Granted Patent B2

US 12,439,161 · App. 17/891,657 · Granted Oct 7, 2025

Inventors: Sandeep Suri (Noida, IN); Abhishek Pathak (Noida, IN); Chirag Bansal (Noida, IN); Abhijeet Padhy (Noida, IN); Ram Pravesh Singh (Noida, IN)

Assignee: Samsung Electronics Co., Ltd.

H04N23/698G06N5/047G06T3/4038G06T7/571G06V10/74G06V10/80H04N23/632H04N23/633H04N23/64H04N23/69H04N23/80G06N3/04H04N23/667

Abstract

Claims (78)

1. A method performed by an electronic device for generating a panoramic image, the method comprising:

obtaining, by the electronic device, a first scene preview by using a zoomed out feature of at least one camera sensor of the electronic device;

obtaining, by the electronic device, at least one second scene preview with a wider view than the first scene preview using the at least one camera sensor;

determining, by the electronic device, whether feature information changes between the first scene preview and the at least one second scene preview;

based on a result of determining whether the feature information changes, performing, by the electronic device, one of:

selecting the first scene preview as the at least one second scene preview with the wider view based on determining that the feature information does not change between the first scene preview and the at least one second scene preview, or

replacing the at least one second scene preview with the first scene preview and obtaining the at least one second scene preview by using at least one increased zoomed out feature of the at least one camera sensor based on determining that the feature information changes between the first scene preview and the at least one second scene preview;

obtaining, by the electronic device based on the at least one second scene preview, a traversal path having a least number of frames and a least change in direction of movement among a plurality of traversal paths;

capturing, by the electronic device, a plurality of frames using the traversal path to generate the panoramic image; and

storing, by the electronic device, the panoramic image.

2. The method of claim 1 , wherein the obtaining, by the electronic device, of the at least one second scene preview comprises:

detecting, by the electronic device, that a user of the electronic device opens a camera application in a panorama mode;

obtaining, by the electronic device, the at least one second scene preview by using the at least one increased zoomed out feature of the at least one camera sensor; and

determining, by the electronic device, whether at least one machine learning (ML) model detected that the feature information changes between the first scene preview and the at least one second scene preview.

3. The method of claim 1 , wherein the obtaining, by the electronic device based on the at least one second scene preview, of the traversal path comprises:

determining, by the electronic device, an orientation of the electronic device using at least one sensor of the electronic device, wherein the orientation comprises one of a landscape mode or a portrait mode;

determining, by the electronic device, a scanning direction on the at least one second scene preview with the orientation using at least one machine learning (ML) model to capture the plurality of frames to generate the panoramic image, wherein the scanning direction comprises the plurality of traversal paths;

determining, by the electronic device, a total number of frames and change in direction of movement for each traversal path;

determining, by the electronic device, at least one frame of the plurality of frames in the at least one second scene preview that is similar to at least one second frame of the plurality of frames by detecting at least one feature of at least one object displayed in the at least one frame being similar to at least one feature of at least one object displayed in the at least one second frame for each traversal path, wherein the at least one feature is detected by the at least one ML model;

generating, by the electronic device, a mapping of similar frames for each traversal path;

removing, by the electronic device, the similar frames based on the mapping;

re-determining, by the electronic device, the total number of frames and the change in direction of movement for each traversal path to recommend the traversal path;

correlating, by the electronic device, the orientation, the scanning direction, and the similar frames; and

recommending, by the electronic device, the traversal path to capture the plurality of frames to generate the panoramic image based on the correlation.

4. The method of claim 1 , wherein the plurality of traversal paths comprises at least one of a single row path, a single column path, a multi-row path, and a multi-column path.

5. The method of claim 3 , wherein the traversal path is recommended to a user of the electronic device by at least one of:

displaying, by the electronic device, a text recommendation on a screen of the electronic device;

displaying, by the electronic device, an image recommendation on the screen of the electronic device; and

playing, by the electronic device, an audio clip for recommendation in the electronic device.

6. The method of claim 1 , wherein the plurality of frames is stitched together to generate the panoramic image using at least one machine learning (ML) model.

7. The method of claim 1 , further comprising:

generating, by the electronic device, the panoramic image,

wherein the panoramic image is generated by:

receiving, by the electronic device, a stitched image for each traversal path from a stitching engine;

receiving, by the electronic device, a mapping of similar frames for each traversal path from a feature analyzer; and

regenerating, by the electronic device, the panoramic image using the received mapping of similar frames for each traversal path, the stitched image, and the at least one second scene preview from memory of the electronic device.

