WO2008035266A2 - Automatic seed point selection - Google Patents
Automatic seed point selection Download PDFInfo
- Publication number
- WO2008035266A2 WO2008035266A2 PCT/IB2007/053730 IB2007053730W WO2008035266A2 WO 2008035266 A2 WO2008035266 A2 WO 2008035266A2 IB 2007053730 W IB2007053730 W IB 2007053730W WO 2008035266 A2 WO2008035266 A2 WO 2008035266A2
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- interest
- region
- seed point
- computer program
- image
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- 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
-
- 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
- G06T2207/10116—X-ray image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20092—Interactive image processing based on input by user
- G06T2207/20104—Interactive definition of region of interest [ROI]
-
- 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
- G06T2207/20112—Image segmentation details
- G06T2207/20156—Automatic seed setting
-
- 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
Definitions
- the invention relates to a computer program to find a single seed point in an image.
- seed point usually an individual pixel or voxel within the image provides some initial value or values, for example by providing an initial grey value or set of coordinates for the computer program, algorithm or sub-routine.
- US 2005/0031202 discloses a technique for segmenting a structure of interest within an image, the technique beginning with the step of seed point selection within the structure.
- the seed point can be identified by the user.
- the computer program is arranged to provide a region of interest in the image and further arranged to allow the region of interest to be manually positioned prior to selection of the single seed point, and arranged to automatically select the single seed point according to at least one pre-defined criterion from the pixels delineated by the region of interest.
- Manual selection of a suitable seed point required for the initiation of an image processing algorithm or a sub-routine is prone to error because of inherent small variations in placement of the cursor on the image.
- the invention removes this source of error by providing the user with a region of interest which is moveable within the image. The user controls the position of the region of interest and is able to translate it within the image until it surrounds the portion of the image in which the user expects a suitable seed point to reside. Once the region of interest is suitably positioned, the seed point is automatically identified within the region of interest according to a single pre-defined criterion or several pre-defined criteria.
- the user places a region of interest instead of placing a seed point. It has been found that it is easier for the user to manually align a region of interest to the portion of the image where a suitable seed point could be expected to reside than it is to try to select the suitable seed point from among the pixels in that portion of the image. A region of interest is inherently larger then the single point identifiable by the cursor and intentional placement is therefore correspondingly easier.
- the criterion, or criteria, by which the seed point would have been manually selected allow, when translated into a pre-defined criterion, or criteria, the automatic selection of the correct pixel within the postioned region of interest. Thus the correct seed point is more accurately identified.
- a particularly useful embodiment of the invention is a region of interest that can also be altered in size.
- the user has improved control over the region of interest and can delineate the region of interest around the portion of the image in where a suitable seed point can be expected to reside.
- the pixels within the region of interest provided for the user form a connected space.
- a particularly useful embodiment has been found to be a region of interest with a substantially circular shape. This may be a completely round circle, but may also be a rounded ellipse. Both have been found to be particularly intuitive shapes for the user to understand and to place over the portion of the image which is suitable for seed point selection.
- the advantages of the circle embodiment are particularly apparent when the user also has control over the size of the region of interest.
- a circle whose radius is controllable imparts an impression of zooming towards a single point when the radius is reduced on the screen, and also indicates zooming away from a single point when the radius is increased, and thereby allows the user to gain an understanding of which portion of the image will contain the single seed point, without obscuration behind the cursor.
- the region of interest can also take other shapes, for example an ellipse, a square, a rectangle, or even a line.
- the latter example is particularly advantageous, for example, when a seed point is sought across a section of some object visible in the image, for example an elongated object such as an arterial vessel, a section of bronchus or a section of colon.
- the user can for example position the line across a suitable width of the object and the seed point is identified from the pixels delineated along the line.
- the eventual single seed point is not merely the point at the center of the region of interest, but is chosen according to a single pre-defined criterion, or several pre-defined criteria, from amongst the group of pixels delineated by the region of interest.
- a pre-defined criterion it is perfectly possible for a pre-defined criterion to in fact be the central point in the region there are many other examples which are clinically found to be advantageous.
