• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    The present invention discloses a system (10) and method for imaging based diagnosis. In the present invention, history (record) artifact images and corresponding measures for repairing the artifacts are obtained and stored in the database 12. A database 12 of historical artifact images and corresponding measures are used to diagnose the input artifact image 54 with known faults.
    公开号:KR19990029402A
    申请号:KR1019980035806
    申请日:1998-09-01
    公开日:1999-04-26
    发明作者:라지브 굽타；크리스토퍼 제임스 데일리；발티노 싸비에르 아폰소；래시크랄 펀자랄 사아
    申请人:제이 엘. 차스킨, 버나드 스나이더, 아더엠. 킹；제너럴 일렉트릭 캄파니；
    IPC主号:
    专利说明:

    System and method for performing imaging based diagnosis
    FIELD OF THE INVENTION The present invention relates generally to fault diagnostics, and more particularly to diagnosing faults in images generated by a malfunctioning imaging machine.
    In either industrial or commercial settings, a malfunctioning imaging machine can severely damage a business. Thus, a malfunctioning imaging machine needs to be repaired quickly and accurately. During a malfunction of an imaging machine, typically an ultrasound, computed tomography (CT), or magnetic resonance imaging (MRI) machine, a field engineer is called to diagnose and repair the machine. Typically, the field engineer looks for a log of events that occurred on the machine. The event log contains information such as the type of machine, the type of machine, and information about which customer. The event log also contains error logs of events that occurred during routine operations, as well as any artifact images from the machine and during any malfunctions. Using the accumulated experience in resolving machine malfunctions, the field engineer wants to thoroughly examine the error log and artifact images, looking for signs of failure. The field engineer then attempts to correct any problems that may cause machine malfunctions. If the error log contains a small amount of information and the artifact image generated is well known, this process will proceed as well. But if the error log contains a large amount of inaccurate information and you don't know the cause of the artifact image, it's usually the case for many complex devices. It is difficult for field engineers to diagnose faults quickly. Therefore, there is a need for a system and method for quickly diagnosing machine malfunctions from artifact images with complex error logs and associated unknown causes.
    According to a first embodiment of the present invention, a system for performing imaging based diagnosis is provided. In this embodiment, the database stores a plurality of historical images taken from a plurality of machines. The plurality of historical images includes a plurality of ideal images generated from the plurality of machines using all possible machine settings. Also, the plurality of history images includes a plurality of artifact images generated from the plurality of machines. Each artifact image has a corresponding corrective action and corresponding corrective action to repair the faults. The system also includes a diagnostic unit for diagnosing new artifact images from machines with unknown faults. The diagnostic unit includes a diagnostic image processor that includes means for finding an ideal image from a plurality of historical images that most closely match the new artifact image. The assigning means assigns an artifact category to the new artifact image based on the matched ideal image. The extracting means extracts the artifact feature from the new artifact image according to the assigned category. There is also a diagnostic fault isolator that includes means for generating a plurality of matrices for the extracted artifact features. The assigning means applies a plurality of matrices to identify artifact images from a plurality of history images that most closely match the new artifact image with corrective actions to repair the unknown fault.
    According to a second embodiment of the present invention, a method for performing imaging based diagnosis is provided. In this embodiment, a plurality of history images taken from a plurality of machines are obtained. The plurality of historical images includes a plurality of ideal images generated from the plurality of machines using all possible machine settings. Also, the plurality of history images includes a plurality of artifact images generated from the plurality of machines. Each artifact image knows the associated faults and the corresponding corrective action to repair the faults. At this point, a new artifact image is received from a machine with an unknown fault. An ideal image from a plurality of history images that most closely matches the new artifact image is found. Artifact categories are assigned to new artifact images based on the ideal image that most closely matches the new artifact image. At this time, a plurality of matrices are generated for the artifact category assigned to the new artifact image. A plurality of matrices are used to identify artifact images from a plurality of history images that most closely match the new artifact images.
    1 is a block diagram of imaging based diagnosis in accordance with the present invention;
    FIG. 2 is a flowchart for explaining image processing steps performed by the training unit shown in FIG.
    FIG. 3 is a flow chart illustrating fault isolation processing steps performed by the training unit shown in FIG. 1; FIG.
    4 is a flowchart for explaining image processing steps performed by the diagnostic unit shown in FIG.
    5 is a flow chart illustrating fault isolation processing steps performed by the diagnostic unit shown in FIG.
