Methods and systems for automatic propagation and segmentation in medical images

Abstract

This paper discloses a method and system for automatic propagation segmentation in medical images. In one embodiment, the method uses a segmented region of interest (RoI) in a reference image to determine segmentation parameters and multiple reference points. Furthermore, the method generates multiple translation points on the current image by translating the multiple reference points onto the current image, where the target RoI to be segmented exists. Subsequently, relevant seeds are automatically selected from the translation points based on the segmentation parameters. Finally, multiple sub-segments using the selected relevant seeds are performed to estimate and segment the target RoI in the current image, such that the target RoI is a propagation segmentation of the segmented RoI in the reference image.

Description

Technical Field

[0001] This topic generally relates to the field of image processing technology, and more specifically, but not exclusively, to methods and systems for the automatic propagation of segmentation in medical images. Background Technology

[0002] Medical imaging has become a primary tool for diagnosing a wide range of diseases. Medical imaging is the process of creating visual representations of the body's interior for clinical analysis and medical intervention, as well as for functional analysis of internal organs or tissues. Typically, visual representations can be in the form of images and videos. One of the first and most important processes used to analyze these visual representations is segmentation. Segmentation is the process of dividing an image into distinct, meaningful segments that correspond to different tissue categories, organs, pathological conditions, or other biologically relevant structures.

[0003] Human anatomy is typically composed of multiple types of tissues. Therefore, in certain diagnostic procedures, particularly in tumor diagnosis, individual tissues may appear differently in different patterns. Consequently, using common segmentation procedures, such as computer-based automated tumor segmentation, remains an ongoing challenge in tumor diagnosis.

[0004] Furthermore, most existing segmentation procedures are tailored for segmenting tumors in images with specific patterns and anatomical structures. Therefore, there is still a need for segmentation techniques that are universal across modalities and anatomical structures.

[0005] The information disclosed in the background section of this disclosure is intended only to enhance the understanding of the general background of the invention and should not be construed as an admission of prior art known to those skilled in the art or any form of implication. Summary of the Invention

[0006] This paper discloses a method for automatic propagating segmentation in medical images. The method includes determining one or more segmentation parameters based on analysis of a segmented region of interest (RoI) in a reference image. Furthermore, the method includes determining multiple reference points corresponding to the segmented reference RoI based on one or more morphological characteristics of the segmented reference RoI. After determining the multiple reference points, the method includes generating multiple translation points in the current image by translating each of the multiple reference points onto the current image, where a target RoI to be segmented exists. Furthermore, the method includes automatically selecting relevant seeds in the current image from the multiple translation points based on one or more segmentation parameters. Finally, the method includes performing multiple sub-segments of the selected relevant seeds to estimate and segment the target RoI in the current image, where the target RoI is a propagating segmentation of the segmented RoI in the reference image, and where the pixel intensity of the current image is quantitatively comparable to or normalizable to the pixel intensity of the reference image.

[0007] Furthermore, this disclosure relates to an image segmentation system for automatically propagating segmentation in medical images. The image segmentation system includes a processor and a memory communicatively coupled to the processor. The memory stores processor-executable instructions that, upon execution, cause the processor to determine one or more segmentation parameters based on an analysis of a segmented region of interest (RoI) in a reference image. Furthermore, the instructions cause the processor to determine a plurality of reference points corresponding to the segmented reference RoI based on one or more morphological characteristics of the segmented reference RoI. Subsequently, the instructions cause the processor to generate a plurality of translation points in the current image by translating each of the plurality of reference points onto the current image, where a target RoI to be segmented exists. Furthermore, the instructions cause the processor to automatically select a relevant seed from the plurality of translation points in the current image based on one or more segmentation parameters. Finally, the instructions cause the processor to perform multiple sub-segments of the selected relevant seed to estimate and segment the target RoI in the current image, wherein the target RoI is a propagated segmentation of the segmented RoI in the reference image, and wherein the pixel intensity of the current image is quantitatively comparable to or normalizable to the pixel intensity of the reference image.

[0008] The foregoing overview is illustrative only and is not intended to be limiting in any way. Other aspects, embodiments, and features will become clear from the accompanying drawings and the following detailed description, in addition to the illustrative aspects, embodiments, and features described above. Attached Figure Description

[0009] The accompanying drawings, which are incorporated in and constitute a part of this disclosure, illustrate exemplary embodiments and, together with the description, explain the principles disclosed. In the drawings, the leftmost numerals(s) of the reference numeral(s) identify the figure in which the reference numeral(s) first appear. The same numerals are used throughout the drawings to denote the same features and components. Some embodiments of systems and / or methods according to this subject matter will now be described by way of example only and with respect to the accompanying drawings, in which:

[0010] Figure 1 An exemplary environment for automatically propagating segmentation in medical images according to some embodiments of the present disclosure is shown;

[0011] Figure 2 Detailed block diagrams of an image segmentation system for automatic propagation segmentation according to some embodiments of the present disclosure are shown;

[0012] Figures 3A to 3H The process of propagation segmentation with respect to a reference image and a current image according to an exemplary embodiment of the present disclosure is illustrated;

[0013] Figure 4 A flowchart illustrating a method for automatically propagating segmentation in medical images according to some embodiments of the present disclosure is shown; and

[0014] Figure 5 A block diagram of an exemplary computer system for implementing embodiments consistent with this disclosure is shown.