- the pre-defined criterion, or criteria, for the automatic selection of the seed point are chosen to be those which most suit the application in which the invention is used.
- the pre-defined criteria are an automation of the required criteria as they would be applied for manual selection of the seed point. So for example, in contrast enhanced images, where any objects of interest might be expected to take up contrast and appear in the resultant image with the subsequently expected range of grey values, the pre-defined criterion could be expected to be selection of the most representative pixel within this range of grey values.
- a suitable seed point for a segmentation algorithm arranged to segment areas of high contrast would be a pixel which itself has a high, contrast enhanced, grey value.
- a suitable pre-defined criterion for automatic selection of the seed point within the positioned region of interest would be selection of the pixel with the maximum grey value.
- a seed point could be selected by use of a filter.
- Filters to identify certain attributes within medical images are known, for example 'vesselness' filters, which are sub-routines which search for groups of pixels or voxels representing vessels, usually blood vessels, within an image according to various a priori criteria. Such filters can be used to automatically select a suitable seed point within the region of interest.
- filters sometimes known as 'blobness' filters can be used to identify objects within an image which is characterized by a roundness or 'blob'-like shape. Such filters are highly useful for identifying polyps in colon data sets and lung nodules in lung data sets.
- Such filters can also be highly advantageously used to automatically select a seed point for use as input value for an image algorithm.
- a weighting function is applied to the pixels delineated by the region of interest prior to automatic selection of the single seed point. This allows the user to favour automatic seed selection in pre-defined portions of the region of interest.
- a particularly advantageous embodiment involves use of a weighting function which applies a pixel weight in proportion to the proximity of the pixel to the centre of the region of interest, in other words, one which weights in favour of pixels situated towards the centre of the region of interest. This provides a particularly intuitive embodiment of the invention because it allows automatic selection of the single seed point in the portion of the region of interest which the user recognises as the centre.
- the user when he is aware of the criterion or criteria used to automatically select the seed point is empowered to position the region of interest evenly around the portion of the image where he most expects a suitable seed point to reside with confidence that the seed point is more likely to be selected from the pixels in the centre of the region of interest.
- This embodiment in effect, allows automatic seed selection to be biased towards the centre of the region of interest.
- This embodiment is found to be particularly useful in the case when the region of interest has the shape of a circle.
- the weighting function can be applied directly to the grey values of the pixels or to the filter responses in the case when a filter is applied. One manner in which this can be performed is to multiply the grey value of each pixel by a value dependent on the distance from the position of the pixel to the calculated centre of the region of interest.
- a particularly useful weighting function is found to be a Gaussian function centred around the centre of the region of interest.
- one manner in which this can be performed is to multiply the grey value of each pixel by a value of the Gaussian function, which value is a function of the distance from the position of the pixel to the calculated centre of the region of interest.
- the Gaussian values may be further multiplied by a scaling factor to increase or reduce the weighting given to the centre of the region of interest and the rate at which this weight increases.
- the invention has been described in terms of its application to a two dimensional image it is equally applicable to medical images which show volume renditions.
- the region of interest may take the shape of a two dimensional representation of a sphere.
- the invention also concerns a workstation arranged to identify a single seed point in a medical image, and arranged to provide a region of interest in the image and further arranged to allow the region of interest to be manually positioned before selection of the single seed point, and further arranged to automatically select the single seed point according to at least one pre-defined criterion from the pixels delineated by the region of interest.
- a workstation has the advantage that it is arranged to perform the invention.
- the workstation is coupled to a computer mouse comprising a rotatable control device, whereby the computer mouse is arranged to provide user control of the region of interest, and further arranged to allow control of the size of the region of interest via the rotatable control device.
- a computer mouse comprising a rotatable control device
- the computer mouse is arranged to provide user control of the region of interest, and further arranged to allow control of the size of the region of interest via the rotatable control device.
- a suitable rotatable control device is a mouse wheel.
