Patent EP0671682A2 successfully refreshed from Google Patents.
US7512286B2
Expired / Lapsed
US
⏱️ 0.0 years remaining
Assessing image quality
Assignee: Hewlett Packard Development Co LP
•
Filing: 2003-10-27
•
Granted: 2009-03-31
•
Expires: 2023-10-27
Abstract
Abstract Systems and methods of assessing image quality are described. In one scheme for assessing image quality, a target object region is detected in an input image. An image quality feature vector representing the target object region in an image quality feature space is generated. The image quality feature vector is mapped to a measure of image quality. In one scheme for generating an image quality assessment engine, target object regions are detected in multiple input images. Image quality feature vectors representing the target object regions in an image quality feature space are generated. The image quality feature vectors are correlated with respective measures of image quality assigned to the input images. A mapping between image quality feature vectors and assigned measures of image quality is computed.
Inventors
- 👤 Huitao Luo
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-09 11:30
CPC Classifications
H; H04; H04N; H04N17/00; H04N17/004; G; G06; G06T; G06T7/00; G06T7/0002; G06V; G06V10/00; G06V10/98; G06V10/993; G06V40/00; G06V40/10; G06V40/16; G06V40/168; H04N19/00; H04N19/10; H04N19/134; H04N19/136; H04N19/14; H04N19/169; H04N19/17; H04N19/85; G06T2207/00; G06T2207/30; G06T2207/30168
Claims Analysis
Total Claims: 42
Claim 1
Independent Claim
162 words
A method of assessing image quality, comprising operating a computer processor to perform operations comprising: detecting a target object region in an input image, wherein the detecting is performed on image data derived solely from the input image without regard to image data derived from any image other than the input image and the detecting comprises determining that the target object region comprises a target object; generating an image quality feature vector representing the target object region in an image quality feature space, wherein the generating comprises generating the image quality feature vector from image data derived solely from the input image without regard to image data derived from any image other than the input image; mapping the image quality feature vector to a measure of image quality based on a classification model that correlates image quality feature vectors with respective image quality values; and outputting a quantitative assessment of quality of the input image based on the measure of image quality.
Claim 2
Independent Claim
162 words
A method of assessing image quality, comprising operating a computer processor to perform operations comprising: detecting a target object region in an input image, wherein the detecting is performed on image data derived solely from the input image without regard to image data derived from any image other than the input image and the detecting comprises determining that the target object region comprises a target object; generating an image quality feature vector representing the target object region in an image quality feature space, wherein the generating comprises generating the image quality feature vector from image data derived solely from the input image without regard to image data derived from any image other than the input image; mapping the image quality feature vector to a measure of image quality based on a classification model that correlates image quality feature vectors with respective image quality values; and outputting a quantitative assessment of quality of the input image based on the measure of image quality.
Claim 3
Dependent Claim
18 words
The method of claim 1 , wherein the detecting comprises detecting a human face in the input image.
Claim 4
Dependent Claim
26 words
The method of claim 1 , wherein the detecting comprises detecting an object relevant to a person's subjective assessment of image quality in the input image.
Claim 5
Dependent Claim
24 words
The method of claim 1 , wherein the detecting comprises detecting the target object region based on a sub-sampled version of the input image.
Claim 6
Dependent Claim
32 words
The method of claim 4 , wherein the generating comprises generating the image quality feature vector based on a version of the target object region at a resolution of the input image.
Claim 7
Dependent Claim
54 words
The method of claim 1 , wherein the detecting comprises detecting the target object region based on a first set of features of the input image, and the generating comprises generating the image quality feature vector based on a second set of features of the input image different from the first set of features.
Claim 8
Dependent Claim
21 words
The method of claim 6 , wherein the first set of features is substantially decoupled from the second set of features.
Claim 9
Dependent Claim
33 words
The method of claim 1 , wherein the image quality feature space is spanned by multiple features including at least one brightness feature describing a respective brightness characteristic of the target object region.
Claim 10
Dependent Claim
34 words
The method of claim 1 , wherein the image quality feature space is spanned by multiple features including at least one spectral feature describing a respective spatial frequency characteristic of the target object region.
Claim 11
Dependent Claim
24 words
The method of claim 9 , wherein generating the image quality feature vector comprises decomposing the target object region into multiple wavelet transform sub-bands.
Claim 12
Dependent Claim
18 words
The method of claim 10 , wherein each spectral feature describes energy in a respective wavelet transform sub-band.