    The imaging based diagnostic system of the present invention is described in the context of medical imaging devices such as ultrasound, CT, or MRI machines. Although the present invention is described in the context of medical imaging equipment, an imaging based diagnostic system can be used in connection with any imaging device (microprocessor controlled chemical, mechanical, electronic imaging device) that produces an image. 1 shows a block diagram of an imaging based diagnostic system 10 according to the present invention. The image based diagnostic system 10 includes a database 12 of historical images, a training unit 14, and a diagnostic unit 16. The training unit 14 includes an image processor 18 and a fault isolator 20. The diagnostic unit 16 also includes an image processor 22 and a fault isolator 24. Both training unit 14 and diagnostic unit 16 are embedded in a computer, such as a workstation. However, other types of computers can be used, such as mainframes, minicomputers, microcomputers, supercomputers, and the like.
    The historical image stored in the database 12 includes a plurality of ideal images 26 of images generated from the plurality of imaging machines 28. A plurality of ideal images 26 of images are generated from the imaging machine using all possible probes and all possible machine default parameter settings. The model of the imaging machine, the probe used, the images taken, the parameter settings on the imaging machine are entered into the database 12 together with the ideal images remotely by a field engineer. Alternatively, the machine generating the image can be programmed to put this information in the image file itself, for example in the header of the image. As such, the information becomes an essential component of the database 12. In the case of images obtained remotely, variables such as model type, probe used, image used, etc., exist in the image itself and are automatically extracted later by the training unit 14. However, parameter settings are not discrete and it is possible to take on infinite combinations of successive values. Thus, it is treated differently from other variables. In the present invention, the number of machine settings is fixed in a finite set. For example, each ideal image from the field is annotated by a field engineer or technician onsite with the appropriate label defining the parameter settings of the imaging machine. Examples of parameter settings for the imaging machine are abdominal setting, chest setting, carotid artery setting.
    In addition to the plurality of ideal images 26, the database 12 receives a plurality of artifact images 30 generated from the plurality of imaging machines 28. Each artifact image 30 is the result of a known fault, such as unplugging the board and installing a fault board or the like. Like the ideal image 26, each of the artifact images 30 has variables that involve the model of the imaging machine, the probe used, the images captured, the parameter settings for the imaging machine, and the like. Again, variables such as model type, probe used, phase used, etc. are present in the image itself and automatically extracted, and parameter setting variables are fixed in a finite set and defined by a field engineer or technician. Alternatively, the machine generating the artifact image may be programmed to put this information in the image file itself, for example in the header of the image.
    In addition to the plurality of artifact images 30, the database 12 receives the keyboard logs 32 and the plurality of error logs generated from the imaging machine 28. The error log and the keyboard log each contain an event log of the imaging machine that occurs during routine operation and any malfunction. Error log and keyboard log show the operation signal of each video machine. For example, one of the error log and the keyboard log contains a sequence of events for an imaging machine with an unplugged board. Another error log and keyboard log contain the sequence of events for the imaging machine with the fault board installed. The plurality of error logs and keyboard logs 32 are stored in the database 12 and used as history cases for documenting software and hardware errors that occur in different imaging machines 28. The description of processing the history cases will be described later in more detail.
    After the plurality of artifact images 30 and the error log and the keyboard log 32 are input to the database 12, the artifact images are divided into a plurality of sets. In particular, the artifact image 30 is divided into M x P x F x S sets, where M is the number of imaging machines, P is the number of probes, F is the number of available phases, and S is the number of machine settings. Some machines can't handle all the probes or machine settings, so it's a set of bins. Dividing the artifact image 30 into sets makes it easier to find history matching for a new artifact image with an unknown fault.
    Historical images in the database 12 are accessed by the training unit 14 via the image processor 18. The image processor 18 processes the plurality of ideal images 26 into a plurality of artifact images 30. 2 shows a flow chart describing the image processing steps performed by the image processor 18. The image processing steps begin at 34 where a plurality of ideal images 26 and a plurality of artifact images 30 are retrieved from database 12. Each artifact image is matched with the corresponding ideal image at 36. The matching process makes the machine type, probe, and machine settings the same for artifacts and ideal images.
