CN106780575A - Non-rigid image registration method based on characteristics of image and Demons - Google Patents
Non-rigid image registration method based on characteristics of image and Demons Download PDFInfo
- Publication number
- CN106780575A CN106780575A CN201611033914.3A CN201611033914A CN106780575A CN 106780575 A CN106780575 A CN 106780575A CN 201611033914 A CN201611033914 A CN 201611033914A CN 106780575 A CN106780575 A CN 106780575A
- Authority
- CN
- China
- Prior art keywords
- image
- registration
- demons
- methods
- rigid
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20016—Hierarchical, coarse-to-fine, multiscale or multiresolution image processing; Pyramid transform
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
Landscapes
- Image Analysis (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The invention discloses a kind of non-rigid image registration method based on characteristics of image and Demons, belong to image processing field, the problem of registration accuracy and registering time requirement cannot simultaneously be met for traditional non-rigid image registration method, the present invention proposes that two kinds (SIFT Demons and SURF Demons) from coarse to fine, non-rigid image registration method by feature level to Pixel-level on the basis of classical Demons methods.Feature point extraction is carried out to floating image and reference picture using SIFT methods or SURF methods first, rough registration (feature level registration) is carried out using the characteristic point extracted；Then, on the basis of rough registration, fine registration (pixel level registration) is carried out using the Demons methods based on optical flow field.Show by substantial amounts of registration experiment, the present invention can realize accurate, quick, stabilization the registration of medical image, overcome the defect that traditional non-rigid registration method is present.
Description
Technical field
The invention belongs to image processing field, and in particular to a kind of non-rigid image based on characteristics of image and Demons is matched somebody with somebody
Quasi- method.
Background technology
Image registration techniques are by different acquisition times (Time), different sensors (Sensor), different acquisition conditions
(Condition) mistake that Same Scene (Scene) or two width or multiple image of same target (Object) is matched
Journey, is widely used in the aspects such as Medical Image Processing, remote sensing image processing., it is necessary to registration in the registration of medical image
Larger nonhomogeneous deformation would generally occur between two images, therefore non-rigid image registration technology is that Medical Image Processing is ground
The hot issue studied carefully.The method of non-rigid image registration is broadly divided into method for registering based on characteristics of image and based on figure at this stage
As the method for registering of gray scale.Method for registering speed based on characteristics of image, but there is office for the unconspicuous image of feature
It is sex-limited；Method for registering based on gradation of image can carry out high-precision registration to the image of small deformation, but the registering time is more long.
Registration based on characteristics of image, be with extract reflection image important information feature be foundation, find registration parameter,
Make similarity measure maximum.The classical way based on characteristics of image has at this stage：SIFT(Scale-invariant Feature
Transform) method and SURF (Speeded-Up Robust Features) method.
Method for registering based on gradation of image, pixel or voxel constitutive characteristic space first with image, then basis
The statistical information of gray value calculates similarity measure, tries to achieve registration parameter, realizes image registration.Based on optical flow field in gray scale method
Modelling is the focus of Recent study, mainly using optical flow estimation, by the meter to each pixel instantaneous velocity vector information
Calculate, estimate deformation field, realize the registration of image.1998, Thirion proposed famous Demons side according to optical flow estimation
Method, Wang He proposed Active Demons (AD) method in 2005, and Demons methods are solved to a certain extent to big
The unworthiness and height of deformation are time consuming nature, introduce Demons (MD) algorithm of multiresolution strategy.Other scholars are to Demons
Method has carried out different improvement, adapts it to the registration of different type image.2007, Tomas Pock et al. were in traditional light
On the basis of flow field model, using total variation regular terms (Total Variation regularization, TV regular terms) and
Shandong nation property data item (L1 norms) builds energy function, it is proposed that a kind of new TV-L1Method, registration effect is more preferable.
The content of the invention
The present invention cannot simultaneously meet registration accuracy with registering time requirement for traditional non-rigid image registration method
A kind of problem, there is provided non-rigid image registration method based on characteristics of image and Demons.
