Back to results
US 12,755,361 B2Granted
Image processing apparatus and image processing method
AssigneeSony Group Corporation
InventorRyota Miyazawa
Life of this patent
Estimating the term…
Abstract
An image processing apparatus is provided with a moire detection unit that detects a pixel region with a different motion out of a pixel region in which the same motion as a motion of a subject is assumed as a change in an in-frame position between images at different times, and generates detection information of moire.
Claims12
An image processing apparatus, comprising:
- a central processing unit (CPU) configured to:
- acquire raw image data, wherein the raw image data comprises current image data and past image data;
- perform an object recognition process on the current image data and the past image data;
- set, based on the performed object recognition process, a target subject from a plurality of subjects in the current image data;
- detect a motion of the set target subject of the plurality of subjects, wherein
- an in-frame position of the target subject in the current image data is different from an in-frame position of the target subject in the past image data, and
- the current image data and the past image data are captured at different time intervals;
- detect a specific motion of a first pixel region of a plurality of pixel regions of the target subject, wherein
- the detected specific motion of the first pixel region is different from the detected specific motion of the target subject;
- determine the first pixel region, within the target subject, with the detected specific motion as a moire portion, wherein the moire portion has moire;
- generate detection information of the moire, wherein
- the generated detection information includes moire presence or absence information and area information indicating the first pixel region of the target subject having the moire;
- and
- perform, based on the generated detection information, a moire reduction process by a low-pass filter (LPF) process on the current image data.
11 more elements
An image processing method, comprising:
- acquiring raw image data, wherein the raw image data comprises current image data and past image data;
- performing an object recognition process on the current image data and the past image data;
- setting, based on the performed object recognition process, a target subject from a plurality of subjects in the current image data;
- detecting a motion of the set target subject of the plurality of subjects, wherein
- an in-frame position of the target subject in the current image data is different from an in-frame position of the target subject in the past image data, and
- the current image data and the past image data are captured at different time intervals;
- detecting a specific motion of a pixel region of a plurality of pixel regions of the target subject, wherein
- the detected specific motion of the pixel region is different from the detected specific motion of the target subject;
- determining the pixel region, within the target subject, with the detected specific motion as a moire portion, wherein the moire portion has moire;
- generating, based on the detection of the pixel region, detection information of the moire, wherein
- the generated detection information includes moire presence or absence information and area information indicating the pixel region of the target subject having the moire;
- and
- performing, based on the generated detection information, a moire reduction process by a low-pass filter (LPF) process on the current image data.
10 more elements
An image processing apparatus, comprising:
- a central processing unit (CPU) configured to:
- perform an object recognition process on current image data and past image data, wherein
- the object recognition process is based on a semantic segmentation process and a pattern recognition process;
- set, based on the performed object recognition process, a target subject from a plurality of subjects in the current image data;
- compare an in-frame position of the target subject in the current image data and an in-frame position of the target subject in the past image data, wherein
- the in-frame position of the target subject in the current image data is different from the in-frame position of the target subject in the past image data, and
- the current image data and the past image data are captured at different time intervals;
- detect, based on the comparison, a specific motion of a pixel region of a plurality of pixel regions of the target subject, wherein
- the detected specific motion of the pixel region is different from a motion of the target subject;
- determine the pixel region, within the target subject, with the detected specific motion as a moire portion, wherein the moire portion has moire;
- generate detection information of the moire;
- and
- perform, based on the generated detection information, a moire reduction process by a low-pass filter (LPF) process on the current image data.
10 more elements
Dependencies are inferred from the claim text.