    The artifact image is registered at 38 with its corresponding ideal image 38 for matching. Normally images are obtained manually by placing a probe on an image from an imaging machine. There is some variation in the image between the image obtained as a result of the manual placement of the probe and the image obtained next. Registration is used to eliminate as many changes as possible. Any remaining unregistered after registration is considered later by the categorization step described later. A pixel-by-pixel comparison of the images obtained at different times by the original registration is possible. In the present invention, registration is achieved by matching artifact images to ideal images. At this time, the image processor 18 processes the areas covered with the reference markers to derive the two-dimensional point that can be matched with the corresponding ideal image. In particular, the image processor 18 takes the center of the reference marker and uses it to point-to-point to match the ideal image. Alternatively, it is possible to perform registration by wrapping the artifact image in the corresponding ideal image so that there is maximum correlation. Lapping is done through perspective, proximity, or rigid body deformation of one image to match another image.
    After registration, each ideal image is removed from the artifact image at 40. In the present invention, the ideal image is removed using a subtraction operation. The subtraction operation is performed pixel by pixel, so that the gray level of the ideal image pixel is taken from that of the artifact image. After this operation, since the final image contains a negative number, the subtracted image is renormalized so that the minimum pixel therein is zero. The subtraction operation results in a subtracted image containing only artifacts. Alternatively, filtering operations are applied to the two images before subtraction to account for any remaining misregistration between the ideal image and the artifact image.
    After subtraction, the artifact category is applied to each subtracted image at 42. In the present invention, the assigned artifact category is based on the eigen space representation of the subtracted artifact image. To determine the covariance matrix, each subtracted image is represented by a vector V of pixel values. For an n × m image, the first n values are n pixels in the first column of the image, the next n values are pixel values in the second column of the image, as described above. The predetermined set of N subtracted images is represented by {V 1 , V 2 ,... V N }. The average of all subtracted images is expressed as V avg . The covariance matrix is defined by the following equation.
    Where i, j '{1,2, ... N}, and Denotes a dot product.
    The covariance matrix is determined and then used to obtain the orthogonal representation and the image basis. Orthogonal representations and image bases are obtained by performing Singular Value Decomposition (SVD) on a covariance matrix. Alternatively, the Karhunen-Loeven Transform (KLT) can be used to determine orthogonal representations and image bases and involves diagonalization of the covariance matrix. For KLT, the covariance matrix is represented by Q, which is defined by the following equation.
    Q = UDT T (2)
    Where U and V are orthogonal to each other, and D is a diagonal matrix. The row of V defines the new image base. It is a characteristic of this new image foundation that the images in it are irrelevant. Other computationally less complex methods can be used to obtain orthogonal representations and image bases. For example, a discrete cosine transform (DCT) is available.
    The determined image basis is used to find a representation for each subtracted artifact image. In particular, each subtracted artifact image is represented as a linear combination of images in a new image base set. Thus, if B 1 , B 2 , .... B N , is based on N images, the historical artifact image I is expressed by the coefficient α 1 , α 2 , .... α N , such that Is characterized.
    I = α 1 B 1 + .... + α n B N (3)
    Where (α 1 , α 2 , ..... α N ) is a point in the N-dimensional space defined by {B 1 , B 2 , .... B N }. Each subtracted artifact image in the history database is represented by this one point. After finding a representation for each subtracted artifact image, the cluster of points that are most spatially close in this four-dimensional space is assigned as an artifact category. Some possible examples of the assigned artifact category are flash light artifacts, TD board artifacts, search light artifacts, and distortion artifacts. These examples illustrate, but are not limited to, examples of artifact categories of the type available in the present invention. For convenience of explanation, examples of the artifact category are shown in FIG. 1 as fault A, fault B, and fault C. FIG.
    After creating the category, image processor 18 extracts a set of artifact features for each artifact at 44. First, the artifact feature is extracted by converting the artifact image generated from the subtraction operation into a Fourier region. When the artifact image is converted into a Fourier region, it becomes a spectral signal of the artifact. Many category specific features that can be measured include image homogeneity, signal to noise ratio, modulation transfer function, resolution, distortion, signal attenuation, and texture characteristics. The invention is not limited to this category specific feature, other features may be measured if desired.