The present invention is achieved by the following technical solutions, and the present invention is comprised the following steps：
The first step, feature point extraction is carried out using SIFT methods or SURF methods to floating image and reference picture, is utilized
The characteristic point of extraction carries out rough registration (feature level registration).
Second step, on the basis of rough registration, fine registration (Pixel-level is carried out using the Demons methods based on optical flow field
Registration).
In the first step, SIFT methods or SURF methods use difference of Gaussian function (Difference of
Gaussians, DoG) multiple dimensioned feature extraction is carried out to image.
In the second step Demons methods on multiple yardsticks image is carried out using multi-resolution pyramid by slightly to
Thin registration.
The first step is comprised the following steps that：
Step 1-1：Reference picture S and floating image M is read in, and reference picture S and floating image M are pre-processed；
Step 1-2：Metric space is built, the parameter of difference of Gaussian is set；It is empty yardstick to be built using Hessian matrixes
Between；
Step 1-3：Extreme point is screened, and generates SIFT or SURF description of these extreme points；
Step 1-4：Description tried to achieve using step 3 carries out Feature Points Matching, generates deformation field, and floating image M is entered
Row rough registration, tries to achieve rough registration image M '.
The second step is comprised the following steps that：
Trying to achieve rough registration image M ' and reference picture S using step 1-4 carries out Demons iteration essence registration：
The present invention be it is from coarse to fine, by feature level to Pixel-level non-rigid image registration method (SIFT-Demons and
SURF-Demons).In order to simultaneously is met medical image is carried out high accuracy, the requirement of high speed, present invention employs by slightly to
Carefully, by the registration strategies of feature level to Pixel-level.First using SIFT methods or SURF methods to floating image and reference picture
Feature point extraction is carried out, rough registration (feature level registration) is carried out using the characteristic point extracted；Then, on the basis of rough registration,
Fine registration (pixel level registration) is carried out using the Demons methods based on optical flow field.In order to further optimize registration result, adopt
With multiresolution strategy.First step SIFT methods or SURF methods use difference of Gaussian function (Difference of
Gaussians, DoG) multiple dimensioned feature extraction is carried out to image；Second step Demons methods are existed using multi-resolution pyramid
Registration from coarse to fine is carried out to image on multiple yardsticks.The present invention can realize that the accurate, quick, stabilization of medical image is matched somebody with somebody
Standard, the defect for overcoming traditional non-rigid registration method to exist.
Brief description of the drawings
Fig. 1, Fig. 2, Fig. 3, Fig. 4 are brain MR image registration result and difference.
Fig. 5, Fig. 6, Fig. 7, Fig. 8 are liver MR image registration results and difference.
Specific embodiment
Below in conjunction with the accompanying drawings and specific embodiment the present invention is further detailed explanation.
The present invention realizes that the registration process of medical image sequentially passes through following steps：
Step 1：Reference picture S and floating image M is read in, and S and M are pre-processed；
Step 2：Metric space is built, the parameter of difference of Gaussian is set；Metric space is built using Hessian matrixes；
Step 3：Extreme point is screened, and generates SIFT or SURF description of these extreme points；
Step 4：Description tried to achieve using step 3 carries out Feature Points Matching, generates deformation field, and floating image M is carried out
Rough registration；
Step 5：Trying to achieve rough registration image M ' and reference picture S using step 4 carries out Demons iteration essence registration：
Step 6：Objective evaluation index, the quality of analysis method are calculated using reference picture S and registering image I.
Multiresolution strategy is introduced in the step 2 to step 6 and realizes registration, so can further improve registration
Precision and convergence rate.
Interpretation
1. the accuracy of image registration
The present invention can be relatively accurate realization to the registration of medical image.Fig. 1 to Fig. 8 is to two different medical images
Registering experimental result, Fig. 1-Fig. 4 be brain MR image registration result and difference, Fig. 5-Fig. 8 be liver's MR image registration results
And difference.The present invention has preferable registration effect, Tables 1 and 2 difference to large deformation medical image and small deformation medical image
It is the present invention in Y-PSNR (PSNR, Peak Signal to Noise Ratio), mean square deviation (MSE, Mean Square
Error), coefficient correlation (CC, Correlation Coefficient), mutual information (MI, Mutual Information) and knot
Commented with several traditional the objective of non-rigid registration method in 5 indexs of structure similarity (SSIM, Structural SIMilarity)
Valency compares.
