WO2020077426A1 - Método para automação de teste de resolução em imagens digitais - Google Patents
Método para automação de teste de resolução em imagens digitais Download PDFInfo
- Publication number
- WO2020077426A1 WO2020077426A1 PCT/BR2019/050444 BR2019050444W WO2020077426A1 WO 2020077426 A1 WO2020077426 A1 WO 2020077426A1 BR 2019050444 W BR2019050444 W BR 2019050444W WO 2020077426 A1 WO2020077426 A1 WO 2020077426A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- image
- phantom
- structures
- resolved
- fact
- Prior art date
Links
- 238000012360 testing method Methods 0.000 title claims abstract description 74
- 238000000034 method Methods 0.000 title claims abstract description 54
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 19
- 238000010801 machine learning Methods 0.000 claims abstract description 14
- 238000012549 training Methods 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 6
- 238000002059 diagnostic imaging Methods 0.000 claims description 3
- 230000015572 biosynthetic process Effects 0.000 claims description 2
- 230000002093 peripheral effect Effects 0.000 claims description 2
- 238000012545 processing Methods 0.000 claims description 2
- 230000007480 spreading Effects 0.000 claims description 2
- 230000004044 response Effects 0.000 description 12
- 238000005259 measurement Methods 0.000 description 10
- 239000000243 solution Substances 0.000 description 9
- 238000002595 magnetic resonance imaging Methods 0.000 description 8
- 238000007477 logistic regression Methods 0.000 description 7
- 238000003908 quality control method Methods 0.000 description 7
- NIXOWILDQLNWCW-UHFFFAOYSA-N acrylic acid group Chemical group C(C=C)(=O)O NIXOWILDQLNWCW-UHFFFAOYSA-N 0.000 description 6
- 238000011156 evaluation Methods 0.000 description 6
- 239000004033 plastic Chemical group 0.000 description 6
- 229920003023 plastic Chemical group 0.000 description 6
- 238000007637 random forest analysis Methods 0.000 description 5
- 238000010200 validation analysis Methods 0.000 description 5
- 230000016776 visual perception Effects 0.000 description 5
- FAPWRFPIFSIZLT-UHFFFAOYSA-M Sodium chloride Chemical compound [Na+].[Cl-] FAPWRFPIFSIZLT-UHFFFAOYSA-M 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000012706 support-vector machine Methods 0.000 description 4
- 230000007423 decrease Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000003384 imaging method Methods 0.000 description 3
- 238000002360 preparation method Methods 0.000 description 3
- 238000013442 quality metrics Methods 0.000 description 3
- 238000003325 tomography Methods 0.000 description 3
- 101100008044 Caenorhabditis elegans cut-1 gene Proteins 0.000 description 2
- PXHVJJICTQNCMI-UHFFFAOYSA-N Nickel Chemical compound [Ni] PXHVJJICTQNCMI-UHFFFAOYSA-N 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000003066 decision tree Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 230000004069 differentiation Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 230000004304 visual acuity Effects 0.000 description 2
- 230000003442 weekly effect Effects 0.000 description 2
- ZAMOUSCENKQFHK-UHFFFAOYSA-N Chlorine atom Chemical compound [Cl] ZAMOUSCENKQFHK-UHFFFAOYSA-N 0.000 description 1
- 201000009273 Endometriosis Diseases 0.000 description 1
- 241000282412 Homo Species 0.000 description 1
- DGAQECJNVWCQMB-PUAWFVPOSA-M Ilexoside XXIX Chemical compound C[C@@H]1CC[C@@]2(CC[C@@]3(C(=CC[C@H]4[C@]3(CC[C@@H]5[C@@]4(CC[C@@H](C5(C)C)OS(=O)(=O)[O-])C)C)[C@@H]2[C@]1(C)O)C)C(=O)O[C@H]6[C@@H]([C@H]([C@@H]([C@H](O6)CO)O)O)O.[Na+] DGAQECJNVWCQMB-PUAWFVPOSA-M 0.000 description 1
- 229920002319 Poly(methyl acrylate) Polymers 0.000 description 1
- 206010060862 Prostate cancer Diseases 0.000 description 1
- 208000000236 Prostatic Neoplasms Diseases 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000017531 blood circulation Effects 0.000 description 1
- 210000000988 bone and bone Anatomy 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 239000003795 chemical substances by application Substances 0.000 description 1
- 239000000460 chlorine Substances 0.000 description 1
- 229910052801 chlorine Inorganic materials 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 230000001010 compromised effect Effects 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000037213 diet Effects 0.000 description 1
- 235000005911 diet Nutrition 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 239000012530 fluid Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 238000002347 injection Methods 0.000 description 1