    2, after determining the artifact features for all artifact images, image processor 18 sends features to fault isolator 20 for further processing. 3 shows a flow chart describing the processing steps that fault isolator 20 performs. The fault isolator 20 first retrieves the error log and keyboard log 32 from the database at 46. The error log and keyboard log 32 then combine their corresponding artifact features at 48. The feature of each artifact quantified using various category specific matrices represents the syndrome associated with the actual fault. Error logs and keyboard logs also represent the syndrome associated with the actual fault. These three sources of information are used to create a case for the case based reasoning system. Each combined artifact feature and log set creates a history case at 50. History cases of artifact features and logs are stored in the database at 52 and later diagnostic unit 16 uses them to diagnose a situation of a new problem with a new artifact image generated from an imaging machine with an unknown fault.
    Returning to FIG. 1 again, diagnostic unit 16 receives a new artifact image 54 generated from imaging machine 56 that experiences an unknown fault. In addition, a new error log and keyboard log 58 of events occurring in the imaging machine are sent to the diagnostic unit 16. The new artifact image 54 and the new error log and keyboard log 58 are input from the image processor 22 to the diagnostic unit 16 by a field engineer or remote dial in connection. The image processor 22 processes the new artifact image 54 and the new log and keyboard log 56 with historical cases stored in the database 12.
    4 shows a flow chart setting that describes the image processing steps that image processor 22 performs. After obtaining a new artifact image, image processor 22 searches database 12 to find an ideal image that most closely matches the new artifact image at 60. The image processor then registers the ideal image at 62 with the new artifact image. As mentioned above, registration is accomplished by assigning a reference marker in the new artifact image to the ideal image by mapping the new artifact image, and processing the marker to derive a two-dimensional point that matches the ideal image. After registration, the ideal image is subtracted from the new artifact image at 64 using a subtraction or filtering operation. It is expressed by linear combination of subtracted image same basic set {B 1 , B 2 , .... B N } and is defined by Equation 4 as follows.
    I = β 1 B 1 + .... + β n B N (4)
    Here, the points {β 1 .... β n } represents another point in the space of the history artifact image. The distance from all artifact clusters to the advantage is used to determine which category the input image belongs to. In (66), a new artifact image deducted from the artifact category is allocated. After assigning the artifact category, image processor 22 extracts the artifact feature from the new artifact image deducted at 68, as described above.
    After determining the artifact feature for the new artifact image, the image processor 22 sends the feature to the fault isolator 24 for further processing. 5 shows a flow chart describing the processing steps performed by fault isolator 24. Fault isolator 24 uses the artifact feature extracted at 70 to generate a category specific matrix. Using the category specific matrix, the fault further illustrates the imaging machine 56 to generate the artifact image 54. Next, the error log and keyboard log 58 accompanying the new artifact image 54 are retrieved at 72. The fault isolator 24 then retrieves the history cases from the database 12 for the cases that most closely match the new artifact image at 74. A candidate set of images that most closely matches the new artifact image is generated at 76. Further, at 78, a corrective action is retrieved to repair the fault corresponding to each candidate set. One type of corrective action is to identify the field replaceable unit in the imaging machine 56 that needs to be replaced.
    The candidate set of images and the corresponding corrective action are presented to the field engineer, ranked in order of likelihood of matching a new artifact image at 80. The field engineer then thoroughly examines the candidate set in rank order at 82 to determine if the fault due to the new artifact is correctly identified. If the fault is correctly identified, fault isolator 24 logs a successful diagnosis at 84. On the other hand, if the fault is not correctly identified, it is determined at 86 if there are more candidate sets to evaluate. If there are more candidates, the next candidate is again evaluated at 88 and 82. This step continues until the fault is correctly identified. However, if no candidates correct the fault, a new artifact image 54 and error log and keyboard log 58 are sent to the training unit at 90 and added to the history case to diagnose future faults. As a result, as more cases are added to the cleaning unit 14, the level of accuracy of the imaging based diagnostic system is equalized and there is no need to add more cases to the training unit.
    Thus, there is provided a system and method according to the present invention for performing imaging based diagnosis that fully satisfies goals, advantages, and objectives as described above. While the invention has been described with reference to several embodiments, those skilled in the art will recognize that modifications and variations are possible without departing from the scope of the invention.
