CN113706409A - CBCT image enhancement method and device based on artificial intelligence and storage medium - Google Patents
CBCT image enhancement method and device based on artificial intelligence and storage medium Download PDFInfo
- Publication number
- CN113706409A CN113706409A CN202110948544.0A CN202110948544A CN113706409A CN 113706409 A CN113706409 A CN 113706409A CN 202110948544 A CN202110948544 A CN 202110948544A CN 113706409 A CN113706409 A CN 113706409A
- Authority
- CN
- China
- Prior art keywords
- image
- cbct
- training
- cbct image
- network model
- 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.)
- Pending
Links
- 238000007408 cone-beam computed tomography Methods 0.000 title claims abstract description 111
- 238000000034 method Methods 0.000 title claims abstract description 28
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 12
- 238000003860 storage Methods 0.000 title claims abstract description 8
- 238000012549 training Methods 0.000 claims abstract description 32
- 238000010606 normalization Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 5
- 238000004590 computer program Methods 0.000 claims description 3
- 238000002360 preparation method Methods 0.000 claims description 3
- 238000013135 deep learning Methods 0.000 claims description 2
- 238000012935 Averaging Methods 0.000 claims 1
- 230000008901 benefit Effects 0.000 abstract description 3
- 230000000694 effects Effects 0.000 description 5
- 210000001519 tissue Anatomy 0.000 description 4
- 238000013136 deep learning model Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000001959 radiotherapy Methods 0.000 description 2
- 208000009119 Giant Axonal Neuropathy Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 201000003382 giant axonal neuropathy 1 Diseases 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000001727 in vivo Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 210000004872 soft tissue Anatomy 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- 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
- G06T2207/20081—Training; Learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a CBCT image enhancement method, a device and a storage medium based on artificial intelligence, wherein the CBCT image enhancement method comprises the following steps: s1: preparing paired CBCT and CT training image data; s2: training a learning network by using the prepared pair of CBCT and CT training image data to obtain a network model for converting the CBCT image into the CT image; s3: and inputting the CBCT image into the network model to obtain an enhanced CBCT image. The CBCT image enhanced by the method has the advantages of smaller noise, high contrast ratio and more accurate CT value, and can effectively improve the positioning precision and positioning speed of a clinician.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a CBCT image enhancement method and device based on artificial intelligence and a storage medium.
Background
In order to meet the requirements of physicists and technicians on clinical placement verification of radiotherapy patients, Cone Beam Computed Tomography (CBCT) technology can be adopted to acquire three-dimensional volume images of patients in a treatment room, and then three-dimensional registration is carried out on the three-dimensional volume images and planned CT images to determine the placement deviation of the patients, so that medical staff can correct the placement of the patients according to the placement deviation.
Compared with the CT for diagnosis, the CBCT bulb tube has relatively smaller current and exposure time, the detector has fewer sampling times when the machine frame rotates for one circle, and the CBCT detector is more influenced by scattering and has lower sensitivity compared with the CT detector. Therefore, the CBCT image is more noisy and has lower contrast than the CT image, which affects the positioning accuracy of the doctor.
In the prior art, the CBCT image is usually enhanced by the following method:
1. performing image noise reduction and edge enhancement by using various convolution operators in an image space domain, such as a Gaussian operator, a canny operator, a Laplacian operator and the like;
2. carrying out high-pass filtering in a frequency domain to remove low-frequency information in the image and reduce the noise of the image;
3. and processing the gray level histogram of the image to increase the contrast of the image.
However, the prior art has low robustness, effects of different images cannot be completely consistent, partial image information is inevitably lost in the process of removing noise, the image edge is blurred, and the image CT value may be changed after the image contrast is enhanced.
Disclosure of Invention
In order to solve the technical problems, the invention discloses a CBCT image enhancement method, a device and a storage medium based on artificial intelligence, the CBCT image enhanced by the method has the advantages of smaller noise, high contrast and more accurate CT value, the positioning precision and the positioning speed of a clinician can be effectively improved, the CBCT image is limited by the poorer accuracy of the CT value, the self-adaptive plan cannot be smoothly implemented, the CBCT image processed by the method is more beneficial to the manufacture of the self-adaptive plan, and the radiotherapy effect is further improved.
