Image-processing device

Patent metadata

Field	Value
Publication number	US-2017091908-A1
Application number	US-201515312408-A
Country	US
Kind code	A1
Filing date	May 18, 2015
Priority date	May 19, 2014
Publication date	Mar 30, 2017
Grant date	—

Abstract

According to the aspect of the present invention, an image processing device capable of absolutely removing noises in each frame forming a live image can be provided. Specifically, according to the aspect of the present invention, a target block T can be set in the original image I t and it is searched out where such target block T is imaged in the superimposition target image O t−1 . According to the aspect of the present invention, a tracking of the subject image is executed per block, so that reliability thereof can be extraordinarily improved compared to the conventional method searching the destination every individual pixel.

First claim

What is claimed is:

1 . An image processing device capable of processing a noise reduction relative to an image generated by continuously imaging a subject, comprising:
a target setting means that sets a target block formed with an attention pixel and periphery pixels of said attention pixel among pixels forming an original image imaging the subject; a search means that searches out a destination block most resembling said target block among reference images imaging the subject imaged at a different time from said original image; a superimposing means that generates a fusion block by superimposing a superimposition target block being in the same location of said destination block in a superimposition target image having the subject image in the same location as the reference image to said target block in the original image; an image generation means that generates a noise reduction image having reduced noises being imaged in said original image by superimposing a plurality of respective said fusion blocks in the image one after the other along with setting a plurality of respective said target blocks one after the other while changing the location of the attention pixel; and wherein said image generation means is operative to allow the location of the fusion block in the noise reduction image to be the same as the location of the target block in the original image.

2 . The image processing device according to claim 1 ; wherein:
said image generation means generates the noise reduction image by dividing the pixel value of the pixel in the image by the integrating number indicating how many times said fusion blocks are added following an adding while overlapping said fusion blocks.

3 . The image processing device according to claim 1 , wherein:
said superimposing means superimposes pixels of said target block in said original image and corresponding pixels of said superimposition target block in said superimposition target image by weighting individually every pixel forming said fusion block, and weighting of superimposition is changed so that said superimposition target block cannot be gradually succeeded by said fusion block in accordance with increase of the absolute value of the difference between the pixel value of the pixel belonging to said target block and the pixel value of the corresponding pixels of the superimposition target block or the destination block.

4 . The image processing device according to claim 1 , wherein:
said superimposing means superimposes pixels of said target block in said original image and said superimposition target block in said superimposition target image by weighting individually every pixel forming said fusion block, and weighting of superimposition is changed so that said superimposition target block cannot be gradually succeeded by said fusion block in accordance with increase of the absolute value of the difference between the pixel value in said target block and the pixel value of the corresponding pixels of the superimposition target block or the destination block.

5 . The image processing device according to claim 1 , comprising:
an editing means that sets the same location block in the location of said target block in said reference image and overwrites the output from said search means so that the destination block searched out by said search means becomes said same location block, if said destination block does not significantly resemble said target block compared to said same location block.

6 . The image processing device according to claim 1 , wherein:
a superimposition target image a 1 is a noise reduction image acquired when the image imaged prior to said original image is image-processed and a reference image b 1 is the image imaged prior to said original image.

7 . The image processing device according to claim 1 , wherein:
a superimposition target image a 1 is a noise reduction image acquired when the image imaged prior to said original image is image-processed, and a reference image b 2 is the noise reduction image corresponding to the image imaged prior to said original image.

8 . The image processing device according to claim 1 , wherein:
a superimposition target image a 2 is an image imaged prior to said original image and a reference image b 1 is an image imaged prior to said original image.

9 . The image processing device according to claim 1 , wherein:
a superimposition target image a 2 is the image imaged prior to said original image and a reference image b 2 is a noise reduction image corresponding to the image imaged prior to said original image.

10 . The image processing device according to claim 1 , wherein:
said search means is operative based on two modes including an accuracy priority mode that searches a destination block relative to some attention pixel over a large area of said reference image, and a velocity priority mode that searches the destination block relative to the attention pixel different from the attention pixel that is a target to be processed by said accuracy priority mode based on the search result of said accuracy priority mode in a narrow area of said reference image; and wherein said velocity priority mode searches the destination block over the area enclosing the predicted location that is the predicted destination location of the attention pixel currently being a search target in said reference image is predicted, based on the locational relationship between the attention pixel that is the target to be processed by said accuracy priority mode and the pixels of the destination in said reference image of said attention pixel found by the search due to said accuracy priority mode.

11 . The image processing device according to claim 1 , wherein:
an area of said superimposition target block relative to said superimposing means is narrower than an area of said target block and said destination block.

12 . The image processing device according to claim 1 , wherein:
an area of said superimposition target block relative to said superimposing means is broader than an area of said target block and said destination block.

13 . The image processing device according to claim 1 , further comprising:
an image reducing means that generates a reduced original image and a reduced reference image by reducing said original image and said reference image; a reduced image target block setting means that sets said attention pixel from pixels forming the original image and a reduced original image target block that is a target block in the reduced original image; a reduced image search means that searches out a reduced reference image destination block most resembling said reduced original image target block from said reduced reference images; and a search area setting means that sets up a search area, wherein said search means searches out said destination block in said the reference image on the basis of the location in said reference image corresponding to a location of said reduced reference image destination block in said search area.

14 . The image processing device according to claim 1 , wherein:
said search means determines a resemblance also relative to a plurality of respective rotated candidate blocks when said destination block is searched out from candidate blocks for said destination block in said reference image.

15 . The image processing device according to claim 1 , wherein:
the target block relative to said target block setting means includes an enclave.

16 . The image processing device according to claim 1 , wherein:
the target block relativ

Assignees

• Shimadzu Corp

Inventors

• Takeda Ryo
• Morita Hisanori

Classifications

• G06T2207/20182

  Noise reduction or smoothing in the temporal domain; Spatio-temporal filtering · CPC title

• G06T2207/10121

  Fluoroscopy · CPC title

• G06T5/50

  using two or more images, e.g. averaging or subtraction · CPC title

• G06T2207/10004

  Still image; Photographic image · CPC title

• G06T7/337

  involving reference images or patches · CPC title