- the construction of a computer program, its integration into a workstation and the use of the computer mouse with a rotatable control device according to the invention will be clear to the person skilled in the art of medical image processing once he understands how the invention is to work.
- Fig. 1 shows a workstation suitable for performing the invention.
- Fig. 2 shows a medical image in which the invention can be used.
- Fig. 3 shows a region of interest, as used in the invention.
- Fig. 1 shows a workstation 101 with a screen 102 on which a medical image 103 is displayed.
- the image contains various anatomical structures 104 and a region of interest 105 which is moveable within the image.
- the region of interest 105 can be moved by the mouse 106.
- the user gains control of the region of interest according to the algorithm in operation and moves the region of interest as though it were a cursor.
- the region of interest is situated in the image in such a way that it surrounds the portion of the image in which the user expects a suitable seed point to reside, the user clicks the mouse to allow automatic calculation of the seed point and the initiation of the algorithm which requires the seed point for initialization.
- the size can be controlled by use of a mouse wheel 107.
- Fig. 2 shows a medical image 201 containing an area of tissue 202 which comprises an area of high contrast 203, in other words a recognizable grouping of pixels with high grey values.
- the contrast medium administered prior to the image acquisition is administered by injection so although the high contrast pixels represent tumor they also represent the blood vessels carrying the contrast agent into the area of tissue 202.
- a segmentation algorithm arranged to segment the tumor will commonly start from selection of a seed point. A fully automatic seed point selection cannot be used at this point because it requires the output of the segmentation algorithm to differentiate the pixels contained in the tumor from the pixels contained in the blood vessels visible in the image.
- Fig. 3 shows the same image 301 containing the area of tissue 302 and the area of high contrast 303.
- a region of interest 304 has been situated around the area of high contrast 303 by the user.
- This region of interest 304 can be presented in the image for the user to position and if positioned around the main area of high contrast 303 will allow automatic selection of a seed point lying within the area of high contrast. This seed point will then initialize the segmentation algorithm.
- the criterion for automatic seed point selection need not require the same complexity and rigor as the criteria contained within the segmentation algorithm.
- a simple a priori criterion is sufficient, such as selection of the maximum grey level value when applied to a contrast image.
- the invention also allows improved reproducibility.
- the same seed point is highly likely to be automatically selected even if the user places the region of interest in slightly different positions in successive iterations of the segmentation algorithm.
- the embodiment in which the size as well as the positioning of the region of interest is used is shown to have maximum advantage.
- Alteration of the size of the region of interest allows the user to balance the size so that it includes the region of interest to be segmented while including as little as possible of the surrounding high contrast blood vessels. Thus it is likely that even with small variations in position the same seed point will be automatically selected.
Abstract
A computer program to identify a single seed point in an image is presented in which a manually positionable region of interest is presented to the user in the image and a single seed point is selected according to pre-defined criteria from the pixels delineated by the region of interest. Such seed points are used to initialise, for example, segmentation algorithms. The invention improves accuracy of seed point selection and also increases reproducibility. In an advantageous embodiment the region of interest is also sizable and a workstation and computer mouse with rotatable control device are provided, where the rotatable control device is used to control the size of the region of interest.
Description
Automatic seed point selection
FIELD OF THE INVENTION
The invention relates to a computer program to find a single seed point in an image.
Many computer programs, algorithms and sub-routines used in image analysis, particularly medical image analysis, require initialization with a starting point, known commonly as a seed point. The seed point, usually an individual pixel or voxel within the image provides some initial value or values, for example by providing an initial grey value or set of coordinates for the computer program, algorithm or sub-routine.
BACKGROUND OF THE INVENTION
US 2005/0031202 discloses a technique for segmenting a structure of interest within an image, the technique beginning with the step of seed point selection within the structure. The seed point can be identified by the user.
Correct manual identification of a suitable seed point is prone to error. The user may inadvertently select an incorrect pixel, particularly in images with a wide variation in grey value over a small number of contiguous pixels, for example in images with considerable noise. Semi-automatic algorithms which require user identification of the seed point suffer from the same disadvantage, as do fully automatic algorithms which allow user intervention to select a new seed point upon failure of the algorithm.