Specifically, the invention provides a CBCT image enhancement method based on artificial intelligence, which comprises the following steps: s1: preparing paired CBCT and CT training image data; s2: training a learning network by using the prepared pair of CBCT and CT training image data to obtain a network model for converting the CBCT image into the CT image; s3: and inputting the CBCT image into the network model to obtain an enhanced CBCT image.
In a further aspect, step S1 includes: s11: registering the CBCT image and the CT image by using a registration frame; s12: and carrying out pixel normalization and gray value normalization processing on the matched CBCT image and CT image.
In a further aspect, step S11 includes: s111: carrying out rigid registration on the CBCT image and the CT image so as to correct the global large-scale offset; s112: after rigid registration, elastic registration is further performed on the CBCT image and the CT image so as to ensure that the tissues in the body in the two groups of images correspond one to one.
In a further technical solution, in step S2, the depth learning is performed based on a pix2pix-gans model, where the pix2pix-gans model includes a generator and a discriminator, the generator generates a pseudo CT image based on an input CBCT image, and the discriminator is used for distinguishing the CBCT image from the pseudo CT image.
In a further technical scheme, the generator adopts a U-Net structure, and the arbiter adopts a PatchGAN structure.
In a further technical scheme, the U-Net structure adopts a structure that an input image is firstly down-sampled to a low dimension and then up-sampled to an original resolution, a skip-connection is added, and corresponding feature maps and feature maps with the same size after decode are spliced together according to a channel and used for reserving detail information of pixel levels under different resolutions.
In a further technical scheme, the PatchGAN structure uses GAN to construct high-frequency information, and performs true and false judgment on each patch with the size of N x N of an image, and averages the results of all the patches of one image to output as a final discriminator.
In a further technical solution, in step S12, the pixel size is unified to 1mm, the gray scale value is limited to (-1024,2048), if the pixel gray scale value is less than-1024, it is set to-1024, if the pixel gray scale value is greater than 2048, it is set to 2048, and then the gray scale value is normalized to (-1, 1).
The invention also provides a CBCT image enhancement device based on artificial intelligence, which comprises: the CBCT and CT training image data preparation module is used for preparing paired CBCT and CT training image data; the network model training module is used for training a learning network by using prepared paired CBCT and CT training image data to obtain a network model for converting the CBCT image into a CT image; and the enhanced CBCT image output module is used for inputting the CBCT image into the network model so as to obtain an enhanced CBCT image.
The invention additionally provides a computer storage medium storing a computer program which, when executed, performs the method as described above.
The algorithm in the invention can remove the noise in the image under the condition of completely keeping the original edge, and the denoising effect is better than that of the prior art.
Drawings
FIG. 1 is a schematic flow chart of a CBCT image enhancement method according to the present invention;
FIG. 2 is a flow chart of image registration of the present invention;
FIG. 3A is a CBCT and CT image prior to registration;
FIG. 3B is the result of rigid registration of CBCT and CT images;
FIG. 3C is the result of elastic registration of CBCT and CT images;
FIG. 4 is a schematic diagram of a pix2pix-gans network structure of the present invention;
fig. 5 is a schematic diagram of the u-net network architecture of the generator of the present invention.
Detailed Description
The technical solution of the present invention will be further described with reference to the following specific examples, but the present invention is not limited to these examples.
In the invention, a deep learning model is designed, the CBCT image can be converted into the CT image by using the deep learning model, after the conversion, the quality of the CBCT image is close to that of the CT image, the noise and the artifact in the image obviously disappear, and the image contrast is also enhanced.
As shown in FIG. 1, the CBCT image enhancement method based on artificial intelligence of the invention can comprise the following steps:
1. preparing paired data of CBCT and CT, acquiring CT images through a CT analog positioning machine, wherein the acquisition time of the CBCT is close to that of the CT, matching the overall contour of the CT and the CBCT by using rigid registration and then ensuring that the in vivo tissues in the two groups of images correspond one to one by using elastic registration in order to ensure that the spatial information of the CBCT and the CT images is consistent.
2. And carrying out pixel size normalization and gray value normalization on the well-matched CBCT and CT images.
3. Training a deep learning network by using the prepared CBCT and CT images to obtain a network model for converting the CBCT into the CT images;
4. and when the CBCT is reconstructed again subsequently, inputting the reconstructed CBCT image into the network model to obtain an enhanced image.
