CN114187299A - Efficient and accurate dividing method for ultrasonic positioning tumor images - Google Patents
Efficient and accurate dividing method for ultrasonic positioning tumor images Download PDFInfo
- Publication number
- CN114187299A CN114187299A CN202111254978.7A CN202111254978A CN114187299A CN 114187299 A CN114187299 A CN 114187299A CN 202111254978 A CN202111254978 A CN 202111254978A CN 114187299 A CN114187299 A CN 114187299A
- Authority
- CN
- China
- Prior art keywords
- image
- region
- tumor
- sub
- roi
- 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- Withdrawn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
- 206010028980 Neoplasm Diseases 0.000 title claims abstract description 39
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 238000001454 recorded image Methods 0.000 claims abstract description 4
- 238000006243 chemical reaction Methods 0.000 claims abstract description 3
- 238000002604 ultrasonography Methods 0.000 claims description 11
- 238000005192 partition Methods 0.000 claims description 3
- 230000001788 irregular Effects 0.000 claims description 2
- 230000011218 segmentation Effects 0.000 claims 4
- 230000002708 enhancing effect Effects 0.000 claims 1
- 239000000284 extract Substances 0.000 abstract 1
- 239000000523 sample Substances 0.000 description 12
- 239000003814 drug Substances 0.000 description 5
- 239000003086 colorant Substances 0.000 description 1
- 238000002591 computed tomography Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000002600 positron emission tomography Methods 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000002560 therapeutic procedure Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/12—Edge-based segmentation
-
- 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
- G06T2207/10024—Color image
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10104—Positron emission tomography [PET]
-
- 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
- G06T2207/10132—Ultrasound image
-
- 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
- G06T2207/30096—Tumor; Lesion
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Ultra Sonic Daignosis Equipment (AREA)
Abstract
The invention discloses a high-efficiency and accurate dividing method for an ultrasonic positioning tumor image, which relates to the technical field of medical images and comprises the following steps: s1: the method comprises the steps of shooting an ultrasonic positioning tumor image of a patient, recording the shot image, accessing the recorded image into a storage device of a computer, collecting medical image pictures of the same tumor focus under three dimensions to form a data set, and preprocessing the ultrasonic positioning image; s2: extracting ROI (region of interest) areas of the data set, performing gray level image conversion on the preprocessed image, converting the PET/CT data set image with the false color into a gray level image, extracting the ROI area image, extracting local features, namely the ROI area image corresponding to a focus area, from the global gray level image of the data set according to clinical marks, and normalizing the ROI area image. The invention extracts the texture characteristics in the sub-region with the assistance of a computer and combines the division rule of the sub-region, thereby better representing the boundary of the tumor and simultaneously reducing the labor intensity of doctors.
Description
Technical Field
The invention relates to the technical field of medical images, in particular to an efficient and accurate dividing method for an ultrasonic positioning tumor image.
Background
Ultrasonic medicine is a subject combining acoustics, medicine, optics and electronics, and the application of an acoustic technology for researching a frequency higher than an audible sound frequency in the medical field, namely ultrasonic medicine, including ultrasonic diagnosis, ultrasonic therapy and biomedical ultrasonic engineering, so the ultrasonic medicine has the characteristics of combining medicine, theory and engineering, has wide related contents and has high value in preventing, diagnosing and treating diseases.
The method has important significance and application value for accurately dividing the focus area of the ultrasonic image, can guide a treatment probe to focus high-intensity ultrasonic on the focus area, can improve treatment effect, and can avoid damaging other tissues.
Disclosure of Invention
The invention aims to solve the problems that in the prior art, the division of an ultrasonic image is mostly completed manually and is very complicated, the treatment efficiency is reduced, and huge workload is brought to doctors, and provides an efficient and accurate division method of an ultrasonic positioning tumor image.