[0015] Those skilled in the art will understand that any block diagram herein represents a conceptual view of an illustrative system embodying the principles of the subject matter. Similarly, it should be understood that any flowchart, diagram, state transition diagram, pseudocode, etc., represents various processes that can be substantially represented in a computer-readable medium and executed by a computer or processor, whether or not such computer or processor is explicitly shown. Detailed Implementation

[0016] In this document, the word "exemplary" is used to mean "as an example, instance, or illustration." Any embodiment or implementation of the subject matter described herein as "exemplary" is not necessarily to be construed as superior to or preferred over other embodiments.

[0017] While this disclosure is readily adaptable to various modifications and alternatives, specific embodiments thereof have been illustrated by way of example in the accompanying drawings and will be described in detail below. However, it should be understood that this is not intended to limit this disclosure to the particular forms disclosed, but rather, this disclosure will cover all modifications, equivalents, and alternatives falling within the spirit and scope of this disclosure.

[0018] The terms “comprises,” “comprising,” “includes,” or any other variations thereof are intended to cover non-exclusive inclusion, such that an arrangement, apparatus, or method that includes a list of components or steps includes not only those components or steps but also other components or steps not expressly listed or inherent to such arrangement, apparatus, or method. In other words, a system or apparatus that begins with “comprising…a(a)” does not exclude the presence of other elements or additional elements in the system or method unless further constraints are imposed.

[0019] Embodiments of this disclosure can be used to automatically propagate lesion segmentation through one or more subsequent oncology sessions during periodic scanning of lesions and / or other affected areas.

[0020] Therefore, in one embodiment, this disclosure discloses a method and image segmentation system for automatically propagating segmentation in medical images. In one embodiment, the method of this disclosure includes picking a reference point corresponding to a segmented region of interest (RoI) in a reference image and / or a reference study, and translating the reference point to a current image of the lesion captured in a subsequent session to perform multiple sub-segmentation. Multiple sub-segmentation of the selected relevant reference point on the current image helps to estimate and segment a target RoI in the current image. In one embodiment, the target RoI may be a propagated segmentation of a segmented RoI in a reference image. Therefore, this disclosure facilitates automatic propagating segmentation in medical images.

[0021] In one embodiment, this disclosure performs lesion segmentation over any anatomical region and any modality to infer desired parameters from the segmented lesions. As an example, the parameters that can be inferred may include, but are not limited to, the size, shape, location, intensity distribution, homogeneity / heterogeneity of the lesions, and the appearance of the surrounding background tissue.

[0022] In one embodiment, this disclosure addresses the technical problem of repeatedly segmenting lesions in subsequent scans by automating the propagation of segmentation using information from lesions segmented in a first scan.

[0023] In one embodiment, the methods and image segmentation systems disclosed herein can be used to propagate the segmentation of objects / entities with similar characteristics in multiple studies such as scans, images, volumes, etc., when the segmentation of objects / entities is available in at least one such study.

[0024] In the following detailed description of embodiments of the present disclosure, reference is made to the accompanying drawings, which form a part of the present invention, in which specific embodiments in which the invention may be practiced are illustrated by way of illustration. These embodiments have been described in sufficient detail to enable those skilled in the art to practice the present disclosure, and it should be understood that other embodiments may be utilized and changes may be made without departing from the scope of the present disclosure. Therefore, the following description should not be construed as limiting.

[0025] Figure 1 An exemplary environment for automatically propagating segmentation in medical images is shown according to some embodiments of the present disclosure.

[0026] Environment 100 may include an image segmentation system 101 and a reference database 105 associated with the image segmentation system 101. In one embodiment, the image segmentation system 101 may be a computing device, such as, but not limited to, a desktop computer, laptop computer, smartphone, or server, which may be configured to automatically propagate segmentation in medical images according to embodiments of the present disclosure. In one embodiment, the reference database 105 may be a storage unit for storing patient-related information, patient-related historical medical images and / or historical reference images 103, and other information required for propagating segmentation. In one implementation, the reference database 105 may be part of the image segmentation system 101 and reside within the image segmentation system 101.

[0027] In one embodiment, reference image 103 and current image 107 can be images of entities such as organs, tissues, bones, air cavities, and muscles of an object, captured at different time periods. Here, the object can be a person, animal, etc. Furthermore, as an example, reference image 103 can be an image of an entity captured during an initial study of the entity, from which a reference region of interest (RoI) can be segmented. Current image 107 can be an image of the same entity captured during a subsequent study of the object. In one embodiment, reference image 103 and current image 107 can be captured by an external device such as a computed tomography (CT) scanner or a magnetic resonance imaging (MRI) device. In one embodiment, image segmentation system 101 can receive reference image 103 and current image 107 from a medical device or a reference database using a wired and / or wireless communication interface configured between image segmentation system 101 and the external device. In one implementation, image segmentation system 101 and the external device can be implemented as a single system that performs scanning and segmentation.