The brain MR image of table 1 registration objective evaluation
The liver CT image registration objective evaluations of table 2
2. time of image registration
The image registration time directly affects practical application effect of the invention.The present invention is based on gray scale relative to tradition
Method for registering, the registering time greatly shortens.Table 3 compares for the time with several traditional non-rigid registration methods of the invention.
The registering time of table 3 compares
In sum, the present invention devise two kinds of non-rigid registration methods for being applied to medical image (SIFT-Demons and
SURF-Demons).The method is based on characteristics of image and optical flow field, can well realize the registration to medical image.
Finally it should be noted that above embodiment is merely illustrative of the technical solution of the present invention and unrestricted, although ginseng
The present invention has been described in detail according to preferably embodiment, but protection scope of the present invention is not limited thereto, it is any
Those familiar with the art the invention discloses technical scope in, the modification that can be readily occurred in or equivalent,
Without deviating from the spirit and scope of technical solution of the present invention, should all be included within the scope of the present invention.
Claims (3)
1. a kind of non-rigid image registration method based on characteristics of image and Demons, it is characterised in that comprise the following steps：
The first step, carries out feature point extraction, using extraction using SIFT methods or SURF methods to floating image and reference picture
Characteristic point carry out rough registration；
Second step, on the basis of rough registration, fine registration is carried out using the Demons methods based on optical flow field；
In the first step, SIFT methods or SURF methods carry out multiple dimensioned feature and carry using difference of Gaussian function pair image
Take.
Demons methods are carried out on multiple yardsticks from coarse to fine using multi-resolution pyramid to image in the second step
Registration.
2. the non-rigid image registration method based on characteristics of image and Demons according to claim 1, it is characterised in that
The first step is comprised the following steps that：
Step 1-1：Reference picture S and floating image M is read in, and reference picture S and floating image M are pre-processed；
Step 1-2：Metric space is built, the parameter of difference of Gaussian is set；Metric space is built using Hessian matrixes；
Step 1-3：Extreme point is screened, and generates SIFT or SURF description of these extreme points；
Step 1-4：Description tried to achieve using step 3 carries out Feature Points Matching, generates deformation field, and floating image M is carried out slightly
Registration, tries to achieve rough registration image M '.
3. the non-rigid image registration method based on characteristics of image and Demons according to claim 2, it is characterised in that
The second step is comprised the following steps that：
Trying to achieve rough registration image M ' and reference picture S using step 1-4 carries out Demons iteration essence registration：
For i=1:1:N, N are default maximum iteration
I iteration is carried out using Demons algorithms, deformation field is obtained, line translation is entered to M；
Calculate similarity measure：If meeting condition, break；If being unsatisfactory for continuing iteration；
End。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611033914.3A CN106780575A (en) | 2016-11-23 | 2016-11-23 | Non-rigid image registration method based on characteristics of image and Demons |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611033914.3A CN106780575A (en) | 2016-11-23 | 2016-11-23 | Non-rigid image registration method based on characteristics of image and Demons |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106780575A true CN106780575A (en) | 2017-05-31 |
Family
ID=58970892
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611033914.3A Withdrawn CN106780575A (en) | 2016-11-23 | 2016-11-23 | Non-rigid image registration method based on characteristics of image and Demons |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106780575A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108022261A (en) * | 2017-11-01 | 2018-05-11 | 天津大学 | A kind of improved optical flow field model algorithm |