Because CBCT and CT data of a patient are acquired at different times, and therefore, certain offset and deformation exist, before a model is trained, a registration frame is required to be used for registering CBCT and CT images, and it is ensured that spatial information of the CBCT and CT images is consistent, an elastic registration frame mainly comprises two parts, namely rigid registration and elastic registration, wherein the rigid registration is mainly used for correcting global large-scale offset, and the elastic registration is mainly used for restoring local deformation such as soft tissue. The overall registration flow chart is shown in fig. 2.
The specific registration effect is shown in fig. 3A-3B, where fig. 3A is a CBCT and CT image before registration, a human body part in the image has obvious dislocation, fig. 3B is a result after rigid registration, the main body contour of the CBCT and CT is corrected, but the skin edge and the tissue in the human body have certain deformation, fig. 3C is a result after elastic registration, the local deformation of the human body has disappeared, and the CBCT and CT image after registration have one-to-one correspondence in pixels, so that they can be used for the subsequent model training.
In addition, for training the model, 200 sets of corresponding data of the CBCT and the CT may be prepared, the pixel size is unified to 1mm, the image layer thickness is also 1mm, the gray values of the CBCT and the CT are limited to (-1024,2048), the gray value of the pixel is set to-1024 when the gray value of the pixel is smaller than-1024, the gray value of the pixel is set to 2048 when the gray value of the pixel is larger than 2048, and then the gray value is normalized to (-1, 1).
The deep learning model designed by the invention is based on a pix2pix-gans model, the model consists of a generator G and a discriminator D, a CBCT picture needs to be input into the generator, and the generator generates a false CT image (fake CT). The input of the discriminator is CBCT and fake CT images, and the discriminator needs to distinguish the CBCT from the fake CT.
As shown in fig. 4, pix2pix-GANs uses a generator to solve the low frequency component (texture information) of an image and GAN (generative confrontation network) to solve the high frequency component (image edge information). The generator uses the traditional L1 loss or MSE loss to make the generated picture as similar as possible to the training picture, and GAN is used to construct the details of the high frequency part.
As shown in fig. 5, a U-Net network structure is used as a generator, the U-Net adopts an (Encoder-Decoder) structure that down-samples an input image to a low dimension, then up-samples the input image to an original resolution, and adds a skip-connection, and the corresponding feature maps and the feature maps with the same size after the decode are spliced together according to a channel (context) to retain the detail information of the pixel level at different resolutions. The effect of U-Net on detail improvement is very obvious.
On the other hand, whether or not the picture is generated is determined using PatchGAN as a discriminator. The idea of PatchGAN is that using GAN to construct high frequency information does not require the entire picture to be input into a discriminator, which can make a true or false discrimination for each patch of size N x N of the image. As different patches may be considered independent of each other. The pix2pix cuts a picture into different patch sizes of N x N, the discriminator judges whether each patch is true or false, and the results of all the patches of the picture are averaged to be output as a final discriminator.
PatchGAN uses a full convolution small network with NxN input, the probability that each pixel in the last layer outputs true after sigmoid is output, and then BCEloss is used for calculating final loss. This has the advantage that because the input dimension is greatly reduced, the parameters are reduced, the operation speed is faster than that of directly inputting, and the graph with any size can be calculated.
In an embodiment of the present invention, there is provided an artificial intelligence based CBCT image enhancement method, including the steps of: s1: preparing paired CBCT and CT training image data; s2: training a learning network by using the prepared pair of CBCT and CT training image data to obtain a network model for converting the CBCT image into the CT image; s3: and inputting the CBCT image into the network model to obtain an enhanced CBCT image.
In a further aspect, step S1 includes: s11: registering the CBCT image and the CT image by using a registration frame; s12: and carrying out pixel normalization and gray value normalization processing on the matched CBCT image and CT image.
In a further aspect, step S11 includes: s111: carrying out rigid registration on the CBCT image and the CT image so as to correct the global large-scale offset; s112: after rigid registration, elastic registration is further performed on the CBCT image and the CT image so as to ensure that the tissues in the body in the two groups of images correspond one to one.
