Content-aware type-on-path generation along object contours

Patent metadata

Field	Value
Publication number	US-12148203-B2
Application number	US-202217744964-A
Country	US
Kind code	B2
Filing date	May 16, 2022
Priority date	May 16, 2022
Publication date	Nov 19, 2024
Grant date	Nov 19, 2024

Abstract

A method for content-aware type-on-path generation is implemented via a computing system including a processor. The method includes surfacing an image via a graphics GUI of a graphics application and detecting one or more salient objects within the image using a CNN model. The method also includes generating a contour map for each detected salient object and generating a path along the contours of each salient object by applying a defined offset to the corresponding contour map. The method further includes applying input text characters as type-on-path along the generated path based at least on user input received via the graphics GUI.

First claim

What is claimed is:

1. A method for content-aware type-on-path generation, wherein the method is implemented via a computing system comprising a processor, and wherein the method comprises:
surfacing an image via a graphics graphical user interface (GUI) of a graphics application;
detecting a salient object within the image using a convolutional neural network (CNN) model;
generating a contour map for the detected salient object;
generating a path along contours of the salient object by applying a defined offset to the contour map; and
applying input text characters as type-on-path along the generated path based at least on user input received via the graphics GUI.

2. The method of claim 1 , wherein the CNN model comprises a region-based CNN (R-CNN) model.

3. The method of claim 1 , comprising training the CNN model using an image dataset comprising a plurality of object classes.

4. The method of claim 3 , comprising updating the CNN model based on new object classes added to the image dataset.

5. The method of claim 1 , wherein detecting the salient object within the image using the CNN model comprises:
analyzing the image using the CNN model;
generating potential object proposals by computing bounding areas around potential objects within the image;
computing saliency scores for each generated potential object proposal; and
selecting the object with the highest saliency score as the salient object within the image.

6. The method of claim 5 , wherein selecting the object with the highest saliency score as the salient object comprises selecting all objects with saliency scores above a predefined threshold as salient objects within the image.

7. The method of claim 1 , wherein generating the contour map for the detected salient object comprises:
generating a mask for the salient object based on a computed bounding area for the salient object;
extracting the contours of the salient object from the generated mask; and
generating the contour map for the salient object based on the extracted contours.

8. The method of claim 1 , wherein generating the path along the contours of the salient object by applying the defined offset to the contour map comprises:
generating first tangential lines tangentially to the contour map at predefined intervals along the contour map;
generating perpendicular lines perpendicularly to the first tangential lines, wherein a length of each perpendicular line is equal to the defined offset;
marking an end of each perpendicular line as a coordinate of the path; and
joining all the marked coordinates together to generate the path along the contours of the salient object.

9. The method of claim 1 , wherein applying the input text characters as the type-on-path along the generated path comprises:
generating second tangential lines tangentially to the path at predefined intervals along the path; and
applying the input text characters perpendicularly to the second tangential lines.

10. The method of claim 9 , comprising:
applying the input text characters perpendicularly to the second tangential lines at an initial font size;
computing an adjusted font size that enables all the input text characters to be applied evenly along the path;
recalculating the second tangential lines according to the adjusted font size; and
reapplying the input text characters perpendicularly to the second tangential lines at the adjusted font size.

11. The method of claim 1 , comprising:
executing, via a network, the graphics application on a remote computing system; and
surfacing the image on a display device of the remote computing system.

12. The method of claim 1 , comprising:
executing the graphics application locally on the computing system; and
surfacing the image on a display device of the computing system.

13. The method of claim 1 , comprising executing the method for a plurality of detected salient objects.

14. A computing system, comprising:
a processor; and
a computer-readable storage medium operatively coupled to the processor, the computer-readable storage medium comprising computer-executable instructions that, when executed by the processor, cause the processor to:
surface an image via a graphics graphical user interface (GUI) of a graphics application;
detect a salient object within the image using a convolutional neural network (CNN) model;
generate a contour map for the detected salient object;
generate a path along contours of the salient object by applying a defined offset to the contour map; and
apply input text characters as type-on-path along the generated path based at least on user input received via the graphics GUI.

15. The computing system of claim 14 , wherein the computer-readable storage medium comprises computer-executable instructions that, when executed by the processor, cause the processor to:
train the CNN model using an image dataset comprising a plurality of object classes; and
update the CNN model based on new object classes added to the image dataset.

16. The computing system of claim 14 , wherein the computer-readable storage medium comprises computer-executable instructions that, when executed by the processor, cause the processor to detect the salient object within the image using the CNN model by:
analyzing the image using the CNN model;
generating potential object proposals by computing bounding areas around potential objects within the image;
computing saliency scores for each generated potential object proposal; and
selecting the object with the highest saliency score as the salient object within the image.

17. The computing system of claim 14 , wherein the computer-readable storage medium comprises computer-executable instructions that, when executed by the processor, cause the processor to generate the contour map for the detected salient object by:
generating a mask for the salient object based on a computed bounding area for the salient object;
extracting the contours of the salient object from the generated mask; and
generating the contour map for the salient object based on the extracted contours.

18. The computing system of claim 14 , wherein the computer-readable storage medium comprises computer-executable instructions that, when executed by the processor, cause the processor to generate the path along the contours of the salient object by:
generating first tangential lines tangentially to the contour map at predefined intervals along the contour map;
generating perpendicular lines perpendicularly to the first tangential lines, wherein a length of each perpendicular line is equal to the defined offset;
marking an end of each perpendicular line as a coordinate of the path; and
joining all the marked coordinates together to generate the path along the contours of the salient object.

19. A computer-readable storage medium comprising computer-executable instructions that, when executed by a processor, cause the processor to:
surface an image via a graphics graphical user interface (GUI) of a graphics application;
detect a salient object within the image using a convolutional neural network (CNN) model;
generate a contour map for the detected salient object;
generate a path along contours of the salient object by applying a defined offset to the contour map; and
apply input text characters as type-on-path along the generated path based at least on user input received via the graphics GUI.

20. The computer-readable storage medium of claim 19 , wherein the computer-executable instruction

Assignees

• Microsoft Technology Licensing Llc

Inventors

• Sharma Mrinal Kumar

Classifications

• G06V10/774

  Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting · CPC title

• G06V10/42

  Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation · CPC title

• G06V10/7715

  Feature extraction, e.g. by transforming the feature space, e.g. multi-dimensional scaling [MDS]; Mappings, e.g. subspace methods · CPC title

• G06V10/454

  Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN] · CPC title

• G06V10/82 Primary

  using neural networks · CPC title