8. An electronic device for generating a panoramic image, the electronic device comprising:

memory;

at least one processor;

at least one camera sensor; and

panorama processing circuitry operably connected to the memory, the at least one camera sensor, and the at least one processor,

wherein the panorama processing circuitry is configured to:

obtain a first scene preview by using a zoomed out feature of the at least one camera sensor,

obtain at least one second scene preview with a wider view than the first scene preview using the at least one camera sensor,

determine whether feature information changes between the first scene preview and the at least one second scene preview,

based on a result of determining whether the feature information changes, control to perform one of:

selecting the first scene preview as the at least one second scene preview with the wider view based on determining that the feature information does not change between the first scene preview and the at least one second scene preview; or

replacing the at least one second scene preview with the first scene preview and obtaining the at least one second scene preview by using at least one increased zoomed out feature of the at least one camera sensor based on determining that the feature information changes between the first scene preview and the at least one second scene preview,

obtain, based on the at least one second scene preview, a traversal path having a least number of frames and a least change in direction of movement among a plurality of traversal paths,

capture a plurality of frames using the traversal path to generate the panoramic image, and

store the panoramic image in the memory.

9. The electronic device of claim 8 , wherein, to obtain the at least one second scene preview, the panorama processing circuitry is further configured to:

detect that a user of the electronic device opens a camera application in a panorama mode,

obtain the at least one second scene preview by using the at least one increased zoomed out feature of the at least one camera sensor, and

determine whether at least one machine learning (ML) model detected that the feature information changes between the first scene preview and the at least one second scene preview.

10. The electronic device of claim 8 , wherein, to obtain, based on the at least one second scene preview, the traversal path, the panorama processing circuitry is further configured to:

determine an orientation of the electronic device using at least one sensor of the electronic device, wherein the orientation comprises one of a landscape mode or a portrait mode,

determine a scanning direction on the at least one second scene preview with the orientation using at least one machine learning (ML) model to capture the plurality of frames to generate the panoramic image, wherein the scanning direction comprises the plurality of traversal paths,

determine a total number of frames and change in direction of movement for each traversal path,

determine at least one frame of the plurality of frames in the at least one second scene preview that is similar to at least one second frame of the plurality of frames by detecting at least one feature of at least one object displayed in the at least one frame being similar to at least one feature of at least one object displayed in the at least one second frame for each traversal path, wherein the at least one feature is detected by the at least one ML model,

generate a mapping of similar frames for each traversal path,

remove the similar frames based on the mapping,

re-determine the total number of frames and the change in direction of movement for each traversal path to recommend the traversal path,

correlate the orientation, the scanning direction, and the similar frames, and

recommend the traversal path to capture the plurality of frames to generate the panoramic image based on the correlation.

11. The electronic device of claim 8 , wherein the plurality of traversal paths comprises at least one of a single row path, a single column path, a multi-row path, and a multi-column path.

12. The electronic device of claim 10 , wherein the traversal path is recommended to a user of the electronic device by at least one of:

displaying a text recommendation on a screen of the electronic device;

displaying an image recommendation on the screen of the electronic device; and

playing an audio clip for recommendation in the electronic device.

13. The electronic device of claim 8 , wherein the plurality of frames is stitched together to generate the panoramic image using at least one machine learning (ML) model.

14. The electronic device of claim 8 , wherein, to generate the panoramic image, the panorama processing circuitry is further configured to:

receive a stitched image for each traversal path from a stitching engine;

receive a mapping of similar frames for each traversal path from a feature analyzer; and

regenerate the panoramic image using the mapping of similar frames for each traversal path, the stitched image and the at least one second scene preview from memory of the electronic device.

15. The electronic device of claim 8 , wherein a number of frames to capture the at least one second scene preview is 25 frames out of 36 frames when the electronic device is operating in a portrait mode and traversing in a row-wise manner.

16. The electronic device of claim 8 , wherein a number of frames to capture the at least one second scene preview is 21 frames out of 24 frames when the electronic device is operating in a landscape mode and traversing in a row-wise manner.

17. The electronic device of claim 8 , wherein the panoramic image is generated based on a correlation of the least number of frames and the least change in direction of movement among the plurality of traversal paths.

Assignments (1)

Priority Claims (1)

Continuity (2)

References Cited (34)