WO2006123272A2 - Automatic organ mapping method an device - Google Patents
Automatic organ mapping method an device Download PDFInfo
- Publication number
- WO2006123272A2 WO2006123272A2 PCT/IB2006/051448 IB2006051448W WO2006123272A2 WO 2006123272 A2 WO2006123272 A2 WO 2006123272A2 IB 2006051448 W IB2006051448 W IB 2006051448W WO 2006123272 A2 WO2006123272 A2 WO 2006123272A2
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- organ
- image
- organs
- rules
- data
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- 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
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20092—Interactive image processing based on input by user
- G06T2207/20101—Interactive definition of point of interest, landmark or seed
-
- 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
- G06T2207/30061—Lung
Definitions
- the present invention is in the medical field and relates to a method for automatically segmenting organs apparent from a body image.
- the invention is particularly adapted to images obtained by Computed Tomography and Magnetic Resonance Imaging (MRI) and nuclear medicine imaging techniques such as Positron Emission Tomography (PET).
- MRI Computed Tomography
- PET Positron Emission Tomography
- Recent medical technologies have changed practitioners daily work and have given them the tools that facilitate the acquisition of clinical data.
- digital imaging now provides an extraordinary amount of potentially useful data that serves as a basis for diagnosis.
- the amount of available data is such that practitioners often lack the time to thoroughly analyze the full extent of the available images.
- ancillary digital processing has supplanted the practitioner for the performance of dedicated tasks.
- Computer aided segmentation and recognition software is indeed capable of doing a first level analysis, which is often limited to the application of conventional anatomical rules and mathematical algorithms that practitioners would traditionally use. Consequently, practitioners' interventions can be kept to a minimum and focused on the diagnosis itself which appeals to necessary human medical knowledge, experience and judgment.
- segmentation generally involves extracting relevant anatomical objects, e.g. organs, tissues, or regions of interest from images for the purpose of anatomical identification, diagnosis or measurement.
- Segmentation serves as an essential first stage for other tasks such as registration, comparison and detection. These other tasks may be carried out using additional algorithms that further help practitioners to locate and identify abnormalities, e.g. nodules, arteries obstruction, organs deformation. Semi-automatic or fully automatic algorithms allow clinicians to skip healthy body tissues or organs identified as such by the computer algorithms and to focus on the most suspicious objects or regions.
- a possible automatic segmentation method is proposed by Nico Karssemeijer in the paper "A statistical method for automatic labeling of tissues in medical images” published in 1990 In Machine and Vision Applications, Springer- Verlag New York Inc.
- This document discloses a statistically based segmentation method applied to the recognition of organs in abdominal CT scans. The method incorporates prior knowledge of anatomical structure and to do that, it uses a stochastic model that represents abdominal geometry in 3D. A probabilistic model is generated that incorporates the properties of each tissue class with respect to its mean grey- value distribution and geometric position. The tissues are then automatically segmented using a global optimization scheme, trying to match this model with a CT image.
- a problem with the above documents and existing automatic processing algorithms is that they require the intervention of a clinician to be initiated. Indeed, most segmentation algorithms require the input of one or more seeding points or segments which will constitute the starting point for the delineation of the organs.
- An additional problem with the above segmentation method is that it uses a stochastic model of the image, which could not be robust due to large patients variability.
- An object of the invention is therefore to simplify the clinician's tasks and provide an initial fully automatic organs segmentation of a raw image so that automatic advanced processing can be ultimately performed.
- Another object of the invention is to generate a generic mapping of objects apparent from an image without using any prior stochastic model of the body.
- a method is therefore proposed to map organs on a body image.
- the method comprises a first step of determining a body part present on the body image obtained from image data. Then it a set of preset rules associated with the body part and from there, the method further extracts appearance data associated with a first organ belonging to the body part.
- the appearance data is specific to a type of the body image.
- the image data is automatically analyzed and the first organ is then located on the body image based on the retrieved appearance data.