Claim 13
Dependent Claim
33 words
The method of claim 1 , wherein the image quality feature space is spanned by multiple features including at least one noise feature describing a respective noise characteristic of the target object region.
Claim 14
Dependent Claim
25 words
The method of claim 12 , wherein the generating comprises computing the noise feature based on a measure of noise in the target object region.
Claim 15
Dependent Claim
36 words
The method of claim 12 , wherein the generating comprises computing the noise feature based on a measure of spatial homogeneity of spectral features each describing a respective spatial frequency characteristic of the target image region.
Claim 16
Dependent Claim
31 words
The method of claim 1 , wherein the mapping comprises mapping the image quality feature vector to a measure of image quality in accordance with a machine learning based classification process.
Claim 17
Dependent Claim
33 words
The method of claim 15 , wherein the mapping comprises mapping the image quality feature vector to a measure of image quality in accordance with a radial basis function based machine learning process.
Claim 18
Dependent Claim
33 words
The method of claim 15 , wherein the mapping comprises mapping the image quality feature vector to a measure of image quality in accordance with a mixture of Gaussian based machine learning process.
Claim 19
Independent Claim
157 words
A system for assessing image quality, comprising computing hardware operable to perform operations comprising: detecting a target object region in an input image based on image data derived solely from the input image without regard to image data derived from any image other than the input image, wherein in the detecting the computing hardware is operable to perform operations comprising determining that the target object region comprises a target object; generating an image quality feature vector representing the target object region in an image quality feature space based on image data derived solely from the input image without regard to image data derived from any image other than the input image; mapping the image quality feature vector to a measure of image quality based on a classification model that correlates image quality feature vectors with respective image quality values; and outputting a quantitative assessment of quality of the input image based on the measure of image quality.
Claim 20
Independent Claim
157 words
A system for assessing image quality, comprising computing hardware operable to perform operations comprising: detecting a target object region in an input image based on image data derived solely from the input image without regard to image data derived from any image other than the input image, wherein in the detecting the computing hardware is operable to perform operations comprising determining that the target object region comprises a target object; generating an image quality feature vector representing the target object region in an image quality feature space based on image data derived solely from the input image without regard to image data derived from any image other than the input image; mapping the image quality feature vector to a measure of image quality based on a classification model that correlates image quality feature vectors with respective image quality values; and outputting a quantitative assessment of quality of the input image based on the measure of image quality.
Claim 21
Dependent Claim
29 words
The system of claim 18 , wherein in detecting the target object region the target object region detection module performs a human face detection process on the input image.
Claim 22
Dependent Claim
25 words
The system of claim 18 , wherein the feature extraction module detects the target object region based on a sub-sampled version of the input image.
Claim 23
Dependent Claim
33 words
The system of claim 18 , wherein the image quality feature space is spanned by multiple features including at least one brightness feature describing a respective brightness characteristic of the target object region.
Claim 24
Dependent Claim
34 words
The system of claim 18 , wherein the image quality feature space is spanned by multiple features including at least one spectral feature describing a respective spatial frequency characteristic of the target object region.
Claim 25
Dependent Claim
31 words
The system of claim 22 , wherein the feature extraction module is operable to generate the image quality feature vector by decomposing the target object region into multiple wavelet transform sub-bands.
Claim 26
Dependent Claim
18 words
The system of claim 23 , wherein each spectral feature describes energy in a respective wavelet transform sub-band.
Claim 27
Dependent Claim
33 words
The system of claim 18 , wherein the image quality feature space is spanned by multiple features including at least one noise feature describing a respective noise characteristic of the target object region.
Claim 28
Dependent Claim
26 words
The system of claim 25 , wherein the feature extraction module computes the noise feature based on a measure of noise in the target object region.
Claim 29
Dependent Claim
37 words
The system of claim 25 , wherein the feature extraction module computes the noise feature based on a measure of spatial homogeneity of spectral features each describing a respective spatial frequency characteristic of the target image region.
Claim 30
Dependent Claim
33 words
The system of claim 18 , wherein the image quality assessment module maps the image quality feature vector to a measure of image quality in accordance with a machine learning based classification process.
Claim 31
Dependent Claim
35 words
The system of claim 28 , wherein the image quality assessment module maps the image quality feature vector to a measure of image quality in accordance with a radial basis function based machine learning process.
Claim 32
Dependent Claim
35 words
The system of claim 28 , wherein the image quality assessment module maps the image quality feature vector to a measure of image quality in accordance with a mixture of Gaussian based machine learning process.