- 239000007924 injection Substances 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000000873 masking effect Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 210000003205 muscle Anatomy 0.000 description 1
- 208000010125 myocardial infarction Diseases 0.000 description 1
- 229910052759 nickel Inorganic materials 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 239000002985 plastic film Substances 0.000 description 1
- 229920006255 plastic film Polymers 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 239000011734 sodium Substances 0.000 description 1
- 229910052708 sodium Inorganic materials 0.000 description 1
- 239000011780 sodium chloride Substances 0.000 description 1
- 238000010561 standard procedure Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000010998 test method Methods 0.000 description 1
- 210000001519 tissue Anatomy 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- 238000011282 treatment Methods 0.000 description 1
- 238000011269 treatment regimen Methods 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01T—MEASUREMENT OF NUCLEAR OR X-RADIATION
- G01T1/00—Measuring X-radiation, gamma radiation, corpuscular radiation, or cosmic radiation
- G01T1/16—Measuring radiation intensity
- G01T1/169—Exploration, location of contaminated surface areas
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/58—Testing, adjusting or calibrating thereof
- A61B6/582—Calibration
- A61B6/583—Calibration using calibration phantoms
Definitions
- the present invention relates to a method for automating resolution testing on digital images and, more specifically, a method that emulates human visual perception in the task of identifying structures in an image with low contrast against a troubled background.
- the method of the present invention can, for example, be used for the automation of quality tests of magnetic resonance tomographs.
- Magnetic resonance tomography (2D or 3D) is particularly useful because it offers a huge range of image contrasts without the use / injection of enhancement agents.
- the magnetic resonance tomography can be programmed to produce images that reveal fractures in bones, as in the X-ray modality, but they can also be prepared to highlight the differences between muscle and fat; or to register structures with slight differences in composition in relation to their surroundings. This ability of MRI tomography to allow small structures to be resolved in low contrast, makes it particularly useful in the radiological evaluation of meniscal tears, myocardial infarctions, prostate cancer, endometriosis, etc; to name a few examples.
- the MRI scanner must be routinely subjected to quality tests, which aims to ensure that the scanner is generating images according to its specifications; that meet minimum quality standards, such as those recommended by the American College of Radiology (ACR).
- ACR American College of Radiology
- the ACR proposes an extensive quality control program and issues certificates in the RM modality, among others. In the United States, all institutions that serve the government must adhere to the ACR recommendations and present an accreditation certificate (s). In the rest of the world, the program is adopted by institutions that recognize the importance of monitoring the quality of the radiological images they produce, as well as the adequacy of the ACR proposal.
- the quality control of medical images is essential for the development and application of so-called quantitative analysis methods, which allow the establishment of more objective diagnostic criteria and personalized medical recommendations.
- Most of the methods applied in clinical practice are still qualitative, that is, they form images in which the pixels vary in a gray scale of an arbitrary unit, with no medical / physiological significance. In these images what matters is the relationship between the signals and the pattern of their distribution.
- Quantitative imaging methods are those that allow you to generate images whose pixels bring objective information about the tissue under analysis, such as degree of stiffness, levels of certain substances / molecules, blood flow rate, etc. In these images, the pixel signals vary within a range of values with medical / physiological significance.
- Deviations in quality measurements indicate that the clinical images generated by the CT scanner may be compromised and that it needs maintenance / calibration.
- Quality tests are performed on images of an object of known geometry and composition (which is called phantom), and include measures of distortion, shading and resolution.