- the method retrieves anatomical rules from the preset rules including hierarchical organs positioning relative to the first organ and further analyzes the image data and locates a second organ and other organs on the body image in a vicinity of the first organ based on the anatomical rules.
- an automatic segmentation algorithm runs on at least one of the identified organs and results in the delineation of a segmented organ.
- the segmented organ is now ready for further automatic processing by an ancillary processing algorithm.
- the above method is run on raw images obtained from, e.g., patients' full body, thorax or abdominal scans.
- the above steps permit to automatically draw a map of objects that are clearly apparent from the picture and thus prepare the image for further processing later on, e.g. advanced segmentation or volumetric measurement.
- a clinician traditionally first visualizes the image, locates the lungs on the image and enters image points or a specific region of interest and from there, the measurement application semi-automatically or automatically delineates the lungs and calculates the volume.
- the thorax scan is automatically recognized and the automatic algorithm localizes the lungs without clinician's input.
- Organs delineation is subsequently carried out based on physical rules that can include normal organs relative positioning or abnormal organs relative positioning due to traumas for example. Processing of the image can be done immediately after the image is taken without having to wait for the clinician's presence. Ideally the organ mapping algorithm of the invention is run systematically when the image is registered.
- Fig. 1 is a flowchart diagram of a method of the invention.
- a patient's thoracic image is obtained by Computed Tomography using a conventional CT device.
- Several objects e.g. organs, arteries, tissues, etc... are present on the image and an objective of a mapping algorithm of the invention is to individually delineate the objects so that further processing can be carried out on the individual objects.
- an objective of a mapping algorithm of the invention is to individually delineate the objects so that further processing can be carried out on the individual objects.
- the invention is illustrated with a body image obtained by computed tomography, a similar processing algorithm applying the principles of the invention also runs on images obtained using other well-known imaging technologies such as MRI.
- Fig.1 shows a flowchart diagram 500 that illustrates an exemplary hierarchical processing algorithm of the invention.
- a raw image such as image 100 serves as the algorithm input.
- the image 100 would be partitioned into sub-areas or else, the clinician would select a region of interest that contains the suspected pathology. Segmentation would then be carried out starting from seeding point or segments manually entered.
- Diagram 500 includes a first processing stage 200 depicted by box 200.
- Stage 200 includes first step 210 that includes determining an initial organ or object 110 on image 100.
- image 100 is a CT patient's thoracic scan.
- the initial object 110 may be identified on image 100 based on preset rules.
- the same initial organ may be set by default so that the processing algorithm always looks for the same initial organ for a given type of body image.
- the initial organ 110 for a thoracic scan may be set to the lungs which often is the biggest and most recognizable organ.
- Rules are set to enable a quick and reliable localization of the predefined initial organ 110, e.g. the lungs in case of CT images. These rules may includes the general shape of the initial object, and models of the initial object may be made available to the algorithm for comparison with image 100.
- Models encompass normal shapes of healthy organs and abnormal shapes for injured or unhealthy organs which are for example bigger than usual or which suffered partial ablation.
- the above rules may additionally contain gray levels, position of the initial object on the image and/or a combination of the above. For example, the level of gray of the lungs in a CT thoracic scan is quite predictable.
- the patient's lungs are filled with air and are thus represented by dark pixels ranging from -1000 to - ⁇ 00 HU.
- the lungs will appear darker than other organs or tissues and efforts to localize the lungs can therefore by focused by rule in the darker regions. Confusion may however arise between the lungs filled with air and the outside of the body also represented by darker regions on image 100. Additional rules may then be set to avoid the confusion. For example, the search for the initial object 110 will automatically skip regions in the image borders.
- the lungs are possibly found by setting threshold on the pixel values and restricting the search to a central area.
- the respective sizes of objects fulfilling the primary criteria of the lungs, or any other predefined initial object 110 are compared with the size of the lungs typically observed for the patient's health condition, gender and age.
