Methods and systems for detecting regions in digital images

Abstract

Embodiments of the present invention comprise systems, methods and devices for detection of image regions of various content types using a masking condition and an entropy measure.

Description

FIELD OF THE INVENTION

Embodiments of the present invention comprise methods and systems for automatically detecting regions in digital images.

BACKGROUND

The content of a digital image can have considerable impact on the compression of the digital image, both in terms of compression efficiency and compression artifacts. Pictorial regions in an image are not efficiently compressed using compression algorithms designed for the compression of text. Similarly, text images are not efficiently compressed using compression algorithms that are designed and optimized for pictorial content. Not only is compression efficiency affected when a compression algorithm designed for one type of image content is used on a different type of image content, but the decoded image may exhibit annoying compression artifacts.

Further, image enhancement algorithms designed to sharpen text, if applied to pictorial image content, may produce visually annoying artifacts in some areas of the pictorial content. In particular, those areas of the pictorial content containing strong edges may be affected. While smoothing operations may enhance a natural image, the smoothing of text regions is seldom desirable.

The detection of regions of a particular content type in a digital image can improve compression efficiency, reduce compression artifacts, and improve image quality when used in conjunction with a compression algorithm or image enhancement algorithm designed for the particular type of content.

The semantic labeling of image regions based on content is also useful in document management systems and image databases.

Reliable and efficient detection of regions of pictorial content type and other image regions in digital images is desirable.

SUMMARY

Embodiments of the present invention comprise methods and systems for classifying an image into regions using a masking condition and an entropy measure.

The foregoing and other objectives, features, and advantages of the invention will be more readily understood upon consideration of the following detailed description of the invention taken in conjunction with the accompanying drawings.

BRIEF DESCRIPTION OF THE SEVERAL DRAWINGS

FIG. 1 is an example of an image comprising a multiplicity of regions of different content type;

FIG. 2 is a diagram of an exemplary region-detection system (prior art);

FIG. 3 is an exemplary histogram showing feature value separation;

FIG. 4 is an exemplary histogram showing feature value separation;

FIG. 5 is a diagram showing exemplary embodiments of the present invention comprising a masked-entropy calculation from a histogram;

FIG. 6 is a diagram showing an exemplary embodiment of masked-image generation;

FIG. 7 is a diagram showing an exemplary embodiment of histogram generation;

FIG. 8 is diagram showing pixel classification comprising an image window;

FIG. 9 is a diagram showing block classification comprising an image window;

FIG. 10 is a diagram showing exemplary embodiments of the present invention comprising a masked-entropy calculation from a histogram using confidence levels;

FIG. 11 is a diagram showing an exemplary embodiment of masked-image generation using confidence levels; and

FIG. 12 is a diagram showing an exemplary embodiment of histogram generation using confidence levels.

DETAILED DESCRIPTION OF EXEMPLARY EMBODIMENTS

Embodiments of the present invention will be best understood by reference to the drawings, wherein like parts are designated by like numerals throughout. The figures listed above are expressly incorporated as part of this detailed description.

It will be readily understood that the components of the present invention, as generally described and illustrated in the figures herein, could be arranged and designed in a wide variety of different configurations. Thus, the following more detailed description of the embodiments of the methods and systems of the present invention is not intended to limit the scope of the invention but it is merely representative of the presently preferred embodiments of the invention.

Elements of embodiments of the present invention may be embodied in hardware, firmware and/or software. While exemplary embodiments revealed herein may only describe one of these forms, it is to be understood that one skilled in the art would be able to effectuate these elements in any of these forms while resting within the scope of the present invention.

FIG. 1 shows an image 10 comprising three regions: a pictorial region 12, a text region 14, and a graphics region 16. For many image processing, compression, document management, and other applications, it may be desirable to detect various regions in an image. Exemplary regions may include: a pictorial region, a text region, a graphics region, a half-tone region, a continuous-tone region, a color region, a black-and-white region, a region best compressed by Joint Photographic Experts Group (JPEG) compression, a region best compressed by Joint Bi-level Image Experts Group (JBIG) compression, a background region, and a foreground region.