- a good quality image must represent this object with known dimensions and characteristics. And a supposedly high-resolution image should allow you to distinguish small details in that object.
- the signal in the field of view must be homogeneous - with subliminal shading / glows - and free of artifacts.
- Resolving means differentiating.
- a well-resolved image is one that allows the observer to see details of the recorded scene or object, i.e., better differentiate its elements. So, when it comes to detection, resolving is to assert that there is actually something, a structure, in a given region of the image that is different from the background, or its surroundings, and that it should represent, with some degree of confidence that is automatically established by our brains, a true structure at the origin.
- contrast i.e., the relative difference between the captured signal of the structure in the field of view and the background signal of the image. The lower the contrast, the lower the resolution / differentiation power.
- the other determinants of the resolution of an image are quality parameters, such as signal-to-noise level (which reflects the instrument's sensitivity) and signal scattering (which reflects other limitations and imperfections).
- the spatial resolution power of an image is an image quality metric that indicates the minimum size that a structure it must have to be appreciated by someone who observes / "reads" the image.
- the resolution power of high-contrast structures in a digital image is exactly the size of its pixel. If it is not possible to differentiate two structures the size of a pixel, which are a pixel away and emit a high signal, against a background considered dark, the image must be with serious quality problems.
- the resolution power of structures with signals very close to the background signal is less and decreases with contrast. In part, this can be considered a matter of semantics since a lower contrast structure is, as the name implies, less differentiable. But only in part, given that the power of differentiation does not depend solely and exclusively on the difference in signal between the structure and the background, but also on other parameters of image quality.
- a high resolution image allows an observer to solve, that is, to differentiate / detect small structures in the image.
- the resonance tomograph must generate images with minimal resolution power, which satisfy a certain acceptance criterion. Acceptance criteria may differ for systems of different configurations. It is expected, of course, that 3T scanners will generate images with greater resolving power than 15T devices, even if with a lower level of signal homogeneity. Therefore, tomographs with different magnetic fields are treated as different instruments; and are tested by observing different acceptance criteria, according to their potential and limitations.
- the weekly image quality tests proposed by the ACR involve collecting images in a multi-purpose phantom and evaluating a portion of the images in the series. To this is added the preparation time - both with the careful placement of the phantom in the center of the tomograph and with the programming of the acquisition protocol. The evaluation of the images must be performed immediately after acquisition, in order to allow the repetition of the test / acquisition if necessary. Not infrequently, during the evaluation, faults are detected that are the result of simple misplacement of the phantom inside the equipment - a trivial error that must be remedied immediately. To prevent this type of rework, many use a support to help position the phantom. In addition to reducing the preparation time and the frequency of errors, the use of support also increases the reproducibility of the tests. It is worth mentioning that, even if a support is adopted, it is recommended that the user does not leave the images to analyze them later.
- the image quality tests recommended by the ACR to be repeated weekly are those of geometric distortion, spatial resolution, presence of artifacts, central frequency and precision of the displacement to the magnet isocenter.
- the central frequency tests are limited to recording the value of the resonance frequency at the time of the test.
- the isocenter test is limited to measuring whether the phantom region where the reference lasers were marked appears in the image with the Z coordinate close to 0.
- Geometric distortion tests require the operator to measure distances - more precisely, the distance height and internal diameter of the phantom cylinder - using virtual rulers handled on the screen using a mouse. The operator also evaluates the series of images for the presence of artifacts, of any kind, which he considers significant to mention. In resolution tests, the operator must indicate whether or not he is able to differentiate structures of different dimensions and contrasts, against the background of the image, which can be quite troubled.
- the present invention achieves these and other objectives through a method for automating resolution testing on digital images, which comprises: providing a digital image with at least one structure that must be discerned in the image; extract characteristics from the digital image data, said characteristics including characteristics related to at least one structure that must be discerned; and classify the at least structure as resolved or not resolved through a machine learning algorithm, in which the algorithm was trained based on solvability predictors that include characteristics extracted from an image database of previously obtained structures and classification data of these structures as resolved or not resolved, such classification being performed by at least one human operator.