- additional algorithm modules are run in stages 220 and 230 in which a set of second hierarchical level of organs is identified.
- This second set of organs is extracted from image 100 based on predefined anatomical rules such as relative organ positioning, i.e. below, above, etc....
- the anatomical rules may first include usual spatial localization and by consequence, organs are searched in a priori locations on image 100.
- algorithm module 220 seeks to localize the liver and operates on the known rules that the liver is located below the right lung with a possible extension on the left side.
- Algorithm module 230 proceeds on the basis of similar rules in order to find the heart, spleen and/or kidneys respectively.
- the processing algorithm of the invention may further include additional modules dedicated to other organs.
- a three dimensional active object segmentation is done on each identified organ.
- a first 3DAO segmentation algorithm 310 is run on the identified lungs 110 and whose shape was coarsely determined in the previous stage.
- Known automatic organ fine segmentation may be used. Reference is made to the published paper "Efficient Model-based quantification of left ventricle function in 3D echography", Olivier Gerard, Antoine Collet Billon, Jean-Michel Rouet, Miarie Jacob, Maxim Fradkin, Cyril Allouche, IEEE Transactions on Medical Imaging, September 2002, Volume 21 , Number 9.
- a finer segmentation is possibly carried out to determine the fine inner structure of certain organs.
- a finer segmentation is done to determine the internal structure of the lungs 110.
- the finer segmentation may include module 410 that is specific to lobes airways delineation and a module 420 specific to the determination of pulmonary arteries. Finer segmentation may be done for selected organs of interest only.
- further processing is run on the image. Further processing may be performed using any commercially available dedicated medical applications such as lung nodule finder applications, polyp detectors, measurements calculators, and similar pathology or organs specific applications.
- the invention is concerned with a fully automatic and hierarchical mapping algorithm based on a priori knowledge of the human body and the patient's history and the goal of the mapping algorithm of the invention is to generate data that one or more further commercially available applications can use for measurement and detection purposes.
- the invention also pertains to a device for carrying out the above described algorithm.
- a device includes an image acquisition unit that acquires the image either directly from the patient's body in which case, the device includes CT unit or the like.
- the device may also receive the image from an external imaging unit and the image acquisition unit merely registers the image. Input data respecting the patient general health condition, past traumas, age and the like may also be provided to the device and may control the selection or the extraction of the preset rules for a given patient.
- the device further includes a storage unit for storing all preset rules and all physical or anatomical rules mentioned above that are representative of or associated with the type of image 100, the initial object 110 and the anatomical localization of the organs. Storage unit may also comprise data that is enough to generate the preset rules instead of storing all preset rules representative of all situations.
- the device further includes processing unit and a segmentation algorithm for carrying out the algorithm as described above in reference to Fig.l. Preferably, the device performs the mapping of the invention while the image is registered in the digital database of patients' images so that the image is pre-processed when further processing is needed.
- the above exemplary embodiment proposes to carry out the delineation of the individual organs in parallel to the organs localization.
- the invention also encompasses implementations where the organs are first localized throughout the image and in a subsequent step, the organs are individually delineated.
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Image Processing (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The present invention relates to a device for automatically segmenting and preparing an image for further processing. The device automatically maps the organs present on the picture according to preset rules that are representative of conventional physical and anatomical knowledge. The rules include the definition of an initial organ that the device needs to look for first and then, once the device has localized the initial organ, the device follows anatomical rules and localizes further organs. Ultimately, automatic segmentation is done on one or more of the organs that have been localized. The image is then pre-processed for potential further processing.
