IMAGE PROCESSING APPARATUS AND METHOD

• US 20130004085A1
• Filed: 06/29/2012
• Published: 01/03/2013
• Est. Priority Date: 07/01/2011
• Status: Active Grant

• Current Assignee: Toshiba Medical Systems Corporation (Canon Inc.)
• Original Assignee: Toshiba Medical Systems Corporation (Canon Inc.), Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: PIPER, Jim, PLAKAS, Costas, BAI, Lu, WAN, Jun, DAI, Jing, WANG, Shaobin
• Granted Patent: US 8,792,729 B2
• Time in Patent Office: Days
• US Class Current: 382/209
• CPC Class Codes: G06T 3/14 Transformations for image registration, e.g. adjusting or mapping for alignment of images

Abstract

An image processing apparatus may include: a first registration device for performing, by taking a first input image of two overlapped input images having an overlapped as a reference image, a first registration on a second input image, to find, in the second input image, a second pixel which is matched with each first pixel located in the overlapped area of the reference image; an output pixel location determination device for calculating, a location of an output pixel which is located in the overlapped area of the output image and corresponds to the first pixel, wherein the locations of the first and second pixels are respectively weighted, and the shorter the distance from the first pixel to a non-overlapped area of the reference image is, the greater a weight of the location of the first pixel is; and an output pixel value determination device for calculating a pixel value.

Claims

1. An image processing apparatus, capable of stitching two overlapped input images having an overlapped area into an output image, comprising:

a first registration device configured to perform, by taking a first input image of the two overlapped input images as a reference image, a first registration on a second input image of the two overlapped input images, to find, in the second input image, a second pixel which is matched with each first pixel located in the overlapped area of the reference image;

an output pixel location determination device configured to calculate, based on the locations of the first and second pixels, a location of an output pixel which is located in the overlapped area of the output image and corresponds to the first pixel, wherein the locations of the first and second pixels are respectively weighted, and the shorter the distance from the first pixel to a non-overlapped area of the reference image is, the greater a weight of the location of the first pixel is, and the smaller a weight of the location of the second pixel is; and

an output pixel value determination device configured to calculate a pixel value of the output pixel according to a pixel value of at least one of the first and second pixels.

2. The image processing apparatus according to claim 1, wherein the weight of the location of the first pixel is in a nonlinear function relation with the distance from the first pixel to the non-overlapped area of the reference image.

3. The image processing apparatus according to claim 2, wherein the shorter the distance from the first pixel to the non-overlapped area of the reference image or the non-overlapped area of the second input image is, the smaller a variation gradient of the weights is.

4. The image processing apparatus according to claim 3, wherein the output pixel location determination device is further configured to calculate the weight b({right arrow over (X)}) of the location of the second pixel according to the following formula or a mathematical transformation based on the following formula and take (1−

b({right arrow over (X)})) as the weight of the location of the first pixel;

5. The image processing apparatus according to claim 1, wherein the output pixel value determination device is configured to calculate a weighted sum of the pixel values of the first and second pixels, as the pixel value of the output pixel, wherein the shorter the distance from the output pixel to the non-overlapped area of the reference image is, the greater the weight of the pixel value of the first pixel is, and the shorter the distance from the output pixel to the non-overlapped area of the second input image is, the greater the weight of the pixel value of the second pixel is.

6. The image processing apparatus according to claim 5, wherein the weights of the pixel values of the first and second pixels are in a nonlinear function relation with the distance from the output pixel to the non-overlapped area of the reference image, and the shorter the distance from the output pixel to the non-overlapped area of the reference image or the non-overlapped area of the second input image is, the smaller a variation gradient of the weights is.

7. The image processing apparatus according to claim 6, wherein the output pixel value determination device is further configured to calculate the weight b({right arrow over (Y)}) of the pixel value of the second pixel according to the following formula or a mathematical transformation based on the following formula and take (1−

b({right arrow over (Y)})) as the weight of the pixel value of the first pixel;

8. The image processing apparatus according to claim 1, wherein the first registration device is configured to perform the first registration by using a non-rigid registration method.

9. The image processing apparatus according to claim 1, further comprising:

a rigid registration device configured to perform a rigid registration on the second input image by taking the reference image as a reference before the first registration is performed by the first registration device and outputting the second input image subjected to the rigid registration to the first registration device,

wherein the first registration device performs the first registration on the second input image subjected to the rigid registration by taking the reference image as a reference.

10. An image processing method, for stitching two overlapped input images having an overlapped area into an output image, comprising:

performing, by taking a first input image of the two overlapped input images as a reference image, a first registration on a second input image of the two overlapped input images, to find, in the second input image, a second pixel which is matched with each first pixel located in the overlapped area of the reference image;

calculating, based on the locations of the first and second pixels, a location of an output pixel which is located in the overlapped area of the output image and corresponds to the first pixel, wherein the locations of the first and second pixels are respectively weighted, and the shorter the distance from the first pixel to a non-overlapped area of the reference image is, the greater a weight of the location of the first pixel is, and the smaller a weight of the location of the second pixel is; and

calculating a pixel value of the output pixel according to a pixel value of at least one of the first and second pixels.

11. The image processing method according to claim 10, wherein the weight of the location of the first pixel is in a nonlinear function relation with the distance from the first pixel to the non-overlapped area of the reference image.