SUMMARY OF THE INVENTION
It is an object of the invention to provide a computer program to allow more accurate manual selection of a suitable seed point.
This is achieved according to the invention whereby the computer program is arranged to provide a region of interest in the image and further arranged to allow the region of interest to be manually positioned prior to selection of the single seed point, and arranged to automatically select the single seed point according to at least one pre-defined criterion from the pixels delineated by the region of interest.
Manual selection of a suitable seed point required for the initiation of an image processing algorithm or a sub-routine is prone to error because of inherent small variations in placement of the cursor on the image. This may have a large effect on the output of the image processing algorithm when, for example, the value of the grey level of the selected seed point is used as an input value for the algorithm and the grey level values vary widely from pixel to pixel over the region in which the suitable seed point can be expected to be situated. In such cases it is found empirically that repeated user selection of seed point for sequential iterations of the algorithm often produce varying output results due to an inherent spread in user defined input seed point selection. The invention removes this source of error by providing the user with a region of interest which is moveable within the image. The user controls the position of the region of interest and is able to translate it within the image until it surrounds the portion of the image in which the user expects a suitable seed point to reside. Once the region of interest is suitably positioned, the seed point is automatically identified within the region of interest according to a single pre-defined criterion or several pre-defined criteria.
In effect, the user places a region of interest instead of placing a seed point. It has been found that it is easier for the user to manually align a region of interest to the portion of the image where a suitable seed point could be expected to reside than it is to try to select the suitable seed point from among the pixels in that portion of the image. A region of interest is inherently larger then the single point identifiable by the cursor and intentional placement is therefore correspondingly easier. The criterion, or criteria, by which the seed point would have been manually selected allow, when translated into a pre-defined criterion, or criteria, the automatic selection of the correct pixel within the postioned region of interest. Thus the correct seed point is more accurately identified. It has been found that a particularly useful embodiment of the invention is a region of interest that can also be altered in size. In this embodiment the user has improved control over the region of interest and can delineate the region of interest around the portion of the image in where a suitable seed point can be expected to reside.
The pixels within the region of interest provided for the user form a connected space.
A particularly useful embodiment has been found to be a region of interest with a substantially circular shape. This may be a completely round circle, but may also be a rounded ellipse. Both have been found to be particularly intuitive shapes for the user to understand and to place over the portion of the image which is suitable for seed point
selection. The advantages of the circle embodiment are particularly apparent when the user also has control over the size of the region of interest. A circle whose radius is controllable imparts an impression of zooming towards a single point when the radius is reduced on the screen, and also indicates zooming away from a single point when the radius is increased, and thereby allows the user to gain an understanding of which portion of the image will contain the single seed point, without obscuration behind the cursor.
The region of interest can also take other shapes, for example an ellipse, a square, a rectangle, or even a line. The latter example is particularly advantageous, for example, when a seed point is sought across a section of some object visible in the image, for example an elongated object such as an arterial vessel, a section of bronchus or a section of colon. In this case the user can for example position the line across a suitable width of the object and the seed point is identified from the pixels delineated along the line.
It is important to realize that the eventual single seed point is not merely the point at the center of the region of interest, but is chosen according to a single pre-defined criterion, or several pre-defined criteria, from amongst the group of pixels delineated by the region of interest. Although it is perfectly possible for a pre-defined criterion to in fact be the central point in the region there are many other examples which are clinically found to be advantageous.
In general, the pre-defined criterion, or criteria, for the automatic selection of the seed point are chosen to be those which most suit the application in which the invention is used. The pre-defined criteria are an automation of the required criteria as they would be applied for manual selection of the seed point. So for example, in contrast enhanced images, where any objects of interest might be expected to take up contrast and appear in the resultant image with the subsequently expected range of grey values, the pre-defined criterion could be expected to be selection of the most representative pixel within this range of grey values. To continue the example, if contrast enhanced objects in the image have a high grey value and are subsequently seen to be very bright within the image, a suitable seed point for a segmentation algorithm arranged to segment areas of high contrast would be a pixel which itself has a high, contrast enhanced, grey value. As such, a suitable pre-defined criterion for automatic selection of the seed point within the positioned region of interest would be selection of the pixel with the maximum grey value.