- the classification of the training base data is performed by at least two operators with different levels of experience and in which the levels of experience of the operators are included as predictors of resolvability.
- Predictors of resolvability also include data on a source device for the image provided.
- the method of the present invention can use a test phantom from a medical imaging machine.
- the step of providing an image can comprise a step of providing an image of a phantom captured by a medical image acquisition machine.
- the extracted data is image data from at least one image from at least one cut of the phantom; and the image database of previously obtained structures is an image database of previously obtained phantom sections.
- the solvability predictors also include data about the medical image acquisition machine and data about the medical image acquisition conditions.
- the image base can be formed through the collection of images on different medical image acquisition machines; and the processing of the collected images to generate new images with characteristics of resolution, contrast, homogeneity and signal spreading varied.
- the medical image acquisition machine is an MRI machine
- the phantom is an ACR type phantom.
- the phantom is a phantom of the type comprising a plurality of different diameter holes arranged in radial triplets, in which: the image data is image data from at least one image of at least one axial slice of the phantom comprising an axial slice of the plurality of different holes, the image comprising a plurality of structures that can be discerned, such structures being defined by the image of the holes; the extraction of characteristics of the structures defined by the holes; and resolvability predictors include features extracted from an image database of axial phantom slices previously obtained and from classification data, by a human operator, of the holes as resolved or unsolved.
- the extraction of characteristics from the structures is performed based on four regions of interest: a region of the hole, a region peripheral to the hole, a region adjacent to the hole and a region in line that passes through the three holes of each triplet.
- an image signal is obtained for each region of interest and information about average and standard deviation of the signal is extracted.
- Figure 1 - is an image of the ACR phantom used for a low contrast resolution test according to an embodiment of the method of the present invention
- Figure 2 - is an image showing different masks that can be used in the resolution tests according to embodiments of the method of the present invention.
- Figure 3 - is an image of an example screen of the interface for training the algorithm of the method of the present invention.
- the present invention is based on a method for automating the quality test of magnetic resonance equipment that, using a machine learning algorithm, is capable of emulating the ability of human visual perception and similarly judging to a human, whether a particular structure can be seen / detected / resolved in the image or not.
- the machine learning algorithm is trained using image data of structures associated with classification data of that structure, and the classification used for training was performed by human operators.
- test object image data is obtained, where the images show a plurality of structures that can be discerned from an image background.
- Features are extracted from the image data obtained - at least some of the features being related to the structures that could be discerned - and the structures are classified as resolved or not resolved through a machine learning algorithm.
- the classification of image structures as resolved or unresolved is carried out by operators with different experience levels, and the predictors also include data on the operators' experience levels.
- Predictors may also include data on the resonance equipment being tested and other peculiarities of the acquisition, such as, for example, reception band, type of antenna and speed of the gradients.
- the ACR phantom is an acrylic cylinder 190 mm in diameter and 148 mm high / length, containing several acrylic and plastic structures inside, each serving a purpose (type of measurement) different).
- the remainder of the interior of the cylinder is filled with a Nickel, Chlorine and Sodium saline solution.
- a minimum fraction is filled with air, as a way to prevent thermal expansion of the fluid from damaging the casing (for this reason it is normal to see bubbles inside the phantom).
- pixels that represent the saline solution have a maximum signal (white) in the image and pixels that represent acrylic / plastic have a minimum signal (black). Pixels from regions that encompass both materials have an intermediate signal. Per For example, it is normal for pixels to fall on the edge of the acrylic structures, the pixel being darker the more it is displaced inside the structure. This does not mean, however, that pixels entirely inside the structure are perfectly black. Noise and homogeneous slice of signal reception in the antennas, as well as other types of scattering observed in resonance images, can contaminate the pixel signal, making it clearer.
- the phantom was designed to generate images with two very distinct signal types - that is, of the solution vs. acrylic / plastic (with the exception of regions for low-contrast tests, designed to generate intermediate signals) - but their images are not black and white, but in gray scale, with very high contrast.
- a low contrast resolution test is a test that aims to determine the resolution power of the images generated by an equipment when the imaged structures have low contrast, that is, when the structures that are sought to differentiate have little signal highlight in relation to their surroundings or background. Thus, this test aims to verify the equipment's ability to offer images with sufficient quality to allow differentiating low contrast structures.
