Patent JP2002132225A successfully refreshed from Google Patents.
Patent EP0495480A2 successfully refreshed from Google Patents.
Patent US7171023B2 successfully refreshed from Google Patents.
Patent US7171042B2 successfully refreshed from Google Patents.
US7171042B2
Expired / Lapsed
US
⏱️ 0.0 years remaining
System and method for classification of images and videos
Assignee: Intel Corp
•
Filing: 2000-12-04
•
Granted: 2007-01-30
•
Expires: 2020-12-04
Abstract
Abstract Systems and methods are provided through which a graphic image is classified in terms of being natural versus computer generated, or being a scientific slide presentation versus a comic image. The image is classified by extracting appropriate feature(s) from the image, and using the feature(s) to determine, within a predetermined degree of accuracy, the graphic classification of the image. The classification determination uses a trained model. The trained model is created by using machine learning algorithms such as Neural Networks, Support Vector Machines, and Learning Vector Quantizations. Subsequently, the trained model is used to classify a group of images of unknown classification automatically.
Inventors
- 👤 Alexander J. Hartmann
- 👤 Rainer W. Lienhart
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-09 13:53
CPC Classifications
G; G06; G06V; G06V30/00; G06V30/40; G06V30/41; G06V30/413; G06F; G06F16/00; G06F16/50; G06F16/58; G06F16/583; G06F16/5838
Claims Analysis
Total Claims: 29
Claim 1
Independent Claim
144 words
A method to train image classification, comprising: measuring noise and edge sharpness in a first image; generating a feature vector from the first image by; generating a noise-reduced second image from the first image; calculating a difference between the first image and the second image; generating a noise pixel histogram of the difference to use as a noise feature vector; generating a blurred third image from the first image; calculating another difference between the first image and the third image; generating a sharpness pixel histogram of the another difference to use as a sharpness feature vector; and combining the noise feature vector and the sharpness feature vector as the feature vector; and training a classification model from the feature vector to classify a fourth image as a natural image versus an artificial image from the noise and the edge sharpness in the fourth image.
Claim 2
Independent Claim
144 words
A method to train image classification, comprising: measuring noise and edge sharpness in a first image; generating a feature vector from the first image by; generating a noise-reduced second image from the first image; calculating a difference between the first image and the second image; generating a noise pixel histogram of the difference to use as a noise feature vector; generating a blurred third image from the first image; calculating another difference between the first image and the third image; generating a sharpness pixel histogram of the another difference to use as a sharpness feature vector; and combining the noise feature vector and the sharpness feature vector as the feature vector; and training a classification model from the feature vector to classify a fourth image as a natural image versus an artificial image from the noise and the edge sharpness in the fourth image.
Claim 3
Dependent Claim
22 words
The method of claim 1 , wherein generating the noise-reduced second image further comprises applying a median filter to the first image.
Claim 4
Dependent Claim
22 words
The method of claim 1 , wherein generating the noise-reduced second image further comprises applying a Gaussian filter to the first image.
Claim 5
Dependent Claim
22 words
The method of claim 1 , wherein generating the noise-reduced second image further comprises applying a Wiener filter to the first image.
Claim 6
Dependent Claim
18 words
The method of claim 1 , wherein the first image further comprises a frame in a video stream.
Claim 7
Dependent Claim
26 words
The method of claim 1 , wherein generating a blurred third image from the first image further comprises applying a Gaussian filter to the first image.
Claim 8
Dependent Claim
18 words
The method of claim 1 , wherein the first image further comprises all frames in a video stream.
Claim 9
Independent Claim
135 words
A method to classify a first image as natural image versus artificial image, further comprising: receiving a feature vector of the first image, wherein the feature vector comprises a combination of a noise vector and a sharpness vector of the first image; classifying the first image as natural image versus artificial image from the feature vector; and generating the classification of the image; wherein the classifying in is based on training from the combination of a noise feature vector and a sharpness feature vector, the noise feature vector including a pixel histogram of a difference between a second image and a noise-reduced third image generated from the second image, and the sharpness feature vector including a pixel histogram of a difference between the second image and a blurred fourth image generated from the second image.
Claim 10
Independent Claim
135 words
A method to classify a first image as natural image versus artificial image, further comprising: receiving a feature vector of the first image, wherein the feature vector comprises a combination of a noise vector and a sharpness vector of the first image; classifying the first image as natural image versus artificial image from the feature vector; and generating the classification of the image; wherein the classifying in is based on training from the combination of a noise feature vector and a sharpness feature vector, the noise feature vector including a pixel histogram of a difference between a second image and a noise-reduced third image generated from the second image, and the sharpness feature vector including a pixel histogram of a difference between the second image and a blurred fourth image generated from the second image.