Similarly, selection of the pixels with minimum grey value can also be used. As an alternative example a seed point could be selected by use of a filter. Filters to identify certain attributes within medical images are known, for example
'vesselness' filters, which are sub-routines which search for groups of pixels or voxels representing vessels, usually blood vessels, within an image according to various a priori criteria. Such filters can be used to automatically select a suitable seed point within the region of interest. Similarly, filters sometimes known as 'blobness' filters can be used to identify objects within an image which is characterized by a roundness or 'blob'-like shape. Such filters are highly useful for identifying polyps in colon data sets and lung nodules in lung data sets. Such filters can also be highly advantageously used to automatically select a seed point for use as input value for an image algorithm. In another useful embodiment, a weighting function is applied to the pixels delineated by the region of interest prior to automatic selection of the single seed point. This allows the user to favour automatic seed selection in pre-defined portions of the region of interest. A particularly advantageous embodiment involves use of a weighting function which applies a pixel weight in proportion to the proximity of the pixel to the centre of the region of interest, in other words, one which weights in favour of pixels situated towards the centre of the region of interest. This provides a particularly intuitive embodiment of the invention because it allows automatic selection of the single seed point in the portion of the region of interest which the user recognises as the centre. The user, when he is aware of the criterion or criteria used to automatically select the seed point is empowered to position the region of interest evenly around the portion of the image where he most expects a suitable seed point to reside with confidence that the seed point is more likely to be selected from the pixels in the centre of the region of interest. This embodiment, in effect, allows automatic seed selection to be biased towards the centre of the region of interest. This embodiment is found to be particularly useful in the case when the region of interest has the shape of a circle. The weighting function can be applied directly to the grey values of the pixels or to the filter responses in the case when a filter is applied. One manner in which this can be performed is to multiply the grey value of each pixel by a value dependent on the distance from the position of the pixel to the calculated centre of the region of interest.
A particularly useful weighting function is found to be a Gaussian function centred around the centre of the region of interest. In this case, one manner in which this can be performed is to multiply the grey value of each pixel by a value of the Gaussian function, which value is a function of the distance from the position of the pixel to the calculated centre of the region of interest. The Gaussian values may be further multiplied by a scaling
factor to increase or reduce the weighting given to the centre of the region of interest and the rate at which this weight increases.
The skilled person will understand how to arrange a computer program to perform the invention and how to integrate this program into the normal workings of image analysis software.
Although the invention has been described in terms of its application to a two dimensional image it is equally applicable to medical images which show volume renditions. For example, in the case where a two dimensional representation of a three dimensional space is shown the region of interest may take the shape of a two dimensional representation of a sphere.
The invention also concerns a workstation arranged to identify a single seed point in a medical image, and arranged to provide a region of interest in the image and further arranged to allow the region of interest to be manually positioned before selection of the single seed point, and further arranged to automatically select the single seed point according to at least one pre-defined criterion from the pixels delineated by the region of interest. Such a workstation has the advantage that it is arranged to perform the invention.
In a particularly useful embodiment, the workstation is coupled to a computer mouse comprising a rotatable control device, whereby the computer mouse is arranged to provide user control of the region of interest, and further arranged to allow control of the size of the region of interest via the rotatable control device. This arrangement has the advantage that the user can control the position of the region of interest by positioning the mouse, as is known, while controlling the size of the region of interest by use of the rotatable control device.
A suitable rotatable control device is a mouse wheel. The construction of a computer program, its integration into a workstation and the use of the computer mouse with a rotatable control device according to the invention will be clear to the person skilled in the art of medical image processing once he understands how the invention is to work.
BRIEF DESCRIPTION OF THE DRAWINGS
These and other embodiments of the invention will be further explained with the assistance of the Figs.
Fig. 1 shows a workstation suitable for performing the invention. Fig. 2 shows a medical image in which the invention can be used.