- the test involves the acquisition of four axial cuts in the posterior ACR phantom region, where thin circular plastic films are found, drilled each with 30 diameter holes. varied, which are organized in 10 radial triplets (the term triplets denotes the set of three holes of the same diameter, radially aligned).
- the four axial cuts correspond to cuts 8-1 1 that are obtained with the recommended acquisition protocol.
- Films of plastic found in that position of the phantom have a different thickness in each volume, which is what determines the contrast between the holes and the bottom of the image.
- FIG. 1 shows an example of an image of section 10 of the ACR phantom, with T triplets indicated.
- the ACR low contrast resolution test essentially consists of counting how many of the 10 hole triplets can be resolved in each of the 1 to 8 cuts. Counting is performed from the triplets with larger holes diameter for smaller ones; and from cut 1 1, where the contrast is greater, up to cut 8. The triplet is considered visible when all 3 holes can be clearly detected; and counting should cease as soon as any of the holes in a triplet are no longer visible, even if one or more holes are visible in any of the subsequent triplets. Naturally, the higher the count, the better the image quality, as it allows to solve smaller structures.
- the operator assesses the image's resolution power by assessing the detectability of structures that show little difference in signal with respect to their surroundings.
- a structure is perceived in an image (that is, resolved, detected, visualized) when we are sure that something (that is, the structure) stands out from the background.
- the visual perception of structures depends as much on the size and sign of the structure as on their relationship with their surroundings.
- image problems such as distortions, shading, edge effects and other artifacts are absent
- the resolving power can be modeled reasonably well using empirical equations like the one known by DeVries-Rose law, which relates signal, contrast and background noise to the resolution power of the image.
- DeVries-Rose law which relates signal, contrast and background noise to the resolution power of the image.
- modeling of the detection of structures by a human requires special treatment.
- the present invention models this problem using the machine learning method learning).
- the proposed algorithm can be applied to classify structures as visible or imperceptible in a digital image, according to the level of signal, contrast, noise and other quality metrics such as geometric distortions, shading and signal scattering.
- ROIs are not square, metrics related to signal dispersion in a single dimension were calculated.
- the image signal under each ROI formed vectors, from which the signal's mean and standard deviation (noise) were extracted (the two most important metrics describing the quality of an image); as well as the average of the lower, middle and upper thirds. These metrics were extracted in five different conditions: 1) from the original image, 2) from the normalized image between the values 0 and 1, and 3) from the normalized image in each sector separately.
- Another possibility is to use square masks and use available algorithms to calculate metrics related to the spatial distribution of signals.
- the low-contrast holes in cuts 8 to 1 1 are within a gray circle with black edges.
- the first action is to locate, at each cut, the center and radius (R) of that circle, which will determine the positions of the ROIs.
- Sectors The inner circle is divided into 10 sectors of 36 degrees of opening each. When the procedure is applied to section 8, the sectors are arranged so that one of them (say sector n. 1) is aligned with the vertical axis, in the 90 ° position. When the procedure is applied to cut 9, the sectors rotate 9 o clockwise; in section 10, the rotation is 18 o ; and in section 11, of 27 °.
- Two circumferences of radii corresponding to 40% and 70% of R divide the circle into: 1) a circle less than 40% of R; 2) a crown with an internal radius 40% of R and an external radius 70% of R; and 3) a crown with an internal radius 70% of R and an external radius 100% of R.
- Lines 30 straight segments are formed, 3 between each triplet of holes, in the radial direction and in length corresponding to 40% of R and ending at the edge of the circle.
- the segments are 14 °, 18 ° and 22 ° away from the center of the sectors.
- Circles The formation of circle-type ROIs took place slightly differently.
- the rays of the low contrast holes are known.
- the radii of the holes in each of the 10 triplets are: 7.0mm, 6.0mm, 5.0mm, 4.5mm, 4.0mm, 3.5mm, 3.0mm, 2.5mm, 2.0mm and 1.5mm. To determine the positions, the following strategy was chosen.
