Content aware, spatially adaptive automated thresholding of images

1. A computer-implemented method for adaptively thresholding image pixels, the method comprising:

obtaining at least one image using at least one processor;

detecting, using the at least one processor, a plurality of objects defined by a plurality of pixels within the at least one image;

determining, using the at least one processor, perceptual information associated with each of the plurality of objects;

generating, using the perceptual information, a context for each of the plurality of pixels that define the plurality of objects, wherein generating the context includes generating a saliency map based on a determined Euclidean distance and mean associated with at least a portion of pixels corresponding to a filtered version of the at least one image; and

automatically thresholding, using the context for each respective pixel, the at least one image to generate an output image that represents the at least one image, the output image being a binarized version of the at least one image.

2. The method of claim 1 , wherein the context is further based on the perceptual information associated with at least one of the plurality of objects, the perceptual information including a color variance associated with at least one of the plurality of objects.

3. The method of claim 1 , wherein the context indicates whether a respective pixel is represented in black or white in the output image.

4. The method of claim 1 , further comprising performing a nearest neighbor estimation using the saliency map, the nearest neighbor estimation including:

defining a minimum size to a neighborhood used in the nearest neighbor estimation;

defining a maximum size to the neighborhood used in the nearest neighbor estimation;

estimating, for each pixel in the saliency map, at least one neighbor pixel;

calculating a maximized variance for a plurality of pixels located within the defined minimum size and the defined maximum size; and

binarizing each pixel in the at least one image according to the respective estimation for each pixel, the binarization generating the output image being corrected for degraded image content in the at least one image.

5. The method of claim 1 , wherein the thresholding is spatially adaptable based on the defined minimum size of the neighborhood and the defined maximum size of the neighborhood.

6. The method of claim 1 , wherein the automatic thresholding is performed on an embedded computing device.

7. A computer program product embodied on a non-transitory computer-readable storage medium and comprising an introspection network and instructions that, when executed by at least one computing device, is configured to cause the at least one computing device to: obtain at least one image using at least one processor; detect, using the at least one processor, a plurality of objects defined by a plurality of pixels within the at least one image; determine, using the at least one processor, perceptual information associated with each of the plurality of objects; generate, using the perceptual information, a context for each of the plurality of pixels that define the plurality of objects, wherein generating the context includes generating a saliency map based on a determined Euclidean distance and mean associated with at least a portion of pixels corresponding to a filtered version of the at least one image; and automatically threshold, using the context for each respective pixel, the at least one image to generate an output image that represents the at least one image, the output image being a binarized version of the at least one image.

8. The computer program product of claim 7 , wherein the context is further based on the perceptual information associated with at least one of the plurality of objects, the perceptual information including a color variance associated with at least one of the plurality of objects.

9. The computer program product of claim 7 , wherein the context indicates whether a respective pixel is represented in black or white in the output image.

10. The computer program product of claim 7 , further comprising performing a nearest neighbor estimation using the context, the nearest neighbor estimation including:

defining a minimum size to a neighborhood used in the nearest neighbor estimation;

defining a maximum size to the neighborhood used in the nearest neighbor estimation;

estimating, for each pixel represented in the context, at least one neighbor pixel;

calculating a maximized variance for a plurality of pixels located within the defined minimum size and the defined maximum size; and

binarizing each pixel in the at least one image according to the respective estimation for each pixel, the binarization generating the output image being corrected for degraded image content in the at least one image.

11. The computer program product of claim 10 , wherein the thresholding is spatially adaptable based on the defined minimum size of the neighborhood and the defined maximum size of the neighborhood.

12. The computer program product of claim 7 , wherein the automatic thresholding is performed on an embedded computing device.

13. A system comprising: at least one memory including instructions; and at least one processor that is operably coupled to the at least one memory and that is arranged and configured to execute instructions that, when executed, cause the at least one processor to: receive at least one image; and in response to receiving the at least one image, automatically: detect a plurality of objects defined by a plurality of pixels within the at least one image; determine perceptual information associated with each of the plurality of objects; generate, using the perceptual information, a context for each of the plurality of pixels that define the plurality of objects, wherein generating the context includes generating a saliency map based on a determined Euclidean distance and mean associated with at least a portion of pixels corresponding to a filtered version of the at least one image; threshold, using the context for each respective pixel, the at least one image to generate an output image that represents the at least one image, the output image being a binarized version of the at least one image; and trigger display of the generated output image.

14. The system of claim 13 , wherein the context is further based on the perceptual information associated with at least one of the plurality of objects, the perceptual information including a color variance associated with at least one of the plurality of objects.

15. The system of claim 13 , wherein the context indicates whether a respective pixel is represented in black or white in the output image.

16. The system of claim 13 , further comprising performing a nearest neighbor estimation using the saliency map, the nearest neighbor estimation including:

defining a minimum size to a neighborhood used in the nearest neighbor estimation;

defining a maximum size to the neighborhood used in the nearest neighbor estimation;

estimating, for each pixel in the saliency map, at least one neighbor pixel;

calculating a maximized variance for a plurality of pixels located within the defined minimum size and the defined maximum size; and

binarizing each pixel in the at least one image according to the respective estimation for each pixel, the binarization generating the output image being corrected for degraded image content in the at least one image.

17. The system of claim 16 , wherein the thresholding is spatially adaptable based on the defined minimum size of the neighborhood and the defined maximum size of the neighborhood.