An exemplary region-detection system 20 is shown in FIG. 2. A region-detection system 20 may include a feature extractor 22 and a classifier 24. The feature extractor 22 may measure, calculate, or in some way extract, a feature or features 23 from the image. The classifier 24 may classify portions of the image based on the extracted feature or features 23. The classification 25 produced by the classifier 24 thereby provides detection of image regions.

The effectiveness and reliability of a region-detection system may depend on the feature or features used for the classification. FIG. 3 shows an example of a normalized frequency-of-occurrence plot of feature values for two image regions. The solid line 32 shows the frequency of occurrence of feature values extracted from image samples belonging to one region. The dashed line 34 shows the frequency of occurrence of feature values extracted from image samples belonging to a second region. The strong overlap of these two histograms may indicate that the feature may not be an effective feature for separating image samples belonging to one of these two regions.

FIG. 4 shows another example of a normalized frequency of occurrence plot of feature values for two image regions. The solid line 42 shows the frequency of occurrence of feature values extracted from image samples belonging to one region. The dashed line 44 shows the frequency of occurrence of feature values extracted from image samples belonging to a second region. The wide separation of these two histograms may indicate that the feature will be an effective feature for classifying image samples as belonging to one of these two regions.

Embodiments of the present invention comprise methods and systems for region detection in a digital image. Some embodiments of the present invention comprise methods and systems for region detection in a digital image wherein the separation between feature values corresponding to image regions may be accomplished by masking, prior to feature extraction, pixels in the image for which a masking condition is met. In some embodiments, the masked pixel values may not be used when extracting the feature value from the image.

In exemplary embodiments of the present invention shown in FIG. 5, a masked image 51 may be formed 52 from an input image 50. The masked image may be formed by checking a masking condition at each pixel in the input image 50. An exemplary embodiment shown in FIG. 6, illustrates the formation of the masked image 52. If an input image pixel 60 satisfies the masking condition 62, the value of the pixel at the corresponding location in the masked image may be assigned a value, called a mask-pixel value, 66, indicating that the masking condition is satisfied at that pixel location in the input image. If an input image pixel 60 does not satisfy the masking condition 64, the value of the pixel at the corresponding location in the masked image may be assigned the value of the input pixel in the input image 68. The masked image thereby masks pixels in the input image for which a masking condition is satisfied.

In the exemplary embodiments of the present invention shown in FIG. 5, after forming 52 the masked image 51, a histogram 53 may be generated 54 for a block, also considered a segment, section, or any division, not necessarily rectangular in shape, of the masked image 51. For the purposes of this specification, associated claims, and included drawings, the term block will be used to describe a portion of data of any shape including, but not limited to, square, rectangular, circular, elliptical, or approximately circular. FIG. 7 shows an exemplary embodiment of histogram formation. A histogram with bins corresponding to the possible pixel values of the masked image may be formed according to FIG. 7. In some embodiments, all bins may be initially considered empty with initial count zero. The value of a pixel 70 in the block of the masked image may be compared 71 to the mask-pixel value. If the value of the pixel 70 is equal 72 to the mask-pixel value, then the pixel is not accumulated in the histogram, meaning that no histogram bin is incremented, and if there are pixels remaining in the block to examine 76, then the next pixel in the block is examined 71. If the value of the pixel 70 is not equal 73 to the mask-pixel value, then the pixel is accumulated in the histogram 74, meaning that the histogram bin corresponding to the value of the pixel is incremented, and if there are pixels remaining in the block to examine 77, then the next pixel is examined 71.

When a pixel is accumulated in the histogram 74, a counter for counting the number of non-mask pixels in the block of the masked image is incremented 75. When all pixels in a block have been examined 78, 79, the histogram may be normalized 80. The histogram may be normalized 80 by dividing each bin count by the number of non-mask pixels in the block of the masked image. In alternate embodiments, the histogram may not be normalized and the counter may not be present.

An entropy measure 55 may be calculated 56 for the histogram 53 of a block of the masked image. The entropy measure 55 may be considered an image feature of the input image. The entropy measure 55 may be considered any measure of the form:

$-\sum _{i=1}^{N}h\left(i\right)*f\left(h\left(i\right)\right),$ where N is the number of histogram bins, h(i) is the accumulation or count of bin i, and ƒ(·) may be a function with mathematical characteristics similar to a logarithmic function. The entropy measure 55 may be weighted by the proportion of pixels that would have been counted in a bin, but were masked. The entropy measure is of the form:

$-\sum _{i=1}^{N}w\left(i\right)h\left(i\right)*f\left(h\left(i\right)\right)$ where w(i) is the weighting function. In some embodiments of the present invention, the function ƒ(h(i)) may be log₂(h(i)).