Description
Automatic organ mapping method an device
The present invention is in the medical field and relates to a method for automatically segmenting organs apparent from a body image. The invention is particularly adapted to images obtained by Computed Tomography and Magnetic Resonance Imaging (MRI) and nuclear medicine imaging techniques such as Positron Emission Tomography (PET). Recent medical technologies have changed practitioners daily work and have given them the tools that facilitate the acquisition of clinical data. In particular, digital imaging now provides an extraordinary amount of potentially useful data that serves as a basis for diagnosis. However, the amount of available data is such that practitioners often lack the time to thoroughly analyze the full extent of the available images. To overcome this problem, ancillary digital processing has supplanted the practitioner for the performance of dedicated tasks. Computer aided segmentation and recognition software is indeed capable of doing a first level analysis, which is often limited to the application of conventional anatomical rules and mathematical algorithms that practitioners would traditionally use. Consequently, practitioners' interventions can be kept to a minimum and focused on the diagnosis itself which appeals to necessary human medical knowledge, experience and judgment.
For example, a wide range of organ segmentation software is based on various optimization methods, e.g. cost function minimization and the like and all permit, with varying results, organ delineation with little clinician's involvement. Segmentation generally involves extracting relevant anatomical objects, e.g. organs, tissues, or regions of interest from images for the purpose of anatomical identification, diagnosis or measurement.
Segmentation serves as an essential first stage for other tasks such as registration, comparison and detection. These other tasks may be carried out using additional algorithms that further help practitioners to locate and identify abnormalities, e.g. nodules, arteries obstruction, organs deformation. Semi-automatic or fully automatic algorithms allow clinicians to skip healthy body tissues or organs
identified as such by the computer algorithms and to focus on the most suspicious objects or regions.
A possible automatic segmentation method is proposed by Nico Karssemeijer in the paper "A statistical method for automatic labeling of tissues in medical images" published in 1990 In Machine and Vision Applications, Springer- Verlag New York Inc. This document discloses a statistically based segmentation method applied to the recognition of organs in abdominal CT scans. The method incorporates prior knowledge of anatomical structure and to do that, it uses a stochastic model that represents abdominal geometry in 3D. A probabilistic model is generated that incorporates the properties of each tissue class with respect to its mean grey- value distribution and geometric position. The tissues are then automatically segmented using a global optimization scheme, trying to match this model with a CT image.
A problem with the above documents and existing automatic processing algorithms is that they require the intervention of a clinician to be initiated. Indeed, most segmentation algorithms require the input of one or more seeding points or segments which will constitute the starting point for the delineation of the organs. An additional problem with the above segmentation method is that it uses a stochastic model of the image, which could not be robust due to large patients variability.
An object of the invention is therefore to simplify the clinician's tasks and provide an initial fully automatic organs segmentation of a raw image so that automatic advanced processing can be ultimately performed.
Another object of the invention is to generate a generic mapping of objects apparent from an image without using any prior stochastic model of the body.
These and other aspects of the invention will be apparent from and will be elucidated with reference to the embodiments described hereinafter.
A method is therefore proposed to map organs on a body image. The method comprises a first step of determining a body part present on the body image obtained from image data. Then it a set of preset rules associated with the body part and from there, the method further extracts appearance data associated with a first organ belonging to the body part. The appearance data is specific to a type of the body image. The image data is automatically analyzed and the first organ is then located on the body image based on the retrieved appearance data. The method then retrieves
anatomical rules from the preset rules including hierarchical organs positioning relative to the first organ and further analyzes the image data and locates a second organ and other organs on the body image in a vicinity of the first organ based on the anatomical rules. In a subsequent step, an automatic segmentation algorithm runs on at least one of the identified organs and results in the delineation of a segmented organ. The segmented organ is now ready for further automatic processing by an ancillary processing algorithm.