    权利要求:
    Claims (18)
    [1" claim-type="Currently amended] A system 10 for performing imaging based diagnosis of a machine 56,
    A database 12 containing a plurality of historical images taken by a plurality of machines 28, wherein the plurality of historical images are a plurality of ideal images 26 generated from the plurality of machines 28 using all possible machine settings. And a plurality of artifact images 30 generated from the plurality of machines 28, each artifact image 30 having a known fault associated therewith and a corresponding corrective action to repair the faults. Database,
    With a diagnostic unit 16 for diagnosing a new artifact image 54 from a machine 56 with an unknown fault, the diagnostic unit 16 is ideal for multiple history images that most closely match the new artifact image 54. Means for finding an image, means for assigning an artifact category to a new artifact image based on the matched ideal image 54, means for extracting artifact features from the new artifact image 54 according to the assigned category, and Means for generating a plurality of matrices of extracted artifact features, corrective measures for repairing unknown faults, and a plurality of matrices for identifying artifact images in the plurality of history images that most closely match the new artifact image 54. A diagnostic fault isolator 24 comprising means for applying a System for performing a diagnosis based on images of a machine comprising a diagnostic unit group.
    [2" claim-type="Currently amended] 2. The apparatus of claim 1, further comprising a training unit 14 coupled to the database 12 and the diagnostic unit 16, wherein the training unit 14 comprises a plurality of artifact images 30 and a plurality of ideal images. 26 means for obtaining, means for matching each of the plurality of artifact images 30 to a corresponding ideal image, means for assigning an artifact category to the matched image, and a matched image according to the assigned category. A system for performing imaging based diagnosis of a machine comprising a training image processor (18) comprising means for extracting artifact features.
    [3" claim-type="Currently amended] 3. The machine according to claim 2, wherein the training image processor 18 further comprises means for registering artifact images with unknown faults to corresponding ideal images and means for removing corresponding ideal images from the registered images. System for performing an imaging based diagnosis of a patient.
    [4" claim-type="Currently amended] 3. The imaging based diagnosis of a machine as recited in claim 2, wherein said training unit (14) further comprises a training fault isolator (20) coupled to said training image processor (18) to separate said extracted artifact feature from a history case. System for performing the operation.
    [5" claim-type="Currently amended] 5. The method of claim 4, wherein the database further comprises a plurality of error logs generated from a plurality of machines, each of the plurality of error logs performing an image based diagnosis of a machine containing data indicative of events occurring between operations of the machine. System for doing so.
    [6" claim-type="Currently amended] 6. The system of claim 5, wherein the training fault isolator couples the extracted artifact features and error logs to a history case.
    [7" claim-type="Currently amended] 2. The machine according to claim 1, wherein the diagnostic image processor 22 further comprises means for registering a new artifact image 54 in the matched ideal image and means for removing a corresponding ideal image from the registered image. System for performing imaging based diagnosis.
    [8" claim-type="Currently amended] 2. The machine of claim 1, wherein the diagnostic fault isolator comprises means for receiving an error log generated from a machine with an unknown fault, wherein the error log is an image of a machine containing data indicative of events occurring between operations of the machine. System for performing basic diagnostics.
    [9" claim-type="Currently amended] 9. The system of claim 8, wherein the diagnostic fault isolator uses an error log to generate a plurality of matrices.
    [10" claim-type="Currently amended] The system of claim 1, wherein the diagnostic unit further comprises means for adding a newly identified artifact image and a corresponding corrective action to the plurality of artifact images in the database.
    [11" claim-type="Currently amended] As a method for performing imaging based diagnosis of machine 56,
    In the step of obtaining a plurality of history images taken from the plurality of machines 28, the plurality of history images is a plurality of ideal images 26 and a plurality of machines generated from the plurality of machines 28 using all possible machine settings. A plurality of artifact images 30 generated from (28), each artifact image 30 having an associated fault associated therewith and a corresponding corrective action to repair the faults;
    Receiving a new artifact image 54 from a machine 56 with an unknown fault,
    Finding an ideal image from a plurality of history images that most closely match the new artifact image 54;
    Assigning an artifact category to the new artifact image 54 based on the ideal image that most closely matches the new artifact image 54;
    Generating a plurality of matrices of the artifact category assigned to the new artifact image 54;
    Performing an image based diagnosis of a machine comprising using a plurality of matrices for identifying artifact images in the plurality of historical images that most closely match the new artifact images 54 and corrective measures for repairing unknown faults. How to do it.
    [12" claim-type="Currently amended] The method of claim 11, wherein the obtaining of the plurality of history images comprises:
    Matching each of the plurality of artifact images 30 to a corresponding ideal image taken in the plurality of ideal images 26;
    Assigning an artifact category to each matching image;
    Extracting an artifact feature from each matched image.
