System for natural language assessment of relative color quality
View Patent ↗Embodiments of the invention include a system for providing a natural language objective assessment of relative color quality between a reference and a source image. The system may include a color converter that receives a difference measurement between the reference image and source image and determines a color attribute change based on the difference measurement. The color attributes may include hue shift, saturation changes, and color variation, for instance. Additionally, a magnitude index facility determines a magnitude of the determined color attribute change. Further, a natural language selector maps the color attribute change and the magnitude of the change to natural language and generates a report of the color attribute change and the magnitude of the color attribute change. The output can then be communicated to a user in either text or audio form, or in both text and audio forms.
1. A hardware system for providing a natural language objective assessment of relative color quality between a series of reference images from video images and a series of source images from video images, comprising:
a processor including:
a moving image color appearance model structured to determine a difference measurement between selected ones of the reference images and a corresponding source image of the series of source images;
a color convertor structured to receive the difference measurement and to determine a color attribute change based on the difference measurement;
a magnitude index facility structured to determine a magnitude of the determined attribute change;
a natural language selector structured to generate a natural language report of the color attribute change and of the magnitude of the color attribute change; and
an output module structured to communicate the natural language report to a user of the system.
2. The system for providing a natural language objective assessment of relative color quality between a series of reference images and a series of source images of claim 1 , wherein the color converter is a hue shift determination facility, and in which the color attribute change is hue shift between the selected ones of the reference images and the corresponding source image.
3. The system for providing a natural language objective assessment of relative color quality between a series of reference images and a series of source images of claim 1 , wherein the color converter is a saturation determination facility, and in which the color attribute change is saturation difference between the selected ones of the reference images and the corresponding source image.
4. The system for providing a natural language objective assessment of relative color quality between a series of reference images and a series of source images of claim 1 , wherein the color converter is a color variation determination facility, and in which the color attribute change is a color variety variation between the selected ones of the reference images and the corresponding source image.
5. The system for providing a natural language objective assessment of relative color quality between a series of reference images and a series of source images of claim 1 , wherein the output module is a voice synthesis module structured to accept the natural language report and generate an audio report to the user of the system.
6. The system for providing a natural language objective assessment of relative color quality between a series of reference images and a series of source images of claim 1 , further comprising a region of interest selector structured to accept user input to delineate boundaries to be included in the assessment, wherein the boundaries are of portions of the series of references images and are of portions of the series of source images.
7. The system for providing a natural language objective assessment of relative color quality between a series of reference images and a series of source images of claim 6 , in which the region of interest selector is a spatial selector.
8. The system for providing a natural language objective assessment of relative color quality between a series of reference images and a series of source images of claim 6 , in which the region of interest selector is a temporal selector.
9. A method for providing a natural language assessment of relative color quality between a series of reference images from video images and a series of source images from video images, the method comprising:
determining, via a moving image color appearance model, a color comparison difference measurement between selected ones of the reference images and a corresponding source image of the series of source images, wherein the color comparison difference measurement is within a region of interest in space and/or time;
accepting, at a processor and from the moving image color appearance model, an input of the color comparison difference measurement;
converting, via the processor, the color comparison difference measurement to a two-dimensional color space difference;
determining, via the processor, a color attribute change based on the two-dimensional color space difference;
determining, via the processor, a magnitude of the color attribute change;
mapping, via the processor, the magnitude of the color attribute change to one of a collection of natural language words;
generating, via the processor, from the mapped magnitude, and from the two-dimensional color space difference, the natural language assessment;
outputting, via the processor, the natural language assessment.
10. The method providing a natural language assessment of relative color quality between a series of reference images and a series of source images of claim 9 , wherein the two-dimensional color space accords to CIECAM02.
11. The method providing a natural language assessment of relative color quality between a series of reference images and a series of source images of claim 9 , wherein determining a color attribute change comprises determining a hue shift between the selected ones of the reference images and the corresponding source image.
12. The method providing a natural language assessment of relative color quality between a series of reference images and a series of source images of claim 9 , wherein determining a color attribute change comprises determining a saturation measurement between the selected ones of the reference images and the corresponding source image.
13. The method providing a natural language assessment of relative color quality between a series of reference images and a series of source images of claim 9 , wherein determining a color attribute change comprises determining a color variation measurement between the selected ones of the reference images and the corresponding source image.
14. The method providing a natural language assessment of relative color quality between a series of reference images and a series of source images of claim 9 , further comprising selecting, for assessment, the region of interest.
15. The method providing a natural language assessment of relative color quality between a series of reference images and a series of source images of claim 9 , further comprising generating an audio report of the natural language assessment.