Fig. 3 shows a region of interest, as used in the invention.
DETAILED DESCRIPTION OF THE EMBODIMENTS
Fig. 1 shows a workstation 101 with a screen 102 on which a medical image 103 is displayed. The image contains various anatomical structures 104 and a region of interest 105 which is moveable within the image. The region of interest 105 can be moved by the mouse 106.
In particularly suitable implementation, the user gains control of the region of interest according to the algorithm in operation and moves the region of interest as though it were a cursor. When the region of interest is situated in the image in such a way that it surrounds the portion of the image in which the user expects a suitable seed point to reside, the user clicks the mouse to allow automatic calculation of the seed point and the initiation of the algorithm which requires the seed point for initialization.
In the particularly advantageous embodiment in which the user also has control of the size of the region of interest the size can be controlled by use of a mouse wheel 107.
Fig. 2 shows a medical image 201 containing an area of tissue 202 which comprises an area of high contrast 203, in other words a recognizable grouping of pixels with high grey values. The contrast medium administered prior to the image acquisition is administered by injection so although the high contrast pixels represent tumor they also represent the blood vessels carrying the contrast agent into the area of tissue 202. A segmentation algorithm arranged to segment the tumor will commonly start from selection of a seed point. A fully automatic seed point selection cannot be used at this point because it requires the output of the segmentation algorithm to differentiate the pixels contained in the tumor from the pixels contained in the blood vessels visible in the image.
Fig. 3 shows the same image 301 containing the area of tissue 302 and the area of high contrast 303. In this case a region of interest 304 has been situated around the area of high contrast 303 by the user. This region of interest 304 can be presented in the image for the user to position and if positioned around the main area of high contrast 303 will allow automatic selection of a seed point lying within the area of high contrast. This seed point will then initialize the segmentation algorithm.
The criterion for automatic seed point selection need not require the same complexity and rigor as the criteria contained within the segmentation algorithm. A simple a
priori criterion is sufficient, such as selection of the maximum grey level value when applied to a contrast image.
It can be seen that the invention also allows improved reproducibility. In a case such as that shown in Figs. 2 and 3, the same seed point is highly likely to be automatically selected even if the user places the region of interest in slightly different positions in successive iterations of the segmentation algorithm. In this case the embodiment in which the size as well as the positioning of the region of interest is used is shown to have maximum advantage. Alteration of the size of the region of interest allows the user to balance the size so that it includes the region of interest to be segmented while including as little as possible of the surrounding high contrast blood vessels. Thus it is likely that even with small variations in position the same seed point will be automatically selected.
Although the invention has been described primarily in terms of seed point selection within two dimensional images, one possible application within three dimensional volume imaging can be implemented by presentation to the user of three circular regions of interest in three orthogonal views.
Claims
1. A computer program to identify a single seed point in an image, characterized in that the computer program is arranged to provide a region of interest 304 in the image and further arranged to allow the region of interest to be manually positioned prior to selection of the single seed point, and arranged to automatically select the single seed point according to at least one pre-defined criterion from the pixels delineated by the region of interest.
2. A computer program as claimed in claim 1, characterized in that the computer program string is further arranged to allow the region of interest to be manually sized before selection of the single seed point.
3. A computer program as claimed in either claims 1 or 2, characterized in that the pixels in the region of interest form a connected space.
4. A computer program as claimed in any of the previous claims, characterized in that the region of interest has a substantially circular shape.
5. A computer program as claimed in any of claims 1 to 4, characterized in that the pre-defined criterion is selection of the pixel with the maximal grey value from amongst the pixels delineated by the region of interest.
6. A computer program as claimed in any of claims 1 to 4, characterized in that the pre-defined criterion is selection of the pixel with the minimal grey value from amongst the pixels delineated by the region of interest.
7. A computer program as claimed in any of claims 1 to 4, characterized in that the pre-defined criterion comprises use of a filter.
8. A computer program as claimed in claim 7, characterized in that the filter is arranged to identify portions of the image which represent anatomical vessels.
