Apparatus and method for raw-cost calculation using adaptive window mask

Patent metadata

Field	Value
Publication number	US-9721348-B2
Application number	US-201614992984-A
Country	US
Kind code	B2
Filing date	Jan 11, 2016
Priority date	Oct 23, 2015
Publication date	Aug 1, 2017
Grant date	Aug 1, 2017

Abstract

Disclosed is an apparatus and method for calculating a raw-cost necessary for combining images into one image by matching of stereo images. The raw-cost calculation apparatus includes an image acquirer, a window generator, a window mask generator, a window masker, and a raw-cost calculator. In the raw-cost calculation apparatus and method, a raw cost may be calculated by using an adaptive window mask so that accurate 3D information may be obtained on the boundary of thin structures even when stereo images are matched and combined.

First claim

What is claimed is:

1. An apparatus that performs raw-cost calculation using an adaptive window mask, the apparatus comprising:
an image acquirer configured to acquire image information that includes pattern and texture information of a reference image, and pattern and texture information of a target image, wherein the reference image and the target image comprise a stereo image;
a window generator configured to generate a plurality of windows, each including at least one pixel, by receiving the texture information, included in the acquired image information, and having respective shapes determined depending on types of objects of the texture information of the reference image or target image;
a window mask generator configured to generate a window mask, which has a same shape as a shape of the generated windows, and is capable of converting the pattern information, included in the acquired image information, according to a predetermined algorithm;
a window masker configured to perform masking of the reference image and the target image the by using the generated window mask to convert the pattern information of the reference image and the pattern information of the target image included in the acquired image information; and
a raw-cost calculator configured to match pattern information of the masked reference image and pattern information of the masked target image to calculate a difference, and to calculate a raw cost based on the difference,
wherein the reference image and the target image are combined, using the calculated raw cost, to produce a three-dimensional (3D) image.

2. The apparatus of claim 1 , wherein the window mask generator generates the window mask that includes a set value for masking,
wherein the set value for masking varies depending on a camera setting of the received texture image, a light amount, a state of lighting, and image resolution.

3. The apparatus of claim 2 , wherein the window mask generator generates the window mask that includes all pixels of windows in response to the set value for masking being greater than a predetermined value.

4. The apparatus of claim 1 , wherein by assuming that an absolute difference, calculated by matching windows of the generated reference image and windows of the target image, on an X axis is defined as W, an absolute difference on a Y axis is defined as H, and a maximum absolute difference between images having the X axis and Y axis is defined as D, the raw cost calculator calculates a volume of the raw cost by using the defined values W, H, and D and according to a predetermined algorithm, and calculates a final raw cost by combining the calculated volume of raw cost with adjacent pixels.

5. A method for producing three-dimensional (3D) images by performing a raw-cost calculation using an adaptive window mask, the method comprising:
acquiring image information that includes pattern and texture information of a reference image, and pattern and texture information of a target image, wherein the reference image and the target image comprise a stereo image;
generating a plurality of windows, each including at least one pixel, by receiving the texture information, included in the acquired image information, and having respective shapes determined depending on types of objects of the texture information of the reference image or target image;
generating a window mask, which has a same shape as a shape of the generated windows, and is capable of converting the pattern information, included in the acquired image information, according to a predetermined algorithm;
performing masking of the reference image and the target image by using the generated window mask to convert the pattern information of the reference image and the pattern information of the target image included in the acquired image information;
matching pattern information of the masked reference image and pattern information of the masked target image to calculate a difference, and calculating a raw cost based on the difference; and
producing the 3D image by combining, using the calculated raw cost, the reference image and the target image.

6. The method of claim 5 , wherein the generating of the window mask comprises generating the window mask that includes a set value for masking,
wherein the set value for masking varies depending on a camera setting of the received texture image, a light amount, a state of lighting, and image resolution.

7. The method of claim 6 , wherein the generating of the window mask comprises generating the window mask that includes all pixels of windows in response to the set value for masking being greater than a predetermined value.

8. The method of claim 5 , wherein by assuming that an absolute difference, calculated by matching windows of the generated reference image and windows of the target image, on an X axis is defined as W, an absolute difference on a Y axis is defined as H, and a maximum absolute difference between images having the X axis and Y axis is defined as D, the calculation of the raw cost comprises calculating a volume of the raw cost by using the defined values W, H, and D and according to a predetermined algorithm, and calculating a final raw cost by combining the calculated volume of raw cost with adjacent pixels.

Assignees

• Electronics & Telecommunications Res Inst

Inventors

• Jeong Jae Chan
• Chang Ji Ho
• Choi Seung Min

Classifications

• G06T7/593 Primary

  from stereo images · CPC title

• G06T7/40

  Analysis of texture (depth or shape recovery from texture G06T7/529) · CPC title

• G06V10/759

  Region-based matching · CPC title

• G06V10/7515

  Shifting the patterns to accommodate for positional errors · CPC title

• G06T7/003

  Physics · mapped topic