The above method is run on raw images obtained from, e.g., patients' full body, thorax or abdominal scans. The above steps permit to automatically draw a map of objects that are clearly apparent from the picture and thus prepare the image for further processing later on, e.g. advanced segmentation or volumetric measurement. Thus, for example, in a case where pulmonary volumetric measurements is needed, a clinician traditionally first visualizes the image, locates the lungs on the image and enters image points or a specific region of interest and from there, the measurement application semi-automatically or automatically delineates the lungs and calculates the volume. In the invention, the thorax scan is automatically recognized and the automatic algorithm localizes the lungs without clinician's input. Organs delineation is subsequently carried out based on physical rules that can include normal organs relative positioning or abnormal organs relative positioning due to traumas for example. Processing of the image can be done immediately after the image is taken without having to wait for the clinician's presence. Ideally the organ mapping algorithm of the invention is run systematically when the image is registered.
The present invention will now be described in more detail, by way of example, with reference to the accompanying drawing, wherein: Fig. 1 is a flowchart diagram of a method of the invention.
In an exemplary embodiment of the invention, a patient's thoracic image is obtained by Computed Tomography using a conventional CT device. Several objects, e.g. organs, arteries, tissues, etc... are present on the image and an objective of a mapping algorithm of the invention is to individually delineate the objects so that further processing can be carried out on the individual objects. Although the invention is illustrated with a body image obtained by computed tomography, a
similar processing algorithm applying the principles of the invention also runs on images obtained using other well-known imaging technologies such as MRI.
Fig.1 shows a flowchart diagram 500 that illustrates an exemplary hierarchical processing algorithm of the invention. Initially, a raw image such as image 100 serves as the algorithm input. Conventionally the image 100 would be partitioned into sub-areas or else, the clinician would select a region of interest that contains the suspected pathology. Segmentation would then be carried out starting from seeding point or segments manually entered. Diagram 500 includes a first processing stage 200 depicted by box 200. Stage 200 includes first step 210 that includes determining an initial organ or object 110 on image 100. As mentioned above, image 100 is a CT patient's thoracic scan. The initial object 110 may be identified on image 100 based on preset rules. First, for a given type of body image, that is a thoracic scan, a brain scan, an abdominal scan or a whole body scan or any other body part image, the same initial organ may be set by default so that the processing algorithm always looks for the same initial organ for a given type of body image. Thus, the initial organ 110 for a thoracic scan may be set to the lungs which often is the biggest and most recognizable organ. Rules are set to enable a quick and reliable localization of the predefined initial organ 110, e.g. the lungs in case of CT images. These rules may includes the general shape of the initial object, and models of the initial object may be made available to the algorithm for comparison with image 100. Models encompass normal shapes of healthy organs and abnormal shapes for injured or unhealthy organs which are for example bigger than usual or which suffered partial ablation. The above rules may additionally contain gray levels, position of the initial object on the image and/or a combination of the above. For example, the level of gray of the lungs in a CT thoracic scan is quite predictable.
Indeed, the patient's lungs are filled with air and are thus represented by dark pixels ranging from -1000 to -^00 HU. The lungs will appear darker than other organs or tissues and efforts to localize the lungs can therefore by focused by rule in the darker regions. Confusion may however arise between the lungs filled with air and the outside of the body also represented by darker regions on image 100. Additional rules may then be set to avoid the confusion. For example, the search for the initial object 110 will automatically skip regions in the image borders. The lungs are
possibly found by setting threshold on the pixel values and restricting the search to a central area. In addition, the respective sizes of objects fulfilling the primary criteria of the lungs, or any other predefined initial object 110, are compared with the size of the lungs typically observed for the patient's health condition, gender and age. Once the initial object 110 is roughly localized and segmented on image 100, additional algorithm modules are run in stages 220 and 230 in which a set of second hierarchical level of organs is identified. This second set of organs is extracted from image 100 based on predefined anatomical rules such as relative organ positioning, i.e. below, above, etc.... The anatomical rules may first include usual spatial localization and by consequence, organs are searched in a priori locations on image 100. For example, algorithm module 220 seeks to localize the liver and operates on the known rules that the liver is located below the right lung with a possible extension on the left side. Algorithm module 230 proceeds on the basis of similar rules in order to find the heart, spleen and/or kidneys respectively. The processing algorithm of the invention may further include additional modules dedicated to other organs.
