Method and apparatus for image binarization
Methods and apparatus for generating a binarization threshold are described. A set of input image pixels is processed. Edges including edge pixels are identified. In various embodiments histograms of pixel values and edge pixel values are generated and jointly used in determining an appropriate binarization threshold. In some embodiments a first histogram of edge pixel values is used to determine an interval of pixel values of interest to limit the set of pixel values used for determining a binarization threshold statistically. In some embodiments a first histogram corresponding to edge pixel values and a second histogram corresponding to image pixel values are generated, subjected to one or more smoothing operations, and truncated, based on information derived from the edge histogram. Characteristics of the histograms including, e.g., minima, maxima, are determined, evaluated, and used to generate a best global threshold.
1. An image binarization method, the method comprising:
identifying edges in an image, said edges including edge pixel values;
generating a histogram (H 2 E) of edge pixel values corresponding to the identified edges; and
generating, based on the generated histogram (H 2 E) of edge pixel values, a statistically based binarization threshold, generating the statistically based binarization threshold including:
i) determining from the histogram (H 2 E) of edge pixel values pixel value truncation points L and U; and
ii) generating the statistically based binarization threshold from pixel values in said image which are within a range of pixel values extending between L and U while excluding from use values outside said range;
performing one or more sequential smoothing operations on a histogram (H 1 ) of image pixel values to produce a smoothed image histogram of image pixel values;
performing one or more sequential smoothing operations on said histogram (H 2 E) of edge pixel values to produce a smoothed edge histogram of edge pixel values; and
setting an image binarization threshold to the generated statistically based binarization threshold or a threshold generated from the histogram (H 1 ) of image pixel values, said setting an image binarization threshold including deciding which one of the generated statistically based binarization threshold or the threshold generated from the histogram (H 1 ) of image pixel values to use as the image binarization threshold based at least in part on the smoothed image histogram of image pixel values and the smoothed edge histogram of edge pixel values.
2. The method of claim 1 , wherein at least some of said pixel values in said range are non-edge pixel values.
3. The method of claim 2 wherein generating the statistically based binarization threshold further includes using a clustering algorithm applied to the pixel values in said range.
4. The method of claim 1 , further comprising:
determining based on the histogram H 2 E of edge pixel values, an interval of pixel values of interest, said interval of pixel values of interest extending between a minimum pixel value of interest and a maximum pixel value of interest.
5. The method of claim 4 , wherein said interval of pixel values of interest includes a predetermined percentage of edge pixel values.
6. The method of claim 1 , further comprising:
computing a first derivative of said smoothed edge histogram of edge pixel values; and
determining histogram values corresponding to zero crossings of said first derivative of said smoothed edge histogram of edge pixel values.
7. The method of claim 1 , wherein performing one or more sequential smoothing operations on said histogram (H 1 ) of image pixel values includes:
computing a first derivative of said smoothed image histogram of image pixel values; and
determining histogram values corresponding to zero crossings of said first derivative of said smoothed image histogram of image pixel values.
8. The method of claim 7 , further comprising:
limiting zero crossing information used to generate said statistically based binarization threshold to zero crossings which correspond to pixel values in the range of pixel values extending between L and U; and
set the threshold generated from the histogram (H 1 ) of image pixel values to the location of the minimum in said smoothed image histogram.
9. The method of claim 1 , wherein said step of generating a statistically based binarization threshold is further based on pixel values of non-edge image areas as well as said edge pixel values.
10. The method of claim 1 , wherein identifying edges in the image includes:
generating a contrast value for each pixel value in the image; and
identifying as edge pixel values, pixel values which exceed a contrast threshold.
11. The method of claim 10 , wherein the contrast value of a pixel is a function of the value of the pixel and values of pixels in a neighborhood surrounding said pixel.
12. The method of claim 10 , wherein said contrast threshold is determined based on a plurality of contrast values corresponding to pixel values of said image.
13. The method of claim 12 , wherein said contrast threshold is generated from a histogram of said contrast values.
14. The method of claim 13 , wherein said contrast threshold is generated using an OTSU algorithm on the histogram of contrast values to produce said contrast threshold.
15. A non-transitory computer readable medium including processor executable instructions which when executed by a processor control said processor to:
identify edges in an image, said edges including edge pixel values;
generate a histogram (H 2 E) of edge pixel values corresponding to the identified edges;
generate, based on the histogram (H 2 E) of edge pixel values, a statistically based binarization threshold, as part of generating the statistically based binarization threshold, control the processor to:
i) determine from the histogram (H 2 E) of edge pixel values pixel value truncation points L and U; and
ii) generate the statistically based binarization threshold from pixel values in said image which are within a range of pixel values extending between L and U while excluding from use values outside said range;
perform one or more sequential smoothing operations on a histogram (H 1 ) of image pixel values to produce a smoothed image histogram of image pixel values;
perform one or more sequential smoothing operations on said histogram (H 2 E) of edge pixel values to produce a smoothed edge histogram of edge pixel values; and
set an image binarization threshold to the generated statistically based binarization threshold or a threshold generated from the histogram (H 1 ) of image pixel values, said setting the image binarization threshold including deciding which one of the generated statistically based binarization threshold or the threshold generated from the histogram (H 1 ) of image pixel values to use as the image binarization threshold based at least in part on the smoothed image histogram of image pixel values and the smoothed edge histogram of edge pixel values.
16. An apparatus for performing image binarization, the apparatus comprising:
a memory including an image to be binarized; and
a processor configured to:
identify edges in the image, said edges including edge pixel values;
generate a histogram (H 2 E) of edge pixel values corresponding to the identified edges;
generate, based on the histogram (H 2 E) of edge pixel values, a statistically based binarization threshold, as part of being configured to generate the statistically based binarization threshold said processor being further configured to:
i) determine from the histogram (H 2 E) of edge pixel values pixel value truncation points L and U; and
ii) generate the statistically based binarization threshold from pixel values in said image which are within a range of pixel values extending between L and U while excluding from use values outside said range;
perform one or more sequential smoothing operations on a histogram (H 1 ) of image pixel values to produce a smoothed image histogram of image pixel values;
perform one or more sequential smoothing operations on said histogram (H 2 E) of edge pixel values to produce a smoothed edge histogram of edge pixel values; and
set an image binarization threshold to the generated statistically based binarization threshold or a threshold generated from the histogram (H 1 ) of image pixel values, said setting the image binarization threshold including deciding which one of the generated statistically based binarization threshold or the threshold generated from the histogram (H 1 ) of image pixel values to use as the image binarization threshold based at least in part on the smoothed image histogram of image pixel values and the smoothed edge histogram of edge pixel values.
17. The apparatus of claim 16 , wherein said processor is further configured to:
perform a binarization operation on said image using said set image binarization threshold to generate a binarized image; and
transmit said binarized image or output said binarized image to a display.