| CN111489381A (en) * | 2019-01-28 | 2020-08-04 | 临沂大学 | Infant brain nuclear magnetic resonance image group registration method and device |
| CN113674402A (en) * | 2021-08-23 | 2021-11-19 | 浙江大学 | Plant three-dimensional hyperspectral point cloud model generation method, correction method and device |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102169578A (en) * | 2011-03-16 | 2011-08-31 | 内蒙古科技大学 | Non-rigid medical image registration method based on finite element model |
| CN104933716A (en) * | 2015-06-16 | 2015-09-23 | 山东大学（威海） | Non-rigid registration method applied to medical image |
-
2016
- 2016-11-23 CN CN201611033914.3A patent/CN106780575A/en not_active Withdrawn
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102169578A (en) * | 2011-03-16 | 2011-08-31 | 内蒙古科技大学 | Non-rigid medical image registration method based on finite element model |
| CN104933716A (en) * | 2015-06-16 | 2015-09-23 | 山东大学（威海） | Non-rigid registration method applied to medical image |
Non-Patent Citations (1)
| Title |
|---|
| 赵永洁: "《基于光流场模型的非刚性医学图像配准算法研究》", 《中国优秀硕士论文全文数据库》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108022261A (en) * | 2017-11-01 | 2018-05-11 | 天津大学 | A kind of improved optical flow field model algorithm |
| CN111489381A (en) * | 2019-01-28 | 2020-08-04 | 临沂大学 | Infant brain nuclear magnetic resonance image group registration method and device |
| CN111489381B (en) * | 2019-01-28 | 2023-11-10 | 临沂大学 | Infant brain nuclear magnetic resonance image group registration method and device |
| CN113674402A (en) * | 2021-08-23 | 2021-11-19 | 浙江大学 | Plant three-dimensional hyperspectral point cloud model generation method, correction method and device |
| CN113674402B (en) * | 2021-08-23 | 2023-10-31 | 浙江大学 | Plant three-dimensional hyperspectral point cloud model generation method, correction method and device thereof |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106558073A (en) | Based on characteristics of image and TV L1Non-rigid image registration method | |
| Fan et al. | Registration of optical and SAR satellite images by exploring the spatial relationship of the improved SIFT | |
| CN103839265B (en) | SAR image registration method based on SIFT and normalized mutual information | |
| CN110110646B (en) | Gesture image key frame extraction method based on deep learning | |
| CN106991695A (en) | A kind of method for registering images and device | |
| CN103886611A (en) | Image matching method suitable for automatically detecting flight quality of aerial photography | |
| CN109523585A (en) | A kind of multi-source Remote Sensing Images feature matching method based on direction phase equalization | |
| CN104123554B (en) | SIFT image characteristic extracting methods based on MMTD | |
| CN110245593A (en) | A kind of images of gestures extraction method of key frame based on image similarity | |
| Chen et al. | Robust affine-invariant line matching for high resolution remote sensing images | |
| CN110598613B (en) | Expressway agglomerate fog monitoring method | |
| CN104240264A (en) | Height detection method and device for moving object | |
| CN106780575A (en) | Non-rigid image registration method based on characteristics of image and Demons | |
| CN112308873B (en) | Edge detection method for multi-scale Gabor wavelet PCA fusion image | |
| CN102446356A (en) | Parallel and adaptive matching method for acquiring remote sensing images with homogeneously-distributed matched points | |
| Liu et al. | Farmland aerial images fast-stitching method and application based on improved sift algorithm | |
| CN111199245A (en) | Rape pest identification method | |
| CN111062972B (en) | Image tracking method based on image frequency domain conversion | |
| Paturkar et al. | 3D reconstruction of plants under outdoor conditions using image-based computer vision | |
| CN112017221B (en) | Multi-modal image registration method, device and equipment based on scale space | |
| CN104700359A (en) | Super-resolution reconstruction method of image sequence in different polar axis directions of image plane | |
| CN117078726A (en) | Different spectrum image registration method based on edge extraction | |
| CN105783785A (en) | Wavelet-ridge phase extraction method | |
| CN113792788B (en) | Infrared and visible light image matching method based on multi-feature similarity fusion | |
| Priyadharshini et al. | Automatic registration of images with simulated rotation and translation estimation using HAIRIS |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20170531 |