In a further technical solution, in step S2, the depth learning is performed based on a pix2pix-gans model, where the pix2pix-gans model includes a generator and a discriminator, the generator generates a pseudo CT image based on an input CBCT image, and the discriminator is used for distinguishing the CBCT image from the pseudo CT image.
In a further technical scheme, the generator adopts a U-Net structure, and the arbiter adopts a PatchGAN structure.
In a further technical scheme, the U-Net structure adopts a structure that an input image is firstly down-sampled to a low dimension and then up-sampled to an original resolution, a skip-connection is added, and corresponding feature maps and feature maps with the same size after decode are spliced together according to a channel and used for reserving detail information of pixel levels under different resolutions.
In a further technical scheme, the PatchGAN structure uses GAN to construct high-frequency information, and performs true and false judgment on each patch with the size of N x N of an image, and averages the results of all the patches of one image to output as a final discriminator.
In a further technical solution, in step S12, the pixel size is unified to 1mm, the gray scale value is limited to (-1024,2048), if the pixel gray scale value is less than-1024, it is set to-1024, if the pixel gray scale value is greater than 2048, it is set to 2048, and then the gray scale value is normalized to (-1, 1).
The method of the invention can not lose the original information in the image, not only removes the noise and the artifact in the image, but also retains the edge information of the image, enhances the contrast of the image, moreover, in the process from the CBCT to the CT image training model, the model retains part of the spatial information in the CT image, and when the CBCT image is enhanced by using the model, the prior knowledge in the CT can be retained in the CBCT image.
In another embodiment of the present invention, there is also provided an artificial intelligence based CBCT image enhancement apparatus, including: the CBCT and CT training image data preparation module is used for preparing paired CBCT and CT training image data; the network model training module is used for training a learning network by using prepared paired CBCT and CT training image data to obtain a network model for converting the CBCT image into a CT image; and the enhanced CBCT image output module is used for inputting the CBCT image into the network model so as to obtain an enhanced CBCT image.
In other embodiments of the present invention, there is also provided a computer storage medium storing a computer program which, when executed, performs the method as described above.
The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various changes and modifications can be made without departing from the inventive concept of the present invention, and these changes and modifications are all within the scope of the present invention.
Claims (10)
1. A CBCT image enhancement method based on artificial intelligence is characterized by comprising the following steps:
s1: preparing paired CBCT and CT training image data;
s2: training a learning network by using the prepared pair of CBCT and CT training image data to obtain a network model for converting the CBCT image into the CT image;
s3: and inputting the CBCT image into the network model to obtain an enhanced CBCT image.
2. The method according to claim 1, wherein step S1 includes:
s11: registering the CBCT image and the CT image by using a registration frame;
s12: and carrying out pixel normalization and gray value normalization processing on the matched CBCT image and CT image.
3. The method according to claim 2, wherein step S11 includes:
s111: carrying out rigid registration on the CBCT image and the CT image so as to correct the global large-scale offset;
s112: after rigid registration, elastic registration is further performed on the CBCT image and the CT image so as to ensure that the tissues in the body in the two groups of images correspond one to one.
4. The method according to claim 1, wherein in step S2, the deep learning is performed based on a pix2pix-gans model, the pix2pix-gans model includes a generator and a discriminator, the generator generates a pseudo CT image based on the input CBCT image, and the discriminator is used for distinguishing the CBCT image from the pseudo CT image.
5. The method of claim 4, wherein the generator employs a U-Net structure and the arbiter employs a PatchGAN structure.
6. The method of claim 5, wherein the U-Net structure adopts a structure that an input image is down-sampled to a low dimension and then up-sampled to an original resolution, and a skip-connection is added, and corresponding feature maps and feature maps with the same size after decode are spliced together according to a channel to retain detail information at a pixel level at different resolutions.
7. The method of claim 5, wherein the PatchGAN structure uses GAN to construct high frequency information by performing a true-false decision for each patch of size N x N of an image and averaging the results of all the patches for a picture as the final decision device.
8. The method of claim 2, wherein in step S12, the pixel size is unified to 1mm, the gray scale value is limited to (-1024,2048), to-1024 if the pixel gray scale value is less than-1024, to 2048 if the pixel gray scale value is greater than 2048, and then the gray scale value is normalized to (-1, 1).