[0028] In one embodiment, when capturing a reference image 103 of an entity, a technical expert associated with the image segmentation system 101 can manually perform segmentation of the reference image 103 to label and / or obtain a reference RoI from the reference image 103. Subsequently, when capturing the current image 107 of the entity, the current image 107, together with the reference image 103 and the segmented reference RoI, can be provided as input to the image segmentation system 101 for automatically performing segmentation of the current image 107.

[0029] In one embodiment, after receiving a reference image 103, a current image 107, and a reference RoI, the image segmentation system 101 can determine one or more segmentation parameters 103 associated with the reference image 103 based on analysis of the segmented reference RoI in the reference image. In one embodiment, the one or more segmentation parameters determined from the reference image 103 may include, but are not limited to, one or more morphological characteristics of the segmented reference RoI, the position of the segmented reference RoI within the reference image 103, the intensity distribution of the segmented reference RoI, and the background intensity distribution of the background of the segmented reference RoI.

[0030] In one embodiment, one or more morphological characteristics may include, but are not limited to, the minor axis and longest diameter of the segmented reference RoI, and the size and shape of the segmented reference RoI. In one embodiment, the size and shape of the segmented reference RoI may be determined based on the length and coordinates of the minor axis and longest diameter of the segmented reference RoI.

[0031] In one embodiment, the intensity distribution of the segmented reference RoI can be a variation in pixel intensity within a region of the segmented reference RoI. Similarly, the background intensity distribution of the background of the segmented reference RoI can be a variation in pixel intensity within a region surrounding the segmented reference RoI. In one embodiment, the background intensity distribution of the segmented reference RoI can be determined by determining multiple intensity distributions of multiple background pixels / points corresponding to quadrants of the segmented reference RoI relative to the center of the segmented reference RoI. In one embodiment, the intensity distribution range can be determined based on the amount of overlap between the pixel intensities of the corresponding background intensities in multiple background quadrants and the pixel intensities of the segmented reference RoI.

[0032] In one embodiment, when determining one or more segmentation parameters from a segmented reference RoI, the image segmentation system 101 may determine a plurality of reference points corresponding to the segmented reference RoI based on one or more morphological characteristics of the segmented reference RoI. In one embodiment, the plurality of reference points may be points located along the minor axis and the longest diameter of the segmented reference RoI.

[0033] In one embodiment, after determining multiple reference points, the image segmentation system 101 can generate multiple translation points on the current image 107 of the entity, where the target RoI 109 to be segmented exists. In one embodiment, the multiple translation points can be obtained by translating each of the multiple reference points onto the current image 107.

[0034] In one embodiment, when generating multiple translation points, the image segmentation system 101 can automatically select relevant seeds from multiple translation points in the current image 107 based on one or more segmentation parameters.

[0035] In one embodiment, once a relevant seed is selected from a plurality of translation points, the image segmentation system 101 can perform multiple sub-segments of the selected relevant seed to estimate and segment the target RoI 109 in the current image 107. In one embodiment, the target RoI 109 may correspond to a propagation segmentation of the segmented RoI in the reference image 103. In one embodiment, a prerequisite for performing segmentation of the target RoI 109 may be that the pixel intensity of the current image 107 must be quantitatively comparable to or normalized to the pixel intensity of the reference image 103.

[0036] In one embodiment, after segmenting the target RoI 109 from the current image 107, the target RoI 109 can be compared with a segmented reference RoI to determine changes in one or more corresponding segmentation parameters of the segmented reference RoI and the target RoI 109. In one embodiment, during subsequent studies of the entity, the current image 107 of the entity and the segmented target RoI 109 together can be considered as a reference image 103 with a reference segmented RoI.

[0037] Figure 2 A detailed block diagram of an image segmentation system 101 for automatic propagation segmentation according to some embodiments of the present disclosure is shown.

[0038] In one implementation, the image segmentation system 101 may include an I / O interface 201, a processor 203, and a memory 205. The I / O interface 201 may be configured to communicate with one or more sources and / or external devices for receiving a reference image 103 and a current image 107. Furthermore, the I / O interface 201 may be used to connect the image segmentation system 101 to a display interface to display the reference image 103, the current image 107, and segmented regions of the images to a user. In one embodiment, the memory 205 may be communicatively coupled to the processor 203. The processor 203 may be configured to perform one or more functions of the image segmentation system 101.

[0039] In some implementations, the image segmentation system 101 may include data 207 and modules 209 for performing various operations according to embodiments of the present disclosure. In one embodiment, data 207 may be stored in memory 205 and may include, but is not limited to, one or more segmentation parameters 211, a plurality of reference points 213, a plurality of translation points 215, a target region of interest (RoI) 217, and other data 219.

[0040] In some embodiments, data 207 may be stored in memory 205 in the form of various data structures. Additionally, data 207 may be organized using data models such as relational or hierarchical data models. Other data 219 may store temporary data and temporary files generated by module 209 when performing various functions of the image segmentation system 101. As an example, other data 219 may include, but is not limited to, one or more historical or reference images 103 of an entity, morphological characteristics of the segmented RoI, etc.

[0041] In one embodiment, data 207 may be processed by one or more modules 209 of the image segmentation system 101. As used herein, the term module refers to an application-specific integrated circuit (ASIC), electronic circuitry, a processor (shared, dedicated, or grouped) and memory executing one or more software or firmware programs, combinational logic circuitry, and / or other suitable components that provide the aforementioned functionality. In one embodiment, other modules 231 may be used to perform various miscellaneous functions of the image segmentation system 101. It should be understood that such modules 209 may be represented as a single module or a combination of different modules.