In order to achieve the purpose, the invention adopts the following technical scheme:
the high-efficiency and accurate division method for the ultrasonic positioning tumor image comprises the following steps:
s1: the method comprises the steps of shooting an ultrasonic positioning tumor image of a patient, recording the shot image, accessing the recorded image into a storage device of a computer, collecting medical image pictures of the same tumor focus under three dimensions to form a data set, and preprocessing the ultrasonic positioning image;
s2: extracting an ROI (region of interest) region from the data set, performing gray image conversion on the preprocessed image, converting the PET/CT data set image with the false color into a gray image, extracting the ROI region image, extracting local features, namely the ROI region image corresponding to a focus region, from the global gray image of the data set according to clinical marks, and normalizing the ROI region image;
s3: extracting local features to obtain an ROI image of the focus, and collecting the ROI image in a same scale as the original image, wherein each sample subset displays a partition boundary region by using a dotted line.
Preferably, the preprocessing of the ultrasound positioning image comprises denoising processing of the ultrasound positioning image and enhancement processing of the ultrasound positioning image.
Preferably, the negative sample set and the positive sample set of each sample subset are divided into 5 uniform samples by using a dividing algorithm, respectively, so as to obtain a 5-fold crossed sample division set.
Preferably, the image is over-divided into a plurality of sub-regions, the sub-regions are called super-pixels, the sub-regions are extracted with texture features through computer assistance, and finally the sub-regions are clustered according to the texture features and the spatial features to obtain the final division result.
Preferably, a plurality of pixel points are selected as initial clustering centers, the pixel points are evenly distributed on the image, then the pixel points are used as the clustering centers to update the clustering, and the clustering centers are updated repeatedly until a convergence condition is met, wherein the convergence condition is that iteration is carried out to a certain number of times, or the clustering centers are not changed within a certain error limit.
Preferably, the non-connected sub-region is divided into a division set which is adjacent to the sub-region and has the gray mean value closest to the sub-region, so that the formed super-pixel is more homogeneous inside, and the super-pixel boundary is better matched with the tumor boundary.
Preferably, if a region is a tumor region, the region cannot include image boundary pixels, the tumor region is completely represented in the ultrasound image, and the region including image boundary pixels identified by the change of the global gray scale is considered to be incomplete in that the region is truncated by the image boundary pixels.
Preferably, the very irregular region is a sub-region divided from the background region, for each region, the minimum circumscribed ellipse of the region is found, the circumscribed ellipse is compared with the area of the nearest sub-region, and if the ratio is within a set range, the region can be regarded as a tumor region.
Compared with the prior art, the invention has the following advantages:
the invention improves the image definition by preprocessing the ultrasonic positioning image, obtains a crossed sample division set by dividing the ultrasonic image of the focus into a plurality of subset samples, better embodies the boundary of the tumor by extracting the textural features in the sub-region with the assistance of a computer and combining the division rules of the sub-region, and simultaneously reduces the labor intensity of a doctor.
Detailed Description
The high-efficiency and accurate division method for the ultrasonic positioning tumor image comprises the following steps:
s1: the method comprises the steps of shooting an ultrasonic positioning tumor image of a patient, recording the shot image, storing the recorded image into a storage device of a computer, collecting medical image pictures of the same tumor focus under three dimensions to form a data set, and partitioning an area of an opposite tumor when later division processing is facilitated to carry out pretreatment on the ultrasonic positioning image;
the preprocessing of the ultrasonic positioning image comprises denoising processing of the ultrasonic positioning image and enhancement processing of the ultrasonic positioning image, so that the ultrasonic positioning image is clearer, and image blurring caused by external factors of clothes can be eliminated.
S2: extracting an ROI (region of interest) region from the data set, converting a gray level image of a preprocessed image to facilitate the production of an ultrasonic positioning image, converting a PET/CT (positron emission tomography/computed tomography) data set image with pseudo colors into a gray level image, extracting the ROI region image, extracting local features, namely an ROI region image corresponding to a focus region, from a global gray level image of the data set according to clinical markers, and normalizing the ROI region image;
the negative sample set and the positive sample set of each sample subset are divided into 5 uniform parts by adopting a dividing algorithm respectively to obtain 5-fold crossed sample division sets, and the tumor position can be judged aiming at the sub-regions divided subsequently through the division sets.