Then, in a subsequent stage 300 and once the organs have been roughly positioned on picture 100, a three dimensional active object segmentation is done on each identified organ. A first 3DAO segmentation algorithm 310 is run on the identified lungs 110 and whose shape was coarsely determined in the previous stage. Known automatic organ fine segmentation may be used. Reference is made to the published paper "Efficient Model-based quantification of left ventricle function in 3D echography", Olivier Gerard, Antoine Collet Billon, Jean-Michel Rouet, Miarie Jacob, Maxim Fradkin, Cyril Allouche, IEEE Transactions on Medical Imaging, September 2002, Volume 21 , Number 9.
Then in subsequent stage 400, a finer segmentation is possibly carried out to determine the fine inner structure of certain organs. For example, a finer segmentation is done to determine the internal structure of the lungs 110. The finer segmentation may include module 410 that is specific to lobes airways delineation and a module 420 specific to the determination of pulmonary arteries. Finer segmentation may be done for selected organs of interest only.
Ultimately, once a map of the organs present on image 100 has been built, further processing is run on the image. Further processing may be performed using any commercially available dedicated medical applications such as lung nodule finder applications, polyp detectors, measurements calculators, and similar pathology or organs specific applications.
The foregoing merely illustrates the principles of the invention. It will thus be appreciated that those skilled in the art will be able to devise various arrangements which, although not explicitly described or shown herein, embody the principles of the invention and are thus within the spirit and scope of the following claims. It must be noted that the invention is concerned with a fully automatic and hierarchical mapping algorithm based on a priori knowledge of the human body and the patient's history and the goal of the mapping algorithm of the invention is to generate data that one or more further commercially available applications can use for measurement and detection purposes. The invention also pertains to a device for carrying out the above described algorithm. Such a device includes an image acquisition unit that acquires the image either directly from the patient's body in which case, the device includes CT unit or the like. The device may also receive the image from an external imaging unit and the image acquisition unit merely registers the image. Input data respecting the patient general health condition, past traumas, age and the like may also be provided to the device and may control the selection or the extraction of the preset rules for a given patient. The device further includes a storage unit for storing all preset rules and all physical or anatomical rules mentioned above that are representative of or associated with the type of image 100, the initial object 110 and the anatomical localization of the organs. Storage unit may also comprise data that is enough to generate the preset rules instead of storing all preset rules representative of all situations. The device further includes processing unit and a segmentation algorithm for carrying out the algorithm as described above in reference to Fig.l. Preferably, the device performs the mapping of the invention while the image is registered in the digital database of patients' images so that the image is pre-processed when further processing is needed.
Also, the above exemplary embodiment proposes to carry out the delineation of the individual organs in parallel to the organs localization. However, the invention
also encompasses implementations where the organs are first localized throughout the image and in a subsequent step, the organs are individually delineated. In interpreting these claims, it should be understood that: a) the word "comprising" does not exclude the presence of other elements or acts than those listed in a given claim; b) the word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements; c) any reference signs in the claims do not limit their scope; d) several "means" may be represented by the same item or hardware or software implemented structure or function; e) each of the disclosed elements may be comprised of hardware portions (e.g., including discrete and integrated electronic circuitry), software portions (e.g., computer programming), and any combination thereof; f) hardware portions may be comprised of one or both of analog and digital portions; g) any of the disclosed devices or portions thereof may be combined together or separated into further portions unless specifically stated otherwise; and h) no specific sequence of acts is intended to be required unless specifically indicated.
Claims
1. A method for mapping organs on a body image, the method comprising:
- determining a body part present on the body image obtained from image data;
characterized in that the method further comprises:
- from a set of preset rules associated with the body part, extracting appearance data associated with a first organ belonging to the body part, the appearance data being further specific to a type of the body image;
- automatically analyzing the image data and locating the first organ on the body image based on the retrieved appearance data;
- retrieving anatomical rules from the preset rules including hierarchical organs positioning relative to the first organ;
- further analyzing the image data and locating a second organ and other organs on the body image in a vicinity of the first organ based on the anatomical rules; and
- running an automatic segmentation algorithm of at least one of the identified organs delineating a segmented organ, the segmented organ being delineated for further automatic processing by an ancillary processing algorithm.
