Unlock instant, AI-driven research and patent intelligence for your innovation.

Image processing method, image processing apparatus, image processing system, and memory medium

An image processing, image technology, applied in image data processing, image enhancement, image analysis and other directions, can solve problems such as side effects

Pending Publication Date: 2021-10-19
CANON KK
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Side effects inherent to machine learning models can occur when estimating images other than deblurring

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Image processing method, image processing apparatus, image processing system, and memory medium
  • Image processing method, image processing apparatus, image processing system, and memory medium
  • Image processing method, image processing apparatus, image processing system, and memory medium

Examples

Experimental program
Comparison scheme
Effect test

no. 1 approach

[0031] First, a description will be given of the image processing system in the first embodiment of the present invention. In this embodiment, deblurring a captured image including saturated pixels is the purpose of image estimation by a machine learning model. The blur to be corrected is blur due to aberration or diffraction generated by an optical system, an optical low-pass filter, or a pixel opening of an image sensor. However, the present embodiment is not limited thereto, and in correction of blur due to defocus or shake, the effects of the present embodiment can be similarly obtained. Similarly, in image estimation other than deblurring, the effect of the present embodiment can be obtained.

[0032] figure 2 is a block diagram showing the image processing system 100 . image 3is an external view showing the image processing system 100 . The image processing system 100 includes a learning device 101 and an imaging device 102 connected to each other via a wired or wi...

no. 2 approach

[0075] Next, a description will be given of the image processing system in the second embodiment of the present invention. An object of the present embodiment is to suppress overcorrection and correct blur in a captured image even when blur differs from learned blur due to manufacturing variation or the like. However, the present embodiment can be similarly applied to other image estimations.

[0076] Figure 9 is a block diagram showing the image processing system 300 . Figure 10 is an external view showing the image processing system 300 . The image processing system 300 includes a learning device 301 , an imaging device 302 , and an image processing device 103 , and the respective devices are connected via a network or the like. The imaging apparatus 302 includes an optical system 321 and an image sensor 322 , a memory 323 , a communication unit 324 , and a display 325 . A captured image acquired by the optical system 321 and the image sensor 322 includes blur caused b...

no. 3 approach

[0093] Next, a description will be given of an image processing system in a third embodiment of the present invention. In the present embodiment, converting bokeh including defocus blur of saturated pixels is the purpose of estimating an image by a machine learning model. The conversion of bokeh refers to a process of converting the distribution of defocus blur in an out-of-focus area of ​​a captured image into a different distribution. For example, converting the two-line blur caused by the separation of PSF peaks into a circular blur with a flat distribution or into a Gaussian distribution function. Thereby, it is possible to further sharpen the focused subject, or to change the impression of the captured image. The present invention can also be applied to image estimation other than conversion of bokeh in the same manner, and effects can be obtained.

[0094] Figure 13 It is a block diagram showing the image processing system 500 in this embodiment. Figure 14 is an ex...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The present invention provides an image processing method, an image processing apparatus, an image processing system, and a memory medium. The image processing method includes the steps of acquiring the first model output generated based on a captured image and by a first machine learning model, acquiring the second model output generated based on the captured image and by a second machine learning model which is different from the first machine learning model, and generating an estimated image by using the first model output and the second model output, based on a comparison based on the second model output and one of the captured image and first model output.

Description

technical field [0001] The present invention relates to an image processing method that suppresses side effects that inherently occur when an image is estimated based on a captured image by using a machine learning model. Background technique [0002] A machine learning model is capable of estimating an image while achieving a higher effect than theoretically based estimation of the image using assumptions or approximations. In theory-based estimates of images, the effect is diminished by assumed or approximately ignored elements. On the other hand, in the machine learning model, by performing learning using learning data including these elements, it is possible to estimate an image with high effect based on learning data without making assumptions or approximations. [0003] For example, in the technique of deblurring a captured image, elements are saturated pixels of the captured image. In theory-based methods such as the Wiener filter, it is assumed that no saturated pi...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06T5/00G06N20/20G06N3/04H04N5/232
CPCG06N20/20G06T2207/20081G06T2207/20084H04N23/80G06N3/045G06T5/80G06T5/73G06T2207/10024G06N3/084G06V10/82G06V10/751G06V10/764G06T5/60G06T7/70G06N3/08
Inventor 日浅法人
Owner CANON KK
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com