The image is over-divided into a plurality of sub-regions, the sub-regions are called super-pixels, the performance of the natural landscape image is better in time consumption and over-division precision compared with other methods, the sub-regions are extracted with texture features through computer assistance, and finally the sub-regions are clustered according to the texture features and the space characteristics to obtain the final division result, so that the working intensity of manual division of a doctor is reduced, the division speed and precision are improved, and the treatment efficiency is improved.
The method divides the disconnected sub-regions into the division sets which are adjacent to the sub-regions and the gray mean value of which is closest to the sub-regions, so that the formed superpixels are more uniform, the boundaries of the superpixels better accord with the tumor boundaries, the tumor position regions and the bar boundaries can be distinguished more clearly, and the method is particularly important for doctors to issue operation schemes.
S3: extracting local features to obtain an ROI image of the focus, and collecting the ROI image in a same scale as the original image, wherein each sample subset displays a partition boundary region by using a dotted line.
The method comprises the steps of selecting a plurality of pixel points as initial clustering centers, enabling the pixel points to be evenly distributed on an image, then enabling the pixel points to serve as the clustering centers to be updated and clustered, continuously iterating and updating the clustering centers and the clustering centers until a convergence condition is met, wherein the convergence condition is that iteration is carried out for a certain number of times or the clustering centers are not changed within a certain error limit, and after final updating and clustering are completed, forming subsets after region division is completed by taking each clustering center as an origin.
If one region is a tumor region, the region cannot contain image boundary pixels, the tumor region is completely presented in the ultrasonic image, and the region containing the image boundary pixels can be considered to be truncated by the image boundary pixels through the change identification of the global gray scale, so that the region is incomplete, the tumor region only appears in the completed subset region, and the division of the tumor boundary is clearer.
The method comprises the steps of dividing a background region into sub-regions, finding the minimum circumscribed ellipse of each region, comparing the areas of the circumscribed ellipse and the nearest sub-regions, and if the ratio is within a set range, determining that the region is a tumor region, rapidly finding out the tumor position in the divided sub-regions according to gray level changes, facilitating the formulation of an operation scheme by a doctor and improving the treatment efficiency.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and equivalent alternatives or modifications according to the technical solution and the inventive concept of the present invention should be covered by the scope of the present invention.
Claims (8)
1. The method for efficiently and accurately dividing the ultrasonic positioning tumor image is characterized by comprising the following steps of:
s1: the method comprises the steps of shooting an ultrasonic positioning tumor image of a patient, recording the shot image, accessing the recorded image into a storage device of a computer, collecting medical image pictures of the same tumor focus under three dimensions to form a data set, and preprocessing the ultrasonic positioning image;
s2: extracting an ROI (region of interest) region from the data set, performing gray image conversion on the preprocessed image, converting the PET/CT data set image with the false color into a gray image, extracting the ROI region image, extracting local features, namely the ROI region image corresponding to a focus region, from the global gray image of the data set according to clinical marks, and normalizing the ROI region image;
s3: extracting local features to obtain an ROI image of the focus, and collecting the ROI image in a same scale as the original image, wherein each sample subset displays a partition boundary region by using a dotted line.
2. The method for efficient and accurate segmentation of ultrasound localized tumor images as claimed in claim 1, wherein the pre-processing of the ultrasound localized images includes de-noising the ultrasound localized images and enhancing the ultrasound localized images.
3. The method for efficient and accurate segmentation of ultrasound localized tumor images as claimed in claim 1, wherein the segmentation algorithm is employed to segment the negative sample set and the positive sample set of each sample subset into 5 uniform samples, resulting in 5-fold cross sample segmentation sets.
4. The method of claim 3, wherein the image is over-divided into a plurality of sub-regions, which are called super-pixels, texture features are extracted from the sub-regions with the aid of a computer, and the sub-regions are clustered according to the texture features and spatial characteristics to obtain final division results.
5. The method for efficiently and accurately dividing an ultrasound localized tumor image according to claim 1, wherein a plurality of pixel points are selected as initial clustering centers, the pixel points are evenly distributed on the image, the pixel points are used as clustering centers to perform update clustering, and the clustering and clustering centers are continuously updated iteratively until a convergence condition is met, wherein the convergence condition is that iteration is performed to a certain number of times, or the clustering centers are not changed within a certain error limit.