- the positioning of the ROIs is then performed by adjusting the template of positions for each image according to the center and radius (R) of the detected circle.
- additional predictive variables can be used by the method of the present invention. Among them are listed: level of ghosting, signal homogeneity and geometric distortion.
- acquisition conditions such as hardware type (viz. Magnetic field of the MRI scanner and antenna) and reception band (which is not determined by the ACR), could help to improve the performance of the algorithm and, therefore, they are also among the possible explanatory variables (predictors).
- LR Logistic Regression
- SVM Support Vector Machine
- RF Random Forest
- XGB Extreme Gradient Boosting
- NNet Neural Network
- Logistic Regression it is the most used method in problems of binary classification, as the approached in the context of this invention. It was implemented using the standard method of generalized linear model in R.
- Support Vector Machine it is a widely used Machine Learning method and is known to perform well in binary classification problems. This algorithm was implemented using the R. e1071 package.
- Random Forest it is a method that builds a multiplicity of decision trees at the time of training. It was implemented using the R Random Forest package.
- Extreme Gradient Boosting XGB: it is a method that typically makes use of decision trees and optimizes a differentiable loss function. R.'s xgboost package was used.
- Neural Network the use of a feed-forward network with a single hidden layer (hidden layer) of 10 nodes was used. It was implemented using the R. nnet package.
- Table II shows the main parameters used in each method.
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- High Energy & Nuclear Physics (AREA)
- Molecular Biology (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Biomedical Technology (AREA)
- Spectroscopy & Molecular Physics (AREA)
- General Physics & Mathematics (AREA)
- Optics & Photonics (AREA)
- Pathology (AREA)
- Radiology & Medical Imaging (AREA)
- Biophysics (AREA)
- Heart & Thoracic Surgery (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
- Image Analysis (AREA)
Abstract
Description
Claims
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
BRBR1020180712934 | 2018-10-16 | ||
BR102018071293A BR102018071293A2 (pt) | 2018-10-16 | 2018-10-16 | método para automação de teste de resolução em imagens digitais |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2020077426A1 true WO2020077426A1 (pt) | 2020-04-23 |
Family
ID=70282877
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/BR2019/050444 WO2020077426A1 (pt) | 2018-10-16 | 2019-10-16 | Método para automação de teste de resolução em imagens digitais |
Country Status (2)
Country | Link |
---|---|
BR (1) | BR102018071293A2 (pt) |
WO (1) | WO2020077426A1 (pt) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100086182A1 (en) * | 2008-10-07 | 2010-04-08 | Hui Luo | Diagnostic image processing with automatic self image quality validation |
US7729524B2 (en) * | 2005-04-15 | 2010-06-01 | Carestream Health, Inc. | Assessment of radiographic systems and operators using task-based phantom |
US20170039706A1 (en) * | 2014-03-04 | 2017-02-09 | The Trustees Of Columbia University In The City Of New York | Regularization of images |
US9916525B2 (en) * | 2015-10-13 | 2018-03-13 | Siemens Healthcare Gmbh | Learning-based framework for personalized image quality evaluation and optimization |
-
2018
- 2018-10-16 BR BR102018071293A patent/BR102018071293A2/pt not_active IP Right Cessation
-
2019
- 2019-10-16 WO PCT/BR2019/050444 patent/WO2020077426A1/pt active Application Filing
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7729524B2 (en) * | 2005-04-15 | 2010-06-01 | Carestream Health, Inc. | Assessment of radiographic systems and operators using task-based phantom |
US20100086182A1 (en) * | 2008-10-07 | 2010-04-08 | Hui Luo | Diagnostic image processing with automatic self image quality validation |
US20170039706A1 (en) * | 2014-03-04 | 2017-02-09 | The Trustees Of Columbia University In The City Of New York | Regularization of images |