9. A CBCT image enhancement device based on artificial intelligence is characterized by comprising:
the CBCT and CT training image data preparation module is used for preparing paired CBCT and CT training image data;
the network model training module is used for training a learning network by using prepared paired CBCT and CT training image data to obtain a network model for converting the CBCT image into a CT image;
and the enhanced CBCT image output module is used for inputting the CBCT image into the network model so as to obtain an enhanced CBCT image.
10. A computer storage medium, characterized in that a computer program is stored which, when executed, performs the method according to any one of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110948544.0A CN113706409A (en) | 2021-08-18 | 2021-08-18 | CBCT image enhancement method and device based on artificial intelligence and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110948544.0A CN113706409A (en) | 2021-08-18 | 2021-08-18 | CBCT image enhancement method and device based on artificial intelligence and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113706409A true CN113706409A (en) | 2021-11-26 |
Family
ID=78653228
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110948544.0A Pending CN113706409A (en) | 2021-08-18 | 2021-08-18 | CBCT image enhancement method and device based on artificial intelligence and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113706409A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114558251A (en) * | 2022-01-27 | 2022-05-31 | 苏州雷泰医疗科技有限公司 | Automatic positioning method and device based on deep learning and radiotherapy equipment |
| CN114820730A (en) * | 2022-03-08 | 2022-07-29 | 安徽慧软科技有限公司 | CT and CBCT registration method based on pseudo CT |
| CN116993848A (en) * | 2023-07-21 | 2023-11-03 | 澳门理工大学 | CBCT image reconstruction method, CBCT image reconstruction system, CBCT image reconstruction computer device and CBCT image storage medium |
| CN117078612A (en) * | 2023-08-09 | 2023-11-17 | 中山大学肿瘤防治中心（中山大学附属肿瘤医院、中山大学肿瘤研究所） | CBCT image-based rapid three-dimensional dose verification method and device |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103530884A (en) * | 2013-10-21 | 2014-01-22 | 山东师范大学 | Image-guided adaptive algorithm based on edge-preserving multi-scale deformable registration |
| CN104408734A (en) * | 2014-12-11 | 2015-03-11 | 山东师范大学 | Adaptive target area conversion method combining image segmentation and deformation registration technology |
| CN106573152A (en) * | 2014-08-15 | 2017-04-19 | 皇家飞利浦有限公司 | Supervised 4-d dose map deformation for adaptive radiotherapy planning |
| CN107545137A (en) * | 2017-08-16 | 2018-01-05 | 强深智能医疗科技（昆山）有限公司 | Tumour radiotherapy plan intelligent optimization method |
| US20190333219A1 (en) * | 2018-04-26 | 2019-10-31 | Elekta, Inc. | Cone-beam ct image enhancement using generative adversarial networks |
| US20200034948A1 (en) * | 2018-07-27 | 2020-01-30 | Washington University | Ml-based methods for pseudo-ct and hr mr image estimation |
| US20210104080A1 (en) * | 2020-11-26 | 2021-04-08 | Uih America, Inc. | Systems and methods for image processing |
| WO2021068975A1 (en) * | 2019-10-12 | 2021-04-15 | Shanghai United Imaging Healthcare Co., Ltd. | Systems and methods for image reconstruction |
-
2021
- 2021-08-18 CN CN202110948544.0A patent/CN113706409A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103530884A (en) * | 2013-10-21 | 2014-01-22 | 山东师范大学 | Image-guided adaptive algorithm based on edge-preserving multi-scale deformable registration |
| CN106573152A (en) * | 2014-08-15 | 2017-04-19 | 皇家飞利浦有限公司 | Supervised 4-d dose map deformation for adaptive radiotherapy planning |
| CN104408734A (en) * | 2014-12-11 | 2015-03-11 | 山东师范大学 | Adaptive target area conversion method combining image segmentation and deformation registration technology |
| CN107545137A (en) * | 2017-08-16 | 2018-01-05 | 强深智能医疗科技（昆山）有限公司 | Tumour radiotherapy plan intelligent optimization method |
| US20190333219A1 (en) * | 2018-04-26 | 2019-10-31 | Elekta, Inc. | Cone-beam ct image enhancement using generative adversarial networks |