6. The method of claim 4, wherein the non-connected sub-region is divided into a set of adjacent sub-regions with the mean value of gray scale closest to the sub-region, so that the formed super-pixel is more homogeneous and the boundary of the super-pixel better conforms to the boundary of the tumor.
7. The method of claim 1, wherein if a region is a tumor region, the region cannot contain image boundary pixels, the tumor region is completely represented in the ultrasound image, and the region containing the image boundary pixels is identified by the change of the global gray level as if the region is truncated by the image boundary pixels.
8. The method of claim 1, wherein the highly irregular region is a sub-region divided from a background region, and for each region, the minimum circumscribed ellipse is found, and the circumscribed ellipse is compared with the area of the nearest sub-region, and if the ratio is within a predetermined range, the region is considered as a tumor region.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111254978.7A CN114187299A (en) | 2021-10-27 | 2021-10-27 | Efficient and accurate dividing method for ultrasonic positioning tumor images |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111254978.7A CN114187299A (en) | 2021-10-27 | 2021-10-27 | Efficient and accurate dividing method for ultrasonic positioning tumor images |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114187299A true CN114187299A (en) | 2022-03-15 |
Family
ID=80601587
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111254978.7A Withdrawn CN114187299A (en) | 2021-10-27 | 2021-10-27 | Efficient and accurate dividing method for ultrasonic positioning tumor images |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114187299A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117808813A (en) * | 2024-03-01 | 2024-04-02 | 乐恩(北京)医药技术有限公司 | Respiratory system lung nodule tumor cell image detection method and system and storage medium |
-
2021
- 2021-10-27 CN CN202111254978.7A patent/CN114187299A/en not_active Withdrawn
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117808813A (en) * | 2024-03-01 | 2024-04-02 | 乐恩(北京)医药技术有限公司 | Respiratory system lung nodule tumor cell image detection method and system and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110738701B (en) | Tumor three-dimensional positioning system | |
| US20220157047A1 (en) | Feature Point Detection | |
| CN111178369B (en) | Medical image recognition method and system, electronic equipment and storage medium | |
| CN111340825A (en) | Method and system for generating mediastinal lymph node segmentation model | |
| CN111292307A (en) | Digestive system gallstone recognition method and positioning method | |
| CN116503607B (en) | CT image segmentation method and system based on deep learning | |
| Tan et al. | An approach for pulmonary vascular extraction from chest CT images | |
| CN114332132A (en) | Image segmentation method and device and computer equipment | |
| CN113935976A (en) | Method and system for automatically segmenting blood vessels in internal organs by enhancing CT (computed tomography) image | |
| CN111862058A (en) | Efficient and accurate dividing method for ultrasonic positioning fibroid image | |
| Matsopoulos et al. | Thoracic non-rigid registration combining self-organizing maps and radial basis functions | |
| CN113538209A (en) | Multi-modal medical image registration method, registration system, computing device and storage medium | |
| CN114187299A (en) | Efficient and accurate dividing method for ultrasonic positioning tumor images | |
| CN116580033B (en) | Multi-mode medical image registration method based on image block similarity matching | |
| CN112164447B (en) | Image processing method, device, equipment and storage medium | |
| CN116309647B (en) | Method for constructing craniocerebral lesion image segmentation model, image segmentation method and device | |
| CN112686897A (en) | Weak supervision-based gastrointestinal lymph node pixel labeling method assisted by long and short axes | |
| CN109816665B (en) | Rapid segmentation method and device for optical coherence tomography image | |
| CN110858412B (en) | Heart coronary artery CTA model building method based on image registration | |
| CN116258732A (en) | Esophageal cancer tumor target region segmentation method based on cross-modal feature fusion of PET/CT images | |
| CN109671131B (en) | Image correction method, device, medical image equipment and storage medium | |
| Upadhyay et al. | Analysis of Liver MR Images for Cancer Detection using Genetic Algorithm | |
| Zhou et al. | Two stage registration-based automatic left ventricle myocardium segmentation of cardiac 4DCT images | |
| JP5403431B2 (en) | Tomographic image processing method and apparatus | |
| CN117036372B (en) | Robust laser speckle image blood vessel segmentation system and segmentation method |
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: 20220315 |