US9916525B2 (en) * | 2015-10-13 | 2018-03-13 | Siemens Healthcare Gmbh | Learning-based framework for personalized image quality evaluation and optimization |
Non-Patent Citations (1)
Title |
---|
RAMOS, J. E. ET AL.: "Automation of the ACR MRI Low-Contrast Resolution Test Using Machine Learning", EM: 2018 11TH INTERNATIONAL CONGRESS ON IMAGE AND SIGNAL PROCESSING, BIOMEDICAL ENGINEERING AND INFORMATICS (CISP-BMEI, 13 October 2018 (2018-10-13), pages 1 - 6, XP033511679, DOI: 10.1109/CISP-BMEI.2018.8633140 * |
Also Published As
Publication number | Publication date |
---|---|
BR102018071293A2 (pt) | 2020-04-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Mortamet et al. | Automatic quality assessment in structural brain magnetic resonance imaging | |
US11116471B2 (en) | Methods and apparatus for extended low contrast detectability for radiographic imaging systems | |
Rose et al. | Efficient method for volumetric assessment of peak blood flow velocity using 4D flow MRI | |
Tucholka et al. | An empirical comparison of surface-based and volume-based group studies in neuroimaging | |
JP4469594B2 (ja) | 疾患関連組織変化を測定するシステム | |
JP4402436B2 (ja) | 気道測定のための方法及びシステム | |
CN106688014B (zh) | 用于确定放射摄影图像的图像质量的设备和方法 | |
CN111933273A (zh) | 基于人工智能的帕金森病诊断装置及方法 | |
Ding et al. | Reproducible automated breast density measure with no ionizing radiation using fat‐water decomposition MRI | |
Gedamu et al. | Automated quality control of brain MR images | |
Vos et al. | Multi-fiber tractography visualizations for diffusion MRI data | |
Hans et al. | Comparison of three-dimensional visualization techniques for depicting the scala vestibuli and scala tympani of the cochlea by using high-resolution MR imaging | |
George‐Jones et al. | Automated detection of vestibular schwannoma growth using a two‐dimensional U‐Net convolutional neural network | |
Gooijers et al. | White matter organization in relation to upper limb motor control in healthy subjects: exploring the added value of diffusion kurtosis imaging | |
Eck et al. | Prospective evaluation of repeatability and robustness of radiomic descriptors in healthy brain tissue regions in vivo across systematic variations in T2‐weighted magnetic resonance imaging acquisition parameters | |
Gibson et al. | 3D prostate histology reconstruction: An evaluation of image‐based and fiducial‐based algorithms | |
Crombé et al. | Assessment of repeatability, reproducibility, and performances of T2 mapping‐based radiomics features: a comparative study | |
Geng et al. | Automated MR image prescription of the liver using deep learning: Development, evaluation, and prospective implementation | |
KR102144720B1 (ko) | 기계 학습을 통해 분류된 나이그로좀 1 영역과 복수의 학습 모델에 의한 다중 예측 결과를 이용하는 인공지능 기반 파킨슨 병 진단을 위한 정보제공 방법 | |
KR102144719B1 (ko) | 복수의 학습 모델에 의한 다중 예측 결과를 이용하는 인공지능 기반 파킨슨 병 진단 장치 및 방법 | |
Soliman et al. | A realistic phantom for validating MRI‐based synthetic CT images of the human skull | |
Tan et al. | Denoising and Multiple Tissue Compartment Visualization of Multi‐b‐Valued Breast Diffusion MRI | |
Vaishnav et al. | CT metal artifact reduction algorithms: toward a framework for objective performance assessment | |
Digma et al. | Correcting B 0 inhomogeneity-induced distortions in whole-body diffusion MRI of bone | |
Pinto Pereira et al. | Automated registration of diagnostic to prediagnostic x‐ray mammograms: Evaluation and comparison to radiologists’ accuracy |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 19872983 Country of ref document: EP Kind code of ref document: A1 |
|
DPE1 | Request for preliminary examination filed after expiration of 19th month from priority date (pct application filed from 20040101) | ||
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 19872983 Country of ref document: EP Kind code of ref document: A1 |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 19872983 Country of ref document: EP Kind code of ref document: A1 |
|
32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS (EPO FORM 1205A DATED 11.11.2021) |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 19872983 Country of ref document: EP Kind code of ref document: A1 |