9. A computer program as claimed in any of the previous claims, characterised in that a weighting function is applied to the pixels delineated by the region of interest prior to automatic selection of the single seed point.
10. A computer program as claimed in claim 9, characterised in that the weighting function applies a pixel weight in proportion to the proximity of the pixel to the centre of the region of interest.
11. A computer program as claimed in claim 10, characterised in that the weighting function is a Gaussian function.
12. A workstation 101 arranged to identify a single seed point in a medical image 103, characterized in that it is arranged to provide a region of interest 105 in the image and further arranged to allow the region of interest to be manually positioned prior to selection of the single seed point, and arranged to automatically select the single seed point according to at least one pre-defined criterion from the pixels delineated by the region of interest 105.
13. A workstation as claimed in claim 12, characterized in that it is coupled to a computer mouse 106 comprising a rotatable control device 107, the computer mouse 106 arranged to provide user control of the region of interest, and further arranged to allow control of the size of the region of interest via the rotatable control device 107.
14. method to identify a single seed point in an image 301 characterised in that a positionable region of interest 304 is provided in the image 301 and is positioned in the image 301 prior to selection of the single seed point, and further that the single seed point is selected according to at least one pre-defined criterion from the pixels delineated by the region of interest 304.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP06121047 | 2006-09-21 | ||
EP06121047.2 | 2006-09-21 |
Publications (2)
Publication Number | Publication Date |
---|---|
WO2008035266A2 true WO2008035266A2 (en) | 2008-03-27 |
WO2008035266A3 WO2008035266A3 (en) | 2009-01-29 |
Family
ID=39200918
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/IB2007/053730 WO2008035266A2 (en) | 2006-09-21 | 2007-09-17 | Automatic seed point selection |
Country Status (1)
Country | Link |
---|---|
WO (1) | WO2008035266A2 (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090125620A1 (en) * | 2007-11-13 | 2009-05-14 | John Gregory Klincewicz | Assigning telecommunications nodes to community of interest clusters |
EP2116973A1 (en) * | 2008-05-06 | 2009-11-11 | Carestream Health, Inc. | Method for interactively determining a bounding surface for segmenting a lesion in a medical image |
US9536318B2 (en) | 2012-12-03 | 2017-01-03 | Koninklijke Philips N.V. | Image processing device and method for detecting line structures in an image data set |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1426903A2 (en) * | 2002-11-26 | 2004-06-09 | GE Medical Systems Global Technology Company LLC | Computer aided diagnosis of an image set |
-
2007
- 2007-09-17 WO PCT/IB2007/053730 patent/WO2008035266A2/en active Application Filing
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1426903A2 (en) * | 2002-11-26 | 2004-06-09 | GE Medical Systems Global Technology Company LLC | Computer aided diagnosis of an image set |
Non-Patent Citations (4)
Title |
---|
DAMILAKIS J ET AL: "Broadband ultrasound attenuation imaging: algorithm development and clinical assessment of a region growing technique" PHYSICS IN MEDICINE AND BIOLOGY IOP PUBLISHING UK, vol. 47, no. 2, January 2002 (2002-01), pages 315-325, XP002493537 ISSN: 0031-9155 * |
ELENA MARTÍNEZ ET AL: "Region growing" SUBSECTION OF "RETINAL BLOOD VESSEL SEGMENTATION BY MEANS OF MULTISCALE ANALYSIS AND REGION GROWING", [Online] 16 May 2003 (2003-05-16), XP002493539 Retrieved from the Internet: URL:http://turing.iimas.unam.mx/~elena/Projects/segmenta/IEEEMI00-R2/node6.html> [retrieved on 2008-08-26] * |
OUERHANI N ET AL: "Visual attention guided seed selection for color image segmentation" COMPUTER ANALYSIS OF IMAGES AND PATTERNS. 9TH INTERNATIONAL CONFERENCE, CAIP 2001. PROCEEDINGS (LECTURE NOTES IN COMPUTER SCIENCE VOL.2124) SPRINGER-VERLAG BERLIN, GERMANY, 2001, pages 630-637, XP002493538 ISBN: 3-540-42513-6 * |