| US20200034948A1 (en) * | 2018-07-27 | 2020-01-30 | Washington University | Ml-based methods for pseudo-ct and hr mr image estimation |
| WO2021068975A1 (en) * | 2019-10-12 | 2021-04-15 | Shanghai United Imaging Healthcare Co., Ltd. | Systems and methods for image reconstruction |
| US20210104080A1 (en) * | 2020-11-26 | 2021-04-08 | Uih America, Inc. | Systems and methods for image processing |
Non-Patent Citations (1)
| Title |
|---|
| YANG ZHANG等: "《Improving CBCT Quality to CT Level using Deep-Learning with Generative Adversarial Network》", Retrieved from the Internet <URL:ncbi.nlm.nih.gov/pmc/articles/PMC8166936> * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114558251A (en) * | 2022-01-27 | 2022-05-31 | 苏州雷泰医疗科技有限公司 | Automatic positioning method and device based on deep learning and radiotherapy equipment |
| CN114820730A (en) * | 2022-03-08 | 2022-07-29 | 安徽慧软科技有限公司 | CT and CBCT registration method based on pseudo CT |
| CN114820730B (en) * | 2022-03-08 | 2023-04-07 | 安徽慧软科技有限公司 | CT and CBCT registration method based on pseudo CT |
| CN116993848A (en) * | 2023-07-21 | 2023-11-03 | 澳门理工大学 | CBCT image reconstruction method, CBCT image reconstruction system, CBCT image reconstruction computer device and CBCT image storage medium |
| CN116993848B (en) * | 2023-07-21 | 2024-06-11 | 澳门理工大学 | CBCT image reconstruction method, CBCT image reconstruction system, CBCT image reconstruction computer device and CBCT image storage medium |
| CN117078612A (en) * | 2023-08-09 | 2023-11-17 | 中山大学肿瘤防治中心（中山大学附属肿瘤医院、中山大学肿瘤研究所） | CBCT image-based rapid three-dimensional dose verification method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113706409A (en) | CBCT image enhancement method and device based on artificial intelligence and storage medium | |
| CN110570492B (en) | CT artifact suppression method, device and medium based on neural network | |
| CN113487657B (en) | Deep learning-based mode conversion method | |
| US9002134B2 (en) | Multi-scale image normalization and enhancement | |
| CN111325686A (en) | Low-dose PET three-dimensional reconstruction method based on deep learning | |
| CN110827216A (en) | Multi-generator generation countermeasure network learning method for image denoising | |
| CN106530236B (en) | Medical image processing method and system | |
| CN107545584A (en) | The method, apparatus and its system of area-of-interest are positioned in medical image | |
| CN112017131B (en) | CT image metal artifact removing method and device and computer readable storage medium | |
| CN111192228B (en) | Image processing method, device, CT equipment and CT system | |
| CN110599530B (en) | MVCT image texture enhancement method based on double regular constraints | |
| CN111243087B (en) | Three-dimensional reconstruction method and device for fundus blood vessel and electronic equipment | |
| KR102472464B1 (en) | Image Processing Method and Image Processing Device using the same | |
| CN108038840B (en) | Image processing method and device, image processing equipment and storage medium | |
| CN117094917B (en) | Cardiovascular 3D printing data processing method | |
| CN112598649A (en) | 2D/3D spine CT non-rigid registration method based on generation of countermeasure network | |
| CN110458908B (en) | Method for iteratively reconstructing hyper-field CT image based on limited angle | |
| CN113706643A (en) | Homomorphic adaptation learning-based head CT metal artifact correction method | |
| CN109685803B (en) | Left ventricle image segmentation method, device, equipment and storage medium | |
| WO2020212573A1 (en) | Method of metal artefact reduction in x-ray dental volume tomography | |
| CN116402954A (en) | Spine three-dimensional structure reconstruction method based on deep learning | |
| CN116029934A (en) | Low-dose DR image and CT image denoising method | |
| TWI716762B (en) | A method of using neural network to remove tooth image noise | |
| CN113570586A (en) | Method and device for creating and processing CT image of neural network system | |
| CN109272486B (en) | Training method, device and equipment of MR image prediction model and storage medium |
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 |