Extracting multiple documents from single image

1. A computer-implemented method for document image detection, comprising:

producing, using a neural network, a superpixel segmentation map of an input image, wherein each superpixel of the superpixel segmentation map is associated with a list of probability characteristics, such that each probability characteristic of the list of probability characteristics represents a probability of the superpixel belonging to a visual object found in the input image, wherein the visual object is identified by an index of the probability characteristic in the list of probability characteristics;

generating a superpixel binary mask by associating, based on a corresponding list of probability characteristics, each superpixel of the superpixel segmentation map with a class of a predetermined set of classes;

identifying one or more connected components in the superpixel binary mask;

for each connected component of the superpixel binary mask, identifying a corresponding minimum bounding polygon, wherein identifying the minimum bounding polygon further comprises:

responsive to determining that a first number of pixels in a first line of the superpixel binary mask exceeds, by at least a predetermined threshold, a second number of pixels in a second line of the superpixel binary mask which is adjacent to the first line of the superpixel binary mask, utilizing the second line as a candidate boundary of the minimum bounding polygon, wherein the first line is provided by one of: a row of the superpixel binary mask or a column of the superpixel binary mask, and

computing a value of a quality metric for a set of regions of interest that are defined using a plurality of candidate lines comprising the second line of the superpixel binary mask;

creating one or more image dividing lines based on the minimum bounding polygons; and

defining boundaries of one or more objects of interest based on at least a subset of the image dividing lines.

2. The method of claim 1 , wherein the neural network comprises:

a downscale block;

a context block; and

a final classification block.

3. The method of claim 2 , wherein the neural network further comprises a rectifier activation function.

4. The method of claim 1 , further comprising:

cropping each region of interest of one or more regions of interest to produce a corresponding document image.

5. The method of claim 4 , further comprising:

determining whether two or more regions of interest belong to a single multi-part document.

6. The method of claim 1 , wherein the neural network is trained using augmented images.

7. The method of claim 1 , wherein generating the plurality of candidate lines for the minimum bounding polygon further comprises:

utilizing, as a candidate boundary of the minimum bounding polygon, a line traversing a center of the superpixel binary mask.

8. The method of claim 1 , wherein computing a value of a quality metric for the set of regions of interest further comprises:

applying, to the set of regions of interest, a trainable classifier.

9. A system, comprising:

a memory;

a processor, coupled to the memory, the processor configured to:

produce, using a neural network, a superpixel segmentation map of an input image, wherein each superpixel of the superpixel segmentation map is associated with a list of probability characteristics, such that each probability characteristic of the list of probability characteristics represents a probability of the superpixel belonging to a visual object found in the input image, wherein the visual object is identified by an index of the probability characteristic in the list of probability characteristics;

generate a superpixel binary mask by associating, based on a corresponding list of probability characteristics, each superpixel of the superpixel segmentation map with a class of a predetermined set of classes;

identify one or more connected components in the superpixel binary mask, wherein identifying the minimum bounding polygon further comprises:

responsive to determining that a first number of pixels in a first line of the superpixel binary mask exceeds, by at least a predetermined threshold, a second number of pixels in a second line of the superpixel binary mask which is adjacent to the first line of the superpixel binary mask, utilizing the second line as a candidate boundary of the minimum bounding polygon, wherein the first line is provided by one of: a row of the superpixel binary mask or a column of the superpixel binary mask, and

computing a value of a quality metric for a set of regions of interest that are defined using a plurality of candidate lines comprising the second line of the superpixel binary mask;

for each connected component of the superpixel binary mask, identifying a corresponding minimum bounding polygon;

create one or more image dividing lines based on the minimum bounding polygons; and

define boundaries of one or more objects of interest based on at least a subset of the image dividing lines.

10. The system of claim 9 , wherein the neural network comprises:

a downscale block;

a context block; and

a final classification block.

11. The system of claim 10 , wherein the neural network further comprises a rectifier activation function.

12. The system of claim 9 , further comprising:

cropping each region of interest of one or more regions of interest to produce a corresponding document image.

13. The system of claim 12 , further comprising:

determining whether two or more regions of interest belong to a single multi-part document.

14. The system of claim 9 , wherein the neural network is trained using augmented images.

15. A non-transitory computer-readable storage medium comprising executable instructions that, when executed by a computer system, cause the computer system to:

produce, using a neural network, a superpixel segmentation map of an input image, wherein each superpixel of the superpixel segmentation map is associated with a list of probability characteristics, such that each probability characteristic of the list of probability characteristics represents a probability of the superpixel belonging to a visual object found in the input image, wherein the visual object is identified by an index of the probability characteristic in the list of probability characteristics;

generate a superpixel binary mask by associating, based on a corresponding list of probability characteristics, each superpixel of the superpixel segmentation map with a class of a predetermined set of classes;

identify one or more connected components in the superpixel binary mask wherein identifying the minimum bounding polygon further comprises:

responsive to determining that a first number of pixels in a first line of the superpixel binary mask exceeds, by at least a predetermined threshold, a second number of pixels in a second line of the superpixel binary mask which is adjacent to the first line of the superpixel binary mask, utilizing the second line as a candidate boundary of the minimum bounding polygon, wherein the first line is provided by one of: a row of the superpixel binary mask or a column of the superpixel binary mask, and

computing a value of a quality metric for a set of regions of interest that are defined using a plurality of candidate lines comprising the second line of the superpixel binary mask;

for each connected component of the superpixel binary mask, identifying a corresponding minimum bounding polygon;

create one or more image dividing lines based on the minimum bounding polygons; and

define boundaries of one or more objects of interest based on at least a subset of the image dividing lines.