TUDUKI Y ET AL: "Automated seeded region growing algorithm for extraction of cerebral blood vessels from magnetic resonance angiographic data" ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY, 2000. PROCEEDINGS OF THE 22ND ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE 23-28 JULY 2000, PISCATAWAY, NJ, USA,IEEE, vol. 3, 23 July 2000 (2000-07-23), pages 1756-1759, XP010530840 ISBN: 978-0-7803-6465-3 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090125620A1 (en) * | 2007-11-13 | 2009-05-14 | John Gregory Klincewicz | Assigning telecommunications nodes to community of interest clusters |
US8275866B2 (en) * | 2007-11-13 | 2012-09-25 | At&T Intellectual Property I, L.P. | Assigning telecommunications nodes to community of interest clusters |
US8495201B2 (en) | 2007-11-13 | 2013-07-23 | At&T Intellectual Property I, L.P. | Assigning telecommunications nodes to community of interest clusters |
US8914491B2 (en) | 2007-11-13 | 2014-12-16 | At&T Intellectual Property, I, L.P. | Assigning telecommunications nodes to community of interest clusters |
EP2116973A1 (en) * | 2008-05-06 | 2009-11-11 | Carestream Health, Inc. | Method for interactively determining a bounding surface for segmenting a lesion in a medical image |
US8150120B2 (en) | 2008-05-06 | 2012-04-03 | Carestream Health, Inc. | Method for determining a bounding surface for segmentation of an anatomical object of interest |
US9536318B2 (en) | 2012-12-03 | 2017-01-03 | Koninklijke Philips N.V. | Image processing device and method for detecting line structures in an image data set |
Also Published As
Publication number | Publication date |
---|---|
WO2008035266A3 (en) | 2009-01-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20210375011A1 (en) | Image color adjustment method and system | |
JP4879028B2 (en) | Image processing method, image analysis method, and program storage medium | |
CN110050281B (en) | Annotating objects in a learning image | |
US8320989B2 (en) | Region of interest methods and systems for ultrasound imaging | |
EP2724317B1 (en) | System and method for processing a medical image | |
JP6018216B2 (en) | Interactive image annotation system and method | |
US20220309676A1 (en) | Method and device of extracting label in medical image | |
CN110648338B (en) | Image segmentation method, readable storage medium, and image processing apparatus | |
KR102372200B1 (en) | Method and apparatus for classification of lesion based on learning data applying one or more augmentation method in lesion information augmented patch of medical image | |
US20110123086A1 (en) | Systems and methods for enhancing medical images | |
US20090116718A1 (en) | Medical image display device and program | |
EP3198565B1 (en) | Visualizing volumetric image of an anatomical structure | |
EP2898668A1 (en) | System for photograph enhancement by user controlled local image enhancement | |
US20120032953A1 (en) | Automated contrast enhancement for contouring | |
Egger | Refinement-cut: user-guided segmentation algorithm for translational science | |
AU2013383628B2 (en) | Image processing apparatus, program, computer readable medium and image processing method | |
WO2008035266A2 (en) | Automatic seed point selection | |
CN108573523B (en) | Method and system for segmented volume rendering | |
US9558561B2 (en) | Semiautomatic drawing tool for image segmentation | |
CN107329669B (en) | Method and device for selecting human body sub-organ model in human body medical three-dimensional model | |
Egger et al. | Interactive outlining of pancreatic cancer liver metastases in ultrasound images | |
JP2008541824A (en) | Automated organ linking for organ model placement | |
US20240087190A1 (en) | System and method for synthetic data generation using dead leaves images | |
US10664136B2 (en) | Method, system and apparatus for selecting items in a graphical user interface | |
Na et al. | Superpixel-based line operator for retinal blood vessel segmentation |
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: 07826396 Country of ref document: EP Kind code of ref document: A2 |
|
NENP | Non-entry into the national phase in: |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 07826396 Country of ref document: EP Kind code of ref document: A2 |