average chroma score for the image is computed as:

C_Image = (\sum_x \sum_y C(x, y)) / (#pixels)

A second low-level feature for detecting monochromaticity of the image is Hue Entropy.  Hue is one of the main properties of the color that is used to encode a shade of the color.  The monochromaticity of the image is identified by the presence of single hue in the image.  The standard hue computation algorithm is utilized to recover hue of each pixel.  Subsequently, variation in hue property across the pixels is measured by quantizing the hues in different bins and computing entropy of resulting histogram.  For a given image I(x, y) and a chroma score at every pixel C(x, y), the hue is computed as

h(x, y) = [-1 if C(x, y) = 0]

   [(G(x, y) - B(x, y))/C(x, y) mod 6 if M(x, y) = R]

              [(B(x, y) - R(x, y))/C(x, y) + 2 if M(x, y) = G]

              [(R(x, y) - G(x, y))/C(x, y) + 4 if M(x, y) = B]

A normalized range of Hues is between [0, 1] that is computed as: h(x, y) = h(x, y) / 6.  Subseq...