[0042] In one implementation, one or more modules 209 may be stored as instructions executable by processor 203. In another implementation, each of the one or more modules 209 may be a separate hardware unit communicatively coupled to processor 203 to perform one or more functions of image segmentation system 101. The one or more modules 209 may include, but are not limited to, parameter determination module 221, reference point determination module 223, seed generation module 225, seed selection module 227, segmentation module 229, and other modules 231.

[0043] In one embodiment, the parameter determination module 221 may be used to determine one or more segmentation parameters 211 based on the analysis of segmented reference RoIs in the reference image 103. In one embodiment, the most important step in propagating the segmentation of the reference image 103 may be determining one or more segmentation parameters 211, such as, but not limited to, one or more morphological characteristics of the segmented reference RoI, the position of the segmented reference RoI within the reference image 103, the intensity distribution of the segmented reference RoI, and the intensity distribution of the background of the segmented reference RoI.

[0044] In one embodiment, the parameter determination module 221 can determine one or more morphological characteristics of the segmented reference RoI by determining the minor axis and longest diameter of the segmented reference RoI, and then determining the size and shape of the segmented reference RoI based on the length and coordinates of the minor axis and longest diameter of the segmented reference RoI. Furthermore, the position of the segmented reference RoI within the reference image 103 can be determined based on the coordinates of points on the minor axis and longest diameter.

[0045] In one embodiment, the parameter determination module 221 can determine the intensity distribution of the segmented reference RoI by calculating an intensity histogram of the pixels forming the segmented reference RoI. Furthermore, the parameter determination module 221 can derive the heterogeneity of the segmented reference RoI by thresholding the histogram at predetermined frequencies. For example, pixel intensity values ​​with frequencies greater than 60% of the highest frequency values ​​in the histogram can be considered. Throttling can yield multiple pixel intensity ranges, referred to as lesion ranges (Lr). In one embodiment, if the segmented reference RoI is uniform, there may be only one pixel intensity range. On the other hand, if the segmented reference RoI is heterogeneous, histogram thresholding can yield multiple pixel intensity ranges.

[0046] In one embodiment, the parameter determination module 221 can determine the intensity distribution of the background and / or surrounding regions of the segmented reference RoI through a process similar to calculating a histogram. That is, the intensity distribution can be determined by calculating a histogram and then thresholding the histogram at a predetermined frequency. However, this can be accomplished by grouping pixels into four quadrants based on their coordinates, with the center of the segmented reference RoI as the origin of the quadrant. Therefore, there can be four histograms corresponding to each of the four quadrants and four sets of pixel intensity ranges resulting from the thresholding of the histograms. These pixel intensity ranges can be referred to as the background range (BGr).

[0047] In one embodiment, after obtaining Lr and four BGr values, parameter determination module 221 can calculate the neighbor threshold (NT) and background histogram percentage (BbHp). In one embodiment, NT can represent the minimum number of neighboring pixels of a point that must meet the Lr and BGr criteria to be included in the candidate list and / or selected as a reference point. In one embodiment, BgHp can represent the actual background histogram threshold percentage, which can be taken into account when performing intensity comparisons to include relevant reference seeds in the candidate list.

[0048] In one embodiment, the parameter determination module 221 can determine the background parameters using the following method. Initially, for each quadrant, the background histogram can be thresholded using multiple frequency ranges, such as 10%, 30%, 50%, and 70%. Therefore, at the end of the thresholding, BGr can be obtained for each frequency threshold for each quadrant of the background pixels. In one embodiment, the overlap between Lr and each BGr can be calculated, and the values ​​of NT and BgHp can be iteratively derived, as shown in Tables A and B below.

[0049] In one embodiment, the criteria used to derive parameters can be different for different types and / or sources of images. For example, the criteria for deriving parameters from images acquired from computed tomography (CT) scans may be as shown in Table A. Similarly, the criteria for deriving parameters from images acquired from magnetic resonance imaging (MRI) may be as shown in Table B.

[0050]

[0051] Table A: Exported parameters of CT scans

[0052]

[0053] Table B: Exported parameters of MRI scans

[0054] In one embodiment, the overlap percentage (as shown in Tables A and B) can be the percentage of Lr that overlaps with BGr. Furthermore, the overlap count (as shown in Table B) can be the quadrant number of BGr that overlaps with Lr.

[0055] In one embodiment, the reference point determination module 223 may be used to determine a plurality of reference points 213 corresponding to the segmented reference RoI based on one or more morphological characteristics of the segmented reference RoI. In one embodiment, since the longest diameter and minor axis of the segmented reference RoI, determined based on a reference study of reference image 103, lie on the cross-section of the RoI, the longest diameter and minor axis can be considered to cover most of the different intensity regions in the segmented RoI. Therefore, in one embodiment, the plurality of reference points 213 on the segmented reference RoI can be determined as the coordinates of discrete points identified on the longest diameter and minor axis.