2. The method of Claim 1, characterized in that the retrieving steps are further carried out based on personal data associated with the patient whose body part is imaged.
3. The method of Claim 1, characterized in that the body image is of one of the following types: Computed Tomography or Magnetic Resonance Imaging.
4. The method of Claim 1, characterized in that the first organ is localized in a central region of the image.
5. The method of Claim 1, characterized in that the appearance data includes pixel level threshold defining the range of pixels levels of the first organ.
6. The method of Claim 1, characterized in that the anatomical rules include spatial positioning of the second organ relative to the first organ.
7. The method of Claim 1, characterized in that the physical data includes a spatial positioning of the second organ with respect to the first organ and the locating the second organ includes automatically identifying a point on the body image that belongs to the second organ and running a segmentation algorithm to draw a contour of the second organ.
8. The method of Claim 1, characterized in that it comprises an additional step of automatically segmenting the first organ using appropriate segmentation method once the first organ is retrieved and before the further analysis of the image.
9. A device for carrying a method as claimed in Claim 1.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP05300385 | 2005-05-18 | ||
EP05300385.1 | 2005-05-18 |
Publications (2)
Publication Number | Publication Date |
---|---|
WO2006123272A2 true WO2006123272A2 (en) | 2006-11-23 |
WO2006123272A3 WO2006123272A3 (en) | 2007-05-03 |
Family
ID=37431642
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/IB2006/051448 WO2006123272A2 (en) | 2005-05-18 | 2006-05-09 | Automatic organ mapping method an device |
Country Status (1)
Country | Link |
---|---|
WO (1) | WO2006123272A2 (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111819599A (en) * | 2018-01-10 | 2020-10-23 | 消化器官癌症研究所 | A process for automatically segmenting a 3D medical image by one or more neural networks via structured convolution in accordance with the anatomical geometry of the 3D medical image |
US20220198670A1 (en) * | 2020-12-21 | 2022-06-23 | Siemens Healthcare Gmbh | Method and system for automated segmentation of biological object parts in mri |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6430430B1 (en) * | 1999-04-29 | 2002-08-06 | University Of South Florida | Method and system for knowledge guided hyperintensity detection and volumetric measurement |
EP1465109A2 (en) * | 2002-12-10 | 2004-10-06 | Eastman Kodak Company | Method for automated analysis of digital chest radiographs |
-
2006
- 2006-05-09 WO PCT/IB2006/051448 patent/WO2006123272A2/en active Application Filing
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6430430B1 (en) * | 1999-04-29 | 2002-08-06 | University Of South Florida | Method and system for knowledge guided hyperintensity detection and volumetric measurement |
EP1465109A2 (en) * | 2002-12-10 | 2004-10-06 | Eastman Kodak Company | Method for automated analysis of digital chest radiographs |
Non-Patent Citations (4)
Title |
---|
BRAM VAN GINNEKEN ET AL: "Computer-Aided Diagnosis in Chest Radiography: A Survey" IEEE TRANSACTIONS ON MEDICAL IMAGING, IEEE SERVICE CENTER, PISCATAWAY, NJ, US, vol. 20, no. 12, December 2001 (2001-12), XP011036184 ISSN: 0278-0062 * |
CAMARA O ET AL: "Computational modeling of thoracic and abdominal anatomy using spatial relationships for image segmentation" REAL-TIME IMAGING, ACADEMIC PRESS LIMITED, GB, vol. 10, no. 4, August 2004 (2004-08), pages 263-273, XP004609104 ISSN: 1077-2014 * |
MATTHEW S. BROWN, LAURENCE S. WILSON, BRUCE D. DOUST, ROBERT W. GILL, CHANGMING SUN: "Knowledge-based method for segmentation and analysis of lung boundaries in chest X-ray images" COMPUTERIZED MEDICAL IMAGING AND GRAPHICS, vol. 22, no. 6, 1998, pages 463-477, XP002417705 * |
