Methods of object search and recognition

1. A computer-implemented method for searching for elements in an electronic image of a form, the method comprising:

detecting, via instructions operating on a computing device, objects in the electronic image of the form;

allocating a position for each object in the electronic image;

selecting text objects;

recognizing text within the text objects; and

searching for elements in the electronic image of the form based on a structural description of the form, wherein said searching includes:

selecting an element in the structural description for the searching operation;

determining an algorithm for determining search constraints for the searching operation;

determining search constraints using the determined algorithm; and

searching for the element using one or more of the search constraints, wherein searching for the element includes:

searching using a spatial characteristic of a search area for the element;

searching using a parametric characteristic of the element;

searching using an absolute or relative spatial characteristic of the element; and

searching using a result of said text recognition within the text objects;

identifying variants of the element;

estimating, for the obtained variants of the element, a degree of correspondence of the respective variants to the element description.

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

determining an orientation of the electronic image.

3. The method of claim 2 , wherein all or a part of the elements of the structural description are used to determine the orientation of the electronic image.

4. The method of claim 3 , wherein an auxiliary description in brief form is used to determine the orientation of the electronic image.

5. The method of claim 3 , wherein determining the orientation of the electronic image whose objects of the description coincide with the objects in the electronic image with a sufficiently high quality rating is accepted as the correct orientation of the electronic image.

6. The method of claim 1 , wherein the method further comprises:

selecting a type of the form from several types; and wherein the structural description of the form is based on said selected type.

7. The method of claim 6 , wherein an auxiliary description in a brief form is used to determine the form type; and

wherein the method further comprises selecting a structural description from several structural descriptions.

8. The method of claim 7 , wherein the form type which corresponds to the current electronic image is selected based on comparing quality estimates of the structural descriptions.

9. The method of claim 2 , wherein the orientation of the electronic image is determined by evaluating a coincidence of objects in the electronic image with those in the structural description.

10. The method of claim 6 , wherein the type of the form which corresponds to the electronic image of the form is selected based on comparing quality estimates of structural descriptions.

11. The method of claim 1 , wherein an exact or an interval characteristic of the element is used.

12. The method of claim 1 , wherein one or more of the following spatial characteristics of the search area are used: a half plane, a rectangle, a circle, a polygon, or a combination thereof.

13. The method of claim 1 , wherein two or more variants of a compound element which have a best quality estimate are used for further ananlysis.

14. The method of claim 1 , wherein two or more variants of an element which have a best quality estimate are used for further analysis.

15. The method of claim 1 , wherein the method further compromises:

searching for an object corresponding to a next element of the structural description when no objects are detected in a region of the electronic image which is specified for a current element.

16. The method of claim 1 , wherein the method further comprises confirming that an object is missing in a designated search area in the electronic image of the form.

17. The method of claim 1 , wherein the recognizing text within the text objects is performed to a predefined degree sufficient for identifying a structural description for the form.

18. The method of claim 1 , wherein the searching for elements in the electronic image of the form based on a structural description of the form further includes:

estimating a total reliability for each of the obtained variants.

19. The method of claim 1 , wherein the searching for elements in the electronic image of the form based on a structural description of the form further includes:

analyzing variants of combinations of elements, wherein said analyzing variants includes:

generating a quality rating for each variant;

assigning, to each type of element, variants with a quality rating beyond a threshold for further analysis; and

searching for a variant of the element using a total quality rating of its composite parts.

20. The method of claim 19 , wherein quality ratings of variants are analyzed after generating the total quality rating of the variant; and wherein the method further comprises selecting a variant with a best quality rating.

21. The method of claim 19 , wherein searching for a next object is performed when there are no variants for a current element or when a total quality rating is lower than a pedefined level.

22. The method of claim 19 , wherein the seaching for elements in the electronic image of the form based on a structural description of the form further includes:

analyzing quality estimates of variants that previously had a quality rating that failed to exceed the threshold for further analysis.

23. The method of claim 1 , wherein the method further comprises:

generating hypotheses about the presence, completeness of composition or types of composite parts of the element; and

analyzing a correspondence of the composite parts to the hypothesis about the type of the element.