Claim 11
Dependent Claim
46 words
The method of claim 8 , wherein the classification is performed on a set of video frames of a video sequence, and the most likely classification result emerging out of the classification results of the individual frames is taken as the class of the video sequence.
Claim 12
Dependent Claim
115 words
An image classification system comprising: a feature extraction component to extract a feature that distinguishes a frame in a video stream between a natural image versus an artificial image; and a training system operably coupled to the feature extraction component to train a classification model to distinguish the feature based on a combination of a noise feature vector and a sharpness feature vector; wherein the noise feature vector includes a pixel histogram of a difference between a first image and a noise-reduced second image generated from the second image, and the sharpness feature vector includes a pixel histogram of a difference between the first image and a blurred third image generated from the first image.
Claim 13
Dependent Claim
115 words
An image classification system comprising: a feature extraction component to extract a feature that distinguishes a frame in a video stream between a natural image versus an artificial image; and a training system operably coupled to the feature extraction component to train a classification model to distinguish the feature based on a combination of a noise feature vector and a sharpness feature vector; wherein the noise feature vector includes a pixel histogram of a difference between a first image and a noise-reduced second image generated from the second image, and the sharpness feature vector includes a pixel histogram of a difference between the first image and a blurred third image generated from the first image.
Claim 14
Dependent Claim
46 words
The system of claim 10 , wherein the classification is performed on a set of video frames of a video sequence, and the most likely classification result emerging out of the classification results of the individual frames is taken as the class of the video sequence.
Claim 15
Independent Claim
148 words
A system to classify an image comprising: a processor; a storage device coupled to the processor; a training system coupled to the processor to train classification of at least one image as either a natural image or an artificial image based on a combination of a noise feature vector and a sharpness feature vector, the noise feature vector including a pixel histogram of a difference between a first image and a noise-reduced second image generated from the first image, and the sharpness feature vector including a pixel histogram of a difference between the first image and a blurred third image generated from the first image; and a classification component coupled to the processor to classify the at least one image as either a natural image or an artificial image based on the trained classification and utilizing a noise vector and a sharpness vector of the least one image.
Claim 16
Independent Claim
148 words
A system to classify an image comprising: a processor; a storage device coupled to the processor; a training system coupled to the processor to train classification of at least one image as either a natural image or an artificial image based on a combination of a noise feature vector and a sharpness feature vector, the noise feature vector including a pixel histogram of a difference between a first image and a noise-reduced second image generated from the first image, and the sharpness feature vector including a pixel histogram of a difference between the first image and a blurred third image generated from the first image; and a classification component coupled to the processor to classify the at least one image as either a natural image or an artificial image based on the trained classification and utilizing a noise vector and a sharpness vector of the least one image.
Claim 17
Dependent Claim
22 words
The system of claim 12 , wherein the at least one image further comprises at least one frame in a video stream.
Claim 18
Dependent Claim
157 words
A computer-readable medium encoded with a computer program having computer-executable instructions to cause a computer to perform a method comprising: measuring noise and edge sharpness in a first image; generating a feature vector from the first image by; generating a noise-reduced second image from the first image; calculating a difference between the first image and the second image; generating a noise pixel histogram of the difference to use as a noise feature vector; generating a blurred third image from the first image; calculating another difference between the first image and the third image; generating a sharpness pixel histogram of the another difference to use as a sharpness feature vector; and combining the noise feature vector and the sharpness feature vector as the feature vector; and training a classification model from the feature vector to classify a fourth image as a natural image versus an artificial image from the noise and the edge sharpness in the fourth image.
Claim 19
Dependent Claim
157 words
A computer-readable medium encoded with a computer program having computer-executable instructions to cause a computer to perform a method comprising: measuring noise and edge sharpness in a first image; generating a feature vector from the first image by; generating a noise-reduced second image from the first image; calculating a difference between the first image and the second image; generating a noise pixel histogram of the difference to use as a noise feature vector; generating a blurred third image from the first image; calculating another difference between the first image and the third image; generating a sharpness pixel histogram of the another difference to use as a sharpness feature vector; and combining the noise feature vector and the sharpness feature vector as the feature vector; and training a classification model from the feature vector to classify a fourth image as a natural image versus an artificial image from the noise and the edge sharpness in the fourth image.
Claim 20
Dependent Claim
19 words
The computer-readable medium of claim 14 , wherein the first image further comprises a frame in a video stream.