PETAJAN E ET AL: "Robust face feature analysis for automatic speechreading and character animation" AUTOMATIC FACE AND GESTURE RECOGNITION, 1996., PROCEEDINGS OF THE SECOND INTERNATIONAL CONFERENCE ON KILLINGTON, VT, USA 14-16 OCT. 1996, LOS ALAMITOS, CA, USA,IEEE COMPUT. SOC, US, 14 October 1996 (1996-10-14), pages 357-362, XP010200447 ISBN: 0-8186-7713-9 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111819599A (en) * | 2018-01-10 | 2020-10-23 | 消化器官癌症研究所 | A process for automatically segmenting a 3D medical image by one or more neural networks via structured convolution in accordance with the anatomical geometry of the 3D medical image |
US20220198670A1 (en) * | 2020-12-21 | 2022-06-23 | Siemens Healthcare Gmbh | Method and system for automated segmentation of biological object parts in mri |
Also Published As
Publication number | Publication date |
---|---|
WO2006123272A3 (en) | 2007-05-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10964020B2 (en) | Similar case image search program, similar case image search apparatus, and similar case image search method | |
US7876938B2 (en) | System and method for whole body landmark detection, segmentation and change quantification in digital images | |
Graf et al. | 2d image registration in ct images using radial image descriptors | |
US7336809B2 (en) | Segmentation in medical images | |
EP2916738B1 (en) | Lung, lobe, and fissure imaging systems and methods | |
US8588495B2 (en) | Systems and methods for computer aided diagnosis and decision support in whole-body imaging | |
US7382907B2 (en) | Segmenting occluded anatomical structures in medical images | |
JP4874701B2 (en) | Similar image retrieval apparatus and method, and program | |
EP2598033B1 (en) | Knowledge-based automatic image segmentation | |
US20110158491A1 (en) | Method and system for lesion segmentation | |
US20060025671A1 (en) | Image display apparatus, image display method and the program | |
US20110007954A1 (en) | Method and System for Database-Guided Lesion Detection and Assessment | |
JP4640845B2 (en) | Image processing apparatus and program thereof | |
US20220020152A1 (en) | Apparatus for identifying regions in a brain image | |
KR102349515B1 (en) | Tumor automatic segmentation based on deep learning in a medical image | |
US20190012805A1 (en) | Automatic detection of an artifact in patient image data | |
EP2689344B1 (en) | Knowledge-based automatic image segmentation | |
US7809174B2 (en) | Method and system for segmentation of computed tomography image data | |
US20060078184A1 (en) | Intelligent splitting of volume data | |
US9361701B2 (en) | Method and system for binary and quasi-binary atlas-based auto-contouring of volume sets in medical images | |
Liu et al. | Automatic segmentation of the human brain ventricles from MR images by knowledge-based region growing and trimming | |
US20100014722A1 (en) | Methods involving optimizing and mapping images | |
US10307124B2 (en) | Image display device, method, and program for determining common regions in images | |
WO2006123272A2 (en) | Automatic organ mapping method an device | |
JP5192751B2 (en) | Image processing apparatus, image processing method, and image processing program |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
NENP | Non-entry into the national phase in: |
Ref country code: DE |
|
WWW | Wipo information: withdrawn in national office |
Country of ref document: DE |
|
NENP | Non-entry into the national phase in: |
Ref country code: RU |
|
WWW | Wipo information: withdrawn in national office |
Country of ref document: RU |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 06765678 Country of ref document: EP Kind code of ref document: A2 |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 06765678 Country of ref document: EP Kind code of ref document: A2 |