Claim 33
Dependent Claim
159 words
A computer-readable medium storing computer-readable instructions for causing a computer to perform operations comprising: detecting a target object region in an input image based on image data derived solely from the input image without regard to image data derived from any image other than the input image, wherein in the detecting the computer-readable instructions cause the computer to perform operations comprising determining that the target object region comprises a target object; generating an image quality feature vector representing the target object region in an image quality feature space based on image data derived solely from the input image without regard to image data derived from any image other than the input image; and mapping the image quality feature vector to a measure of image quality based on a classification model that correlates image quality feature vectors with respective image quality values; and outputting a quantitative assessment of quality of the input image based on the measure of image quality.
Claim 34
Dependent Claim
159 words
A computer-readable medium storing computer-readable instructions for causing a computer to perform operations comprising: detecting a target object region in an input image based on image data derived solely from the input image without regard to image data derived from any image other than the input image, wherein in the detecting the computer-readable instructions cause the computer to perform operations comprising determining that the target object region comprises a target object; generating an image quality feature vector representing the target object region in an image quality feature space based on image data derived solely from the input image without regard to image data derived from any image other than the input image; and mapping the image quality feature vector to a measure of image quality based on a classification model that correlates image quality feature vectors with respective image quality values; and outputting a quantitative assessment of quality of the input image based on the measure of image quality.
Claim 35
Independent Claim
177 words
A method of generating an image quality assessment engine, comprising operating a computer processor to perform operations comprising: for each of multiple input images detecting a respective target object region in a respective one of the input images based on image data derived solely from the respective input image without regard to image data derived from any image other than the respective input image, wherein the detecting comprises determining that the target object region comprises a target object, and generating a respective image quality feature vector representing the respective target object region in an image quality feature space, wherein the generating comprises generating the respective image quality feature vector from image data derived solely from the respective input image without regard to image data derived from any image other than the respective input image; correlating the image quality feature vectors with respective measures of image quality assigned to the input images; determining a mapping between image quality feature vectors and the assigned measures of image quality; and storing the mapping as computer-readable instructions on a computer-readable medium.
Claim 36
Independent Claim
177 words
A method of generating an image quality assessment engine, comprising operating a computer processor to perform operations comprising: for each of multiple input images detecting a respective target object region in a respective one of the input images based on image data derived solely from the respective input image without regard to image data derived from any image other than the respective input image, wherein the detecting comprises determining that the target object region comprises a target object, and generating a respective image quality feature vector representing the respective target object region in an image quality feature space, wherein the generating comprises generating the respective image quality feature vector from image data derived solely from the respective input image without regard to image data derived from any image other than the respective input image; correlating the image quality feature vectors with respective measures of image quality assigned to the input images; determining a mapping between image quality feature vectors and the assigned measures of image quality; and storing the mapping as computer-readable instructions on a computer-readable medium.
Claim 37
Dependent Claim
27 words
The method of claim 32 , wherein for each respective one of the input images the detecting comprises detecting a human face in the respective input image.
Claim 38
Dependent Claim
33 words
The method of claim 32 , wherein the image quality feature space is spanned by multiple features including at least one brightness feature describing a respective brightness characteristic of the target object region.
Claim 39
Dependent Claim
34 words
The method of claim 32 , wherein the image quality feature space is spanned by multiple features including at least one spectral feature describing a respective spatial frequency characteristic of the target object region.
Claim 40
Dependent Claim
19 words
The system of claim 18 , wherein the computing hardware comprises a computer-readable storage medium and a computer processor.
Claim 41
Dependent Claim
32 words
The method of claim 1 , wherein the determining comprises classifying each of multiple regions of the input image into one of a target object region class and a false alarm class.
Claim 42
Dependent Claim
30 words
The method of claim 1 , wherein the classification model captures a functional relationship between the image quality feature vectors and quantitative assessments of image quality assigned to training images.
References Cited (Backward Citations)
US5400667A
Cited
US6243492B1
Cited
US6556196B1
Cited
US6285797B1
Cited
WO2001033497A1
Cited
US6798919B2
Cited
US20020171852A1
Cited
US7068841B2
Cited
US20030128876A1
Cited
US6992697B2
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2012-03-30 | 2012-09-29 | 2013-03-31 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2016-03-30 | 2016-09-29 | 2017-03-31 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2020-03-30 | 2020-09-29 | 2021-03-31 | $7700.00 | Window Closed / Lapsed |