In the embodiments of the present invention shown in FIG. 5, after calculating 56 the entropy measure 55 for the histogram 53 corresponding to a block of the image centered at a pixel, the pixel may be classified 57 according to the entropy feature 55. The classifier 57 shown in FIG. 5 may be based on thresholding. A threshold may be determined a priori, adaptively, or by any of numerous methods. The pixel may be classified 57 as belonging to one of two regions depending on which side of the threshold the entropy measure 55 falls.

In some embodiments of the present invention, a moving window of pixel values centered, in turn, on each pixel of the image, may be used to calculate the entropy measure for the block containing the centered pixel. The entropy may be calculated from the corresponding block in the masked image. The entropy value may be used to classify the pixel at the location on which the moving window is centered. FIG. 8 shows an exemplary embodiment in which a block of pixels is used to measure the entropy feature which is used to classify a single pixel in the block. In FIG. 8, a block 80 is shown for an image 81. The pixels in the masked image in the corresponding block may be used to calculate the entropy measure. The pixel in the center of the image 82 may be classified according the entropy measure.

In other embodiments of the present invention, the entropy value may be calculated for a block of the image, and all pixels in the block may be classified with the same classification based on the entropy value. FIG. 9 shows an exemplary embodiment in which a block of pixels is used to measure the entropy feature which is used to classify all pixels in the block. In FIG. 9, a block 90 is shown for an image 91. The pixels in the masked image in the corresponding block may be used to calculate the entropy measure. All pixels 92 in the block 90 may be classified according to the entropy measure.

In some embodiments of the present invention, the masking condition may be based on the edge strength at a pixel.

In some embodiments of the present invention, a level of confidence in the degree to which the masking condition is satisfied may be calculated. The level of confidence may be used when accumulating a pixel into the histogram. Exemplary embodiments in which a level of confidence is used are shown in FIG. 10.

In exemplary embodiments of the present invention shown in FIG. 10, a masked image 101 may be formed 102 from an input image 100. The masked image may be formed by checking a masking condition at each pixel in the input image 100. An exemplary embodiment shown in FIG. 11, illustrates the formation of the masked image 102. If an input image pixel 110 satisfies the masking condition 112, the corresponding pixel in the masked image may be assigned a value, mask-pixel value, 116 indicating that the masking condition is satisfied at that pixel. If an input image pixel 110 does not satisfy the masking condition 114, the corresponding pixel in the masked image may be assigned the value of the corresponding pixel in the input image 118. At pixels for which the masking condition is satisfied 112, a further assignment 115 of a confidence value reflecting the confidence in the mask signature signal may be made. The assignment of confidence value may be a separate value for the masked pixels, or the mask-pixel value may be multi-level with the levels representing the confidence. The masked image may mask pixels in the input image for which a masking condition is satisfied, and further identify the level to which the masking condition is satisfied.

In the exemplary embodiments of the present invention shown in FIG. 10, after forming 102 the masked image 101, a histogram 103 may be generated 104 for a block of the masked image 101. FIG. 12 shows an exemplary embodiment of histogram formation 104. A histogram with bins corresponding to the possible pixel values of the masked image may be formed according to FIG. 12. In some embodiments, all bins may be initially considered empty with initial count zero. The value of a pixel 120 in the block of the masked image may be compared 121 to the mask-pixel value. If the value of the pixel 120 is equal 122 to the mask-pixel value, then the pixel is accumulated 123 in the histogram at a fractional count based on the confidence value, and if there are pixels remaining in the block to examine 126, then the next pixel in the block is examined 121. If the value of the pixel 120 is not equal 123 to the mask-pixel value, then the pixel is accumulated in the histogram 124, meaning that the histogram bin corresponding to the value of the pixel is incremented, and if there are pixels remaining in the block to examine 127, then the next pixel in the block is examined 121.

When a pixel is accumulated in the histogram 124, a counter for counting the number of non-mask pixels in the block of the masked image is incremented 125. When all pixels in a block have been examined 128, 129, the histogram may be normalized 130. The histogram may be normalized 130 by dividing each bin count by the number of non-mask pixels in the block of the masked image. In alternate embodiments, the histogram may not be normalized and the counter not be present.