    [13" claim-type="Currently amended] 13. The method of claim 12, further comprising the steps of: registering an artifact image with a known fault to a corresponding ideal image;
    Removing the corresponding ideal image from the registered image.
    [14" claim-type="Currently amended] 12. The method of claim 11, further comprising the step of determining an artifact feature of a new artifact image (54).
    [15" claim-type="Currently amended] 12. The method of claim 11, further comprising: registering a new artifact image 54 with a corresponding ideal image,
    And removing the corresponding ideal image from the registered image.
    [16" claim-type="Currently amended] 12. The method of claim 11, further comprising adding a newly identified artifact image and corresponding corrective action to the plurality of artifact images in the plurality of history images.
    [17" claim-type="Currently amended] 12. The method of claim 11, further comprising the step of inputting an error log from a machine with an unknown fault, wherein the error log performs an image based diagnosis of the machine containing data indicative of events occurring between operations of the machine. How to.
    [18" claim-type="Currently amended] 18. The method according to claim 17, wherein the input error log is used to generate a plurality of matrices of new artifact images.
    类似技术:
    公开号 | 公开日 | 专利标题
    Knoll et al.2019|Assessment of the generalization of learned image reconstruction and the potential for transfer learning
    Sheikh et al.2005|An information fidelity criterion for image quality assessment using natural scene statistics
    Buades et al.2009|Self-similarity driven color demosaicking
    US5463768A|1995-10-31|Method and system for analyzing error logs for diagnostics
    Thévenaz et al.2000|Optimization of mutual information for multiresolution image registration
    DE102012208537B4|2019-03-28|A method of identifying a root cause of an error in a serviced vehicle and performing corrective actions
    US6343236B1|2002-01-29|Method and system for analyzing fault log data for diagnostics
    EP1495384B1|2006-02-22|Method and apparatus for improving fault classifications
    US8055095B2|2011-11-08|Parallel and adaptive signal processing
    EP0887733B1|2003-09-03|Model-based diagnostic system with automated procedures for next test selection
    US5748508A|1998-05-05|Method and device for signal analysis, process identification and monitoring of a technical process
    US7382507B2|2008-06-03|Image quality defect detection from image quality database
    CN100554980C|2009-10-28|Be used for supporting the method for the fault functional unit of recognology device
    US5495540A|1996-02-27|Automatic subarea selection for image registration
    US7650272B2|2010-01-19|Evaluation of Bayesian network models for decision support
    JP4852043B2|2012-01-11|System, device, and method for updating system monitoring model
    JP4361634B2|2009-11-11|System and method for integrating multiple diagnostic related information
    CN100394444C|2008-06-11|Image processing apparatus, image processing method, image processing system
    EP1117044B1|2003-08-20|System model determination for failure detection and isolation in particular in computer systems
    EP0304848A2|1989-03-01|System for diagnosing defects in electronic assemblies
    Varma1999|ICARUS: design and deployment of a case-based reasoning system for locomotive diagnostics
    US20040078171A1|2004-04-22|Diagnostic systems and methods for predictive condition monitoring
    US8527441B2|2013-09-03|Developing fault model from service procedures
    US6343261B1|2002-01-29|Apparatus and method for automatically diagnosing a technical system with efficient storage and processing of information concerning steps taken
    US20150018664A1|2015-01-15|Assessment of Traumatic Brain Injury
    同族专利:
    公开号 | 公开日
    NL1009997C2|2006-06-16|
    JPH11161774A|1999-06-18|
    US6115489A|2000-09-05|
    NL1009997A1|1999-03-04|
    KR100617885B1|2006-12-01|
    DE19829640A1|1999-03-04|
    AU755077B2|2002-12-05|
    JP4179676B2|2008-11-12|
    AU8189598A|1999-03-18|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    法律状态:
    1997-09-02|Priority to US08/921,959
    1997-09-02|Priority to US08/921,959
    1997-09-02|Priority to US8/921,959
    1998-09-01|Application filed by 제이 엘. 차스킨, 버나드 스나이더, 아더엠. 킹, 제너럴 일렉트릭 캄파니
    1999-04-26|Publication of KR19990029402A
    2006-12-01|Application granted
    2006-12-01|Publication of KR100617885B1
    优先权:
    申请号 | 申请日 | 专利标题
    US08/921,959|1997-09-02|
    US08/921,959|US6115489A|1997-09-02|1997-09-02|System and method for performing image-based diagnosis|
    US8/921,959|1997-09-02|
    [返回顶部]