[0056] In one embodiment, the process of obtaining a reference RoI and determining multiple reference points 213 on the reference RoI can be used Figures 3A to 3C The example representation in the text is used to illustrate this.

[0057] Figure 3A Reference image 103, including reference RoI 303, is shown. Figure 3BThe longest diameter 303A and minor axis 303B of the segmented reference RoI 303 are shown. Now, multiple reference points 213 corresponding to the segmented reference RoI 303 can be determined by identifying a set of discrete points on the longest diameter 303A and the minor axis 303B of the segmented reference RoI 303, as shown. Figure 3C As shown.

[0058] In one embodiment, the seed generation module 225 can be used to generate multiple translation points 215 on the current image 107 by translating each of a plurality of reference points 213 onto the current image 107, where the target RoI 109 to be segmented exists. In one embodiment, once all the reference points 213 have been determined and extracted from the reference image 103, the seed generation module 225 can transform each of the plurality of reference points 213 onto the current image 107 to obtain multiple translation points 215 on the current image 107. That is, the reference points 213 determined from the reference image 103 can be transformed into the current image 107 as translation points 215, such as... Figure 3D As shown.

[0059] In one embodiment, the seed selection module 227 can be used to automatically select relevant seeds 311 in the current image 107 from a plurality of translation points 215 based on one or more segmentation parameters 211. Figure 3E The relevant seed 311, which is included in the candidates and selected from the translation point 215, is shown. In one embodiment, the pixel intensity range corresponding to the segmented reference RoI 303 and for each of the four quadrants of the background pixels can be the most critical segmentation parameter 211 for selecting the relevant seed 311 in the current image 107.

[0060] In one embodiment, one of the plurality of translation points 215 "P" may be included as a candidate for the relevant seed only if the seed point "P" satisfies one or more of the following conditions:

[0061] 1. The pixel intensity of point P is within the range defined by Lr.

[0062] 2. The pixel intensity of point P is outside the background intensity range BGr. Here, the background intensity range can be the set of intensity ranges obtained by thresholding the histogram using the BgHp of the background pixels in the quadrant to which point P belongs.

[0063] 3. The number of direct neighbors of point P that satisfies conditions 1 and 2 above is greater than or equal to NT.

[0064] 4. The intensity Ip at point P makes:

[0065] lp <Lr<BGr

[0066] or

[0067] Ip>Lr>BGr

[0068] In one embodiment, point P can be included in the candidate list when point P satisfies all three conditions 1, 2, and 3 above, or when point P satisfies only condition 4 above.

[0069] In one embodiment, if point P is not included in the candidate list, the neighboring points of point P can be checked, and if the neighboring points satisfy conditions 1, 2 and 3 above, they are included in the candidate list.

[0070] In one embodiment, after repeating the above analysis for all points of a plurality of translation points 215, if none of the plurality of translation points 215 are included in the candidate list, then the translation point 215 that only satisfies condition 1 can be included in the candidate list as a relevant seed 311.

[0071] In one embodiment, in the case of CT images, bilateral filtering of pixel intensity at point P can be used for all conditions 1-4 above. However, for MR images, filtering may not be applied, and pixel intensity can be compared directly. Furthermore, in CT images, intensity represents quantitative information, so two CT scans can be compared directly. However, in MR images, since pixel intensity only represents qualitative information, two scans with the same variant may need to be normalized before comparison.

[0072] In one embodiment, segmentation module 229 can be used to perform multiple sub-segments of the selected relevant seed 311 to estimate and segment the target RoI 109 in the current image 107. In one embodiment, multiple sub-segmentation can be considered as an extension of the single-seed segmentation technique, which is used to address the heterogeneity of the reference RoIs of all selected relevant seed points obtained in the above process. In one embodiment, multiple sub-segmentation of the selected relevant seed 311 can be performed using one of the existing multiple sub-segmentation techniques such as parametric methods, level set methods, clustering methods, etc. Alternatively, multiple segmentation can also be performed using region growing techniques, as shown in this disclosure.

[0073] In one embodiment, segmentation module 229 may receive user input indicating whether the target RoI 109 determined in the current image 107 is greater than, less than, or equal to the size of the reference RoI determined in the reference image 103. This user input can be used to limit the area to be searched and / or scanned when performing multiple subsegments.

[0074] In one embodiment, if the target RoI 109 in the current image 107 is larger than the reference RoI, the search region radius can be set to be equal to the longest diameter of the reference RoI.

[0075] In one embodiment, if the target RoI 109 is smaller than the reference RoI, the search area radius can be set to 50% of the longest diameter of the reference lesion.

[0076] In one embodiment, if the target RoI 109 has the same size as the reference RoI, the search area radius can be set to 75% of the longest diameter of the reference lesion.

[0077] That is, the difference between the size of the target RoI 109 and the size of the reference RoI can be correlated and adjusted using one of the above adjustments.

[0078] Figure 3F The diagram illustrates regions obtained by performing multiple sub-segments of the current image 107 using a relevant seed 311. In one embodiment, the multiple sub-segments may disregard some regions of the target RoI 109, which have different intensities compared to the overall intensity of the RoI, when generating the target RoI 109. However, this can be addressed by using an intensity-based k-means clustering region growing technique within the regions defined by the multiple sub-segment outputs.