An entropy measure 105 may be calculated 106 for the histogram of a neighborhood of the masked image as described in the previous embodiments. In the embodiments of the present invention shown in FIG. 10, after calculating 106 the entropy measure 105 for the histogram 103 corresponding to a block of the image centered at a pixel, the pixel may be classified 107 according to the entropy feature 105. The classifier 107 shown in FIG. 10 may be based on thresholding. A threshold may be determined a priori, adaptively, or by any of numerous methods. The pixel may be classified 107 as belonging to one of two regions depending on which side of the threshold the entropy measure 105 falls.

In some embodiments of the present invention, the masking condition may comprise a single image condition. In some embodiments, the masking condition may comprise multiple image conditions combined to form a masking condition.

In some embodiments of the present invention, the entropy feature may be used to separate the image into two regions. In some embodiments of the present invention, the entropy feature may be used to separate the image into more than two regions.

The terms and expressions which have been employed in the foregoing specification are used therein as terms of description and not of limitation, and there is no intention in the use of such terms and expressions of excluding equivalence of the features shown and described or portions thereof, it being recognized that the scope of the invention is defined and limited only by the claims which follow.

Claims

1. A method for calculating an entropy feature in a digital image comprising a plurality of pixels, said method comprising: a) in a processor, applying a masking condition to said digital image thereby identifying masked pixels that satisfy said masking condition;b) in a calculator, calculating an entropy measure associated with a center pixel within a block of pixels in said digital image, wherein all masked pixels within said block contribute less to said entropy measure than pixels, within said block, not identified by said masking condition; andc) assigning a confidence value to a first masked pixel in said block, wherein said confidence value assigned to said first masked pixel reflects a confidence in a mask signature signal at said first masked pixel.

2. A method as described in claim 1 wherein said calculating excludes said all masked pixels within said block.

3. A method as described in claim 1 wherein said contribution of said first masked pixel is related to said assigned confidence value.

4. A method as described in claim 3 wherein said calculating comprises generating a histogram of pixel values.

5. A method as described in claim 1 wherein said calculating comprises generating a histogram of pixel values.

6. A method as described in claim 1 wherein said masking condition comprises edge strength.

7. A system for calculating an entropy feature in a digital image, said system comprising: a) processor for applying a masking condition to said digital image thereby identifying masked pixels that satisfy said masking condition;b) a calculator for calculating an entropy measure associated with a center pixel within a block of pixels in said digital image, wherein all masked pixels within said block contribute less to said entropy measure than pixels, within said block, not identified by said masking condition; andc) a confidence-value assigner for assigning a confidence value to a first masked pixel in said block, wherein said confidence value assigned to said first masked pixel reflects a confidence in a mask signature signal at said first masked pixel.

8. A system as described in claim 7 wherein said calculating excludes said all masked pixels within said block.

9. A system as described in claim 7 wherein said contribution of said first masked pixel is related to said confidence value.

10. A system as described in claim 9 wherein said calculator comprises a histogram generator for generating a histogram of pixel values.

11. A system as described in claim 10 wherein said histogram generator accumulates a pixel value associated with said first masked pixels based on said confidence value.

12. A system as described in claim 7 wherein said calculator comprises a histogram generator for generating a histogram of pixel values.

13. A system as described in claim 12 wherein said histogram generator accumulates pixel values from only pixels, from said block, for which said masking condition is not met.

14. A system as described in claim 7 wherein said masking condition comprises edge strength.

15. A method for detecting a region in a digital image, said method comprising: a) in a region classifier, applying a masking condition to said digital image thereby identifying masked pixels that satisfy said masking condition;b) assigning a confidence value to a first masked pixel within a neighbor of said digital image, wherein said confidence value reflects a confidence in a mask signature signal at said first masked pixel;c) calculating an entropy measure associated with said neighborhood of said digital image, wherein all masked pixels within said neighborhood contribute less to said entropy measure than pixels, within said neighborhood, not identified by said masking condition; andd) classifying said neighborhood of said digital image based on said entropy measure.

16. A method as described in claim 15 wherein said calculating excludes said all masked pixels within said neighborhood.