Claim 21
Dependent Claim
139 words
A computer-readable medium encoded with a computer program having computer-executable instructions to cause a computer to perform a method comprising: receiving a feature vector of a first image, wherein the feature vector is a combination of a noise vector and sharpness vector of the first image; classifying the first image as natural image versus an artificial image from the feature vector; and generating the classification of the image; wherein the classifying is based on training from a combination of a noise feature vector and a sharpness feature vector, the noise feature vector including a pixel histogram of a difference between a second image and a noise-reduced third image generated from the second image, and the sharpness feature vector including a pixel histogram of a difference between the second image and a blurred fourth image generated from the second image.
Claim 22
Dependent Claim
139 words
A computer-readable medium encoded with a computer program having computer-executable instructions to cause a computer to perform a method comprising: receiving a feature vector of a first image, wherein the feature vector is a combination of a noise vector and sharpness vector of the first image; classifying the first image as natural image versus an artificial image from the feature vector; and generating the classification of the image; wherein the classifying is based on training from a combination of a noise feature vector and a sharpness feature vector, the noise feature vector including a pixel histogram of a difference between a second image and a noise-reduced third image generated from the second image, and the sharpness feature vector including a pixel histogram of a difference between the second image and a blurred fourth image generated from the second image.
Claim 23
Dependent Claim
47 words
The computer-readable medium of claim 16 , wherein the classification is performed on a set of video frames of a video sequence, and the most likely classification result emerging out of the classification results of the individual frames is taken as the class of the video sequence.
Claim 24
Independent Claim
139 words
An apparatus to train classification of at least one image, comprising: a processor; a storage device coupled to the processor; and a training system coupled to the processor to train classification of at least one image as either a natural image or an artificial image, wherein the training system further comprises a generator of a noise feature vector and sharpness feature vector of the at least one image; wherein the classification is based on training from a combination of the noise feature vector and the sharpness feature vector, the noise feature vector including a pixel histogram of a difference between a first image and a noise-reduced second image generated from the first image, and the sharpness feature vector including a pixel histogram of a difference between the first image and a blurred third image generated from the first image.
Claim 25
Independent Claim
139 words
An apparatus to train classification of at least one image, comprising: a processor; a storage device coupled to the processor; and a training system coupled to the processor to train classification of at least one image as either a natural image or an artificial image, wherein the training system further comprises a generator of a noise feature vector and sharpness feature vector of the at least one image; wherein the classification is based on training from a combination of the noise feature vector and the sharpness feature vector, the noise feature vector including a pixel histogram of a difference between a first image and a noise-reduced second image generated from the first image, and the sharpness feature vector including a pixel histogram of a difference between the first image and a blurred third image generated from the first image.
Claim 26
Dependent Claim
22 words
The apparatus of claim 18 , wherein the at least one image further comprises at least one frame in a video stream.
Claim 27
Independent Claim
134 words
An apparatus to classify an image, comprising: a processor; a storage device coupled to the processor; and a classification component coupled to the processor to classify a first image as either a natural image or an artificial image, wherein the classification component further comprises a feature vector generator to generate both of a noise vector and a sharpness vector of the first image; wherein the classifying is based on training from a combination of a noise feature vector and a sharpness feature vector, the noise feature vector including a pixel histogram of a difference between a second image and a noise-reduced third image generated from the second image, and the sharpness feature vector including a pixel histogram of a difference between the second image and a blurred fourth image generated from the second image.
Claim 28
Independent Claim
134 words
An apparatus to classify an image, comprising: a processor; a storage device coupled to the processor; and a classification component coupled to the processor to classify a first image as either a natural image or an artificial image, wherein the classification component further comprises a feature vector generator to generate both of a noise vector and a sharpness vector of the first image; wherein the classifying is based on training from a combination of a noise feature vector and a sharpness feature vector, the noise feature vector including a pixel histogram of a difference between a second image and a noise-reduced third image generated from the second image, and the sharpness feature vector including a pixel histogram of a difference between the second image and a blurred fourth image generated from the second image.
Claim 29
Dependent Claim
45 words
The apparatus of claim 20 , wherein the classifying is performed on a set of video frames of a video sequence, and the most likely classification result emerging out the classification results of the individual frames is taken as the class of the video sequence.
References Cited (Backward Citations)
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2010-01-29 | 2010-07-31 | 2011-01-30 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2014-01-29 | 2014-07-31 | 2015-01-30 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2018-01-29 | 2018-07-31 | 2019-01-30 | $7700.00 | Window Closed / Lapsed |