[0079] In one embodiment, the regions obtained by performing multiple sub-segments on the relevant seed 311 can be used to determine the overlapping regions 313 on the current image 107. Subsequently, the overlapping regions 313 can be extracted from the current image 107 and considered as the target RoI 109 corresponding to the current image 107, such as... Figure 3G As shown. That is, Figure 3G The target RoI 109 is shown as extracted by performing the above steps on the current image 107.

[0080] In one embodiment, a comparison between the reference RoI and the target RoI 109 (e.g.) Figure 3H (As shown) helps to determine the RoI change between the reference image 103 and the current image 107 for the same entity.

[0081] Figure 4 A flowchart illustrating a method for automatically propagating segmented medical images according to some embodiments of the present disclosure is shown.

[0082] like Figure 4 As shown, method 400 includes one or more boxes that illustrate a method for using Figure 1 The image segmentation system 101 shown illustrates a method for automatically propagating segmentation in medical images. Method 400 can be described in the general context of computer-executable instructions. Typically, computer-executable instructions can include routines, programs, objects, components, data structures, procedures, modules, and functions that perform specific functions or implement specific abstract data types.

[0083] The order in which method 400 is described is not intended to be construed as limiting, and any number of described method blocks can be combined in any order to implement the method. Furthermore, individual blocks can be removed from the method without departing from the spirit and scope of the subject matter described herein. Moreover, the method can be implemented using any suitable hardware, software, firmware, or a combination thereof.

[0084] In block 401, method 400 includes determining one or more segmentation parameters 211 based on analysis of segmented regions of interest (RoIs) in reference image 103. In one embodiment, determining one or more segmentation parameters 211 may include, but is not limited to, determining one or more morphological characteristics of the segmented reference RoI 303, the location of the segmented reference RoI 303 within reference image 103, the intensity distribution of the segmented reference RoI 303, and the background intensity distribution of the background of the segmented reference RoI 303.

[0085] In one embodiment, the background intensity distribution can be determined using the following method:

[0086] 1. By determining multiple intensity distributions of multiple background pixels / points corresponding to quadrants relative to the center of the segmented reference RoI 303. In one embodiment, the multiple intensity distributions may be determined based on the amount of overlap between the pixel intensity of the corresponding background in the multiple background quadrants and the pixel intensity of the segmented reference RoI 303.

[0087] 2. Using the center of the segmented reference RoI 303 as the origin, the background pixels are divided into multiple background quadrants. Furthermore, the background threshold corresponding to each of the multiple background quadrants is determined based on the overlap between the pixel intensity of the corresponding background quadrant and the pixel intensity of the segmented reference RoI 303. Finally, the background intensity distribution corresponding to each of the multiple background quadrants can be determined based on the corresponding background threshold.

[0088] In one embodiment, the intensity distribution of the segmented reference RoI 303 can be determined by generating a histogram of the pixels in the segmented reference RoI 303 and determining one or more intensity ranges of the segmented reference RoI 303 based on the histogram and a predetermined threshold. In one embodiment, one or more intensity ranges can represent the intensity distribution of the segmented reference RoI 303.

[0089] In block 403, method 400 includes determining a plurality of reference points 213 corresponding to the segmented reference RoI 303 based on one or more morphological characteristics of the segmented reference RoI 303. In one embodiment, determining one or more morphological characteristics of the segmented reference RoI 303 may include, but is not limited to, determining the minor axis 303B and the longest diameter 303A of the segmented reference RoI 303. Furthermore, determining one or more morphological characteristics may include determining the size and shape of the segmented reference RoI 303 based on the length and coordinates of the minor axis 303B and the longest diameter 303A of the segmented reference RoI 303.

[0090] In block 405, method 400 includes generating a plurality of translation points 215 on the current image 107 by translating each of a plurality of reference points 213 onto the current image 107, in which the target RoI 109 to be segmented exists. In one embodiment, the reference image 103 and the current image 107 may be images of a single entity captured at different time periods.

[0091] In box 407, method 400 includes automatically selecting relevant seeds 311 in the current image 107 from a plurality of translation points 215 based on one or more segmentation parameters 211. In one embodiment, if the intensity of one or more translation points 215 falls within the intensity distribution of the segmented reference RoI 303, the relevant seeds 311 in the current image 107 can be automatically selected by selecting one or more translation points 215 from the plurality of translation points 215. Furthermore, the relevant seeds 311 in the current image can be automatically selected based on the following criteria:

[0092] 1. When the intensity of one or more translation points 215 does not fall within the background intensity distribution of the segmented reference RoI 303.

[0093] 2. When the intensity of at least a predetermined number of adjacent pixels of one or more translation points 215 falls within the intensity distribution of the segmented reference RoI 303.

[0094] 3. When the intensity of at least a predetermined number of adjacent pixels of one or more translation points 215 does not fall within the background intensity distribution range of the background.

[0095] 4. When the intensity of one or more translation points 215 is less than the intensity distribution range of the segmented reference RoI 303 and the background intensity distribution range of the background.

[0096] 5. When the intensity of one or more translation points 215 is greater than the highest intensity value of the intensity distribution of the segmented reference RoI 303 and the highest intensity value of the background intensity distribution range of the background.

[0097] In block 409, method 400 includes performing multiple sub-segments of a selected relevant seed 311 to estimate and segment a target RoI 109 in the current image 107. In one embodiment, the target RoI 109 is a propagated segmentation of the RoI of the reference image 103. In one embodiment, a prerequisite for performing segmentation of the target RoI 109 may be that the pixel intensity of the current image 107 must be quantitatively comparable to or normalized to the pixel intensity of the reference image 103.

[0098] Computer System

[0099] Figure 5 A block diagram of an exemplary computer system 500 for implementing embodiments consistent with this disclosure is shown. In one embodiment, the computer system 500 may be an image segmentation system 101 for automatically propagating segmentation in medical images. The computer system 500 may include a central processing unit (“CPU” or “processor”) 502. The processor 502 may include at least one data processor for executing program components for performing user- or system-generated business processes. Users may include people, patients, doctors and / or technicians, persons using the image segmentation system 101, etc. The processor 502 may include dedicated processing units such as an integrated system (bus) controller, a memory management control unit, a floating-point unit, a graphics processing unit, a digital signal processing unit, etc.

[0100] Processor 502 can be configured to communicate with one or more input / output (I / O) devices (511 and 512) via I / O interface 501. I / O interface 501 can employ communication protocols / methods such as, but not limited to, audio, analog, digital, stereo, IEEE-1394, serial bus, Universal Serial Bus (USB), infrared, PS / 2, BNC, coaxial, component, composite, digital video interface (DVI), high-definition multimedia interface (HDMI), radio frequency (RF) antenna, S-Video, video graphics array (VGA), IEEE 802.n / b / g / n / x, Bluetooth, cellular (e.g., Code Division Multiple Access (CDMA), High-Speed ​​Packet Access (HSPA+), Global System for Mobile Communications (GSM), Long Term Evolution (LTE), etc.). Using I / O interface 501, computer system 500 can communicate with one or more I / O devices 511 and 512.

[0101] In some embodiments, processor 502 may be configured to communicate with communication network 509 via network interface 503. Network interface 503 may communicate with communication network 509. Network interface 503 may employ connection protocols, including but not limited to direct connection, Ethernet (e.g., twisted pair 10 / 100 / 1000Base T), Transmission Control Protocol / Internet Protocol (TCP / IP), Token Ring, IEEE 802.11a / b / g / n / x, etc. Using network interface 503 and communication network 509, computer system 500 may communicate with reference database 105 to receive segmented regions of interest (RoI) of reference image 103. Furthermore, communication network 509 may be used to receive reference image 103 and current image 107 from one or more sources such as an X-ray scanner.

[0102] Communication network 509 can be implemented as one of several types of networks, such as an intranet or local area network (LAN) within an organization. Communication network 509 can be a private network or a shared network, representing an association of several types of networks that communicate with each other using multiple protocols (e.g., Hypertext Transfer Protocol (HTTP), Transmission Control Protocol / Internet Protocol (TCP / IP), Wireless Application Protocol (WAP), etc.). Furthermore, communication network 509 can include various network devices, including routers, bridges, servers, computing devices, storage devices, etc.

[0103] In some embodiments, the processor 502 may be configured to connect to the memory 505 (e.g., via the storage interface 504) via the storage interface 504. Figure 5 The RAM 513 and ROM 514 shown communicate with each other. Storage interface 504 can connect to memory 505, including but not limited to memory drives, removable disk drives, etc. Memory 505 uses connection protocols such as Serial Advanced Technology Attachment (SATA), Integrated Drive Electronics (IDE), IEEE-1394, Universal Serial Bus (USB), Fibre Channel, Small Computer System Interface (SCSI), etc. Memory drives may also include magnetic drums, disk drives, magneto-optical drives, optical drives, Redundant Array of Independent Disks (RAID), solid-state storage devices, solid-state drives, etc.

[0104] The memory 505 may store a collection of program or database components, including but not limited to user / application 506, operating system 507, web browser 508, etc. In some embodiments, the computer system 500 may store user / application data 506, such as data, variables, records, etc., described in this disclosure. Such a database can be implemented as a fault-tolerant, relational, scalable, and secure database, such as Oracle or Sybase.

[0105] Operating system 507 facilitates resource management and operation of computer system 500. Examples of operating systems include, but are not limited to, Apple Macintosh OS X, UNIX, Unix-like system distributions (e.g., Berkeley Software Distribution (BSD), FreeBSD, NetBSD, OpenBSD, etc.), Linux distributions (e.g., Red Hat, Ubuntu, K-Ubuntu, etc.), IBM OS / 2, Microsoft Windows (XP, Vista / 7 / 8, etc.), Apple iOS, Google Android, BlackBerry OS, etc.

[0106] User interfaces can facilitate the display, execution, interaction, manipulation, or operation of program components using text or graphical tools. For example, a user interface can provide computer interaction elements, such as cursors, icons, checkboxes, menus, windows, and widgets, on a display system operatively connected to computer system 500. Graphical user interfaces (GUIs) can be used, including but not limited to Apple Macintosh Aqua, IBM OS / 2, Microsoft Windows (e.g., Aero, Metro), Unix X-Windows, and web interface libraries (e.g., ActiveX, Java, JavaScript, AJAX, HTML, Adobe Flash, etc.).

[0107] Furthermore, one or more computer-readable storage media can be used to implement embodiments consistent with the present invention. A computer-readable storage medium refers to any type of physical memory capable of storing processor-readable information or data. Therefore, a computer-readable storage medium can store instructions for execution by one or more processors, including instructions for causing the processor(s)(s) to perform steps or phases consistent with the embodiments described herein. The term "computer-readable medium" should be understood to include tangible articles but exclude carrier waves and transient signals, i.e., non-transient. Examples include random access memory (RAM), read-only memory (ROM), volatile memory, non-volatile memory, hard disk drives, optical disc (CD) ROMs, digital video discs (DVDs), flash drives, magnetic disks, and any other known physical storage media.

[0108] The terms “an embodiment,” “embodiment,” “multiple embodiments,” “the embodiment,” “the multiple embodiments,” “one or more embodiments,” “some embodiments,” and “one embodiment” refer to “one or more (but not all) embodiments of the invention,” unless otherwise expressly stated.

[0109] Unless otherwise expressly stated, the terms "including," "comprising," "having," and their variations mean "including but not limited to." Unless otherwise expressly stated, the list of items does not imply that any or all items are mutually exclusive.

[0110] Unless otherwise expressly stated, the terms "a," "an," and "the" refer to "one or more." The description of embodiments having multiple components communicating with each other does not imply that all of these components are necessary. Rather, a variety of optional components are described to illustrate various possible embodiments of the invention.

[0111] When a single device or item is described herein, it is apparent that more than one device / item (whether or not they cooperate) can be used in place of a single device / item. Similarly, when more than one device or item is described herein (whether or not they cooperate), it is apparent that a single device / item can be used in place of more than one device or item, or a different number of devices / items can be used in place of the number of devices or programs shown. The functionality and / or features of a device can alternatively be embodied by one or more other devices that are not explicitly described as having such functionality / features. Therefore, other embodiments of the invention do not necessarily need to include the device itself.

[0112] Finally, the language used in this specification has been chosen primarily for readability and guidance purposes and may not have been chosen to describe or limit the subject matter of the invention. Therefore, the scope of the invention is intended not to be limited by this detailed description, but rather by any claims published on the application based herein. Thus, the embodiments of the invention are intended to illustrate, rather than limit, the scope of the invention, which is set forth in the appended claims.

[0113] While various aspects and embodiments have been disclosed herein, other aspects and embodiments will be readily apparent to those skilled in the art. The aspects and embodiments disclosed herein are for illustrative purposes and not intended to be limiting; the true scope and spirit are indicated by the appended claims.

[0114] Figure Labels

[0115]

[0116]

[0117]

Claims

1. A method for automatically propagating segmentation in medical images, the method comprising: Based on the analysis of segmented regions of interest (RoIs) in a reference image, one or more segmentation parameters are determined, wherein determining the one or more segmentation parameters includes determining one or more morphological characteristics of the segmented reference RoI, and wherein determining the one or more morphological characteristics of the segmented reference RoI includes determining the minor axis and the longest diameter of the segmented reference RoI. Based on one or more morphological characteristics of the segmented reference RoI, determine multiple reference points corresponding to the segmented reference RoI; Multiple translation points are generated on the current image by translating each of the plurality of reference points onto the current image, where there is a target RoI that needs to be segmented; Based on the one or more segmentation parameters, a relevant seed in the current image is automatically selected from the plurality of translation points, wherein the selection includes selecting the one or more translation points from the plurality of translation points if the intensity of one or more translation points falls within the intensity distribution of the segmented reference RoI. as well as Perform multiple sub-segments of the selected relevant seed to estimate and segment the target RoI in the current image, wherein the target RoI is the propagation segmentation of the segmented RoI in the reference image, and wherein the pixel intensity of the current image is quantitatively comparable to or can be normalized to the pixel intensity of the reference image.

2. The method according to claim 1, wherein the reference image and the current image are images of a single entity captured at different time periods.

3. The method of claim 1, wherein determining the one or more segmentation parameters further comprises determining at least one of the following: the location of the segmented reference RoI within the reference image, the intensity distribution of the segmented reference RoI, and the background intensity distribution of the background of the segmented reference RoI.

4. The method of claim 3, wherein determining the one or more morphological characteristics of the segmented reference RoI further comprises: The size and shape of the segmented reference RoI are determined based on the length and coordinates of the minor axis and the longest diameter of the segmented reference RoI.

5. The method of claim 3, wherein determining the background intensity distribution of the background comprises: Using the center of the segmented reference RoI as the origin, the background pixels are grouped into multiple background quadrants. Based on the overlap between the pixel intensity of the corresponding background in the multiple background quadrants and the pixel intensity of the segmented reference RoI, multiple intensity distributions of multiple background pixels / points corresponding to the quadrants relative to the center of the segmented reference RoI are determined.

6. The method of claim 5, wherein determining the background intensity distribution of the background comprises: A background threshold corresponding to each of the plurality of background quadrants is determined based on the amount of overlap between the pixel intensity of the corresponding background quadrant and the pixel intensity of the segmented reference RoI. as well as Based on the corresponding background threshold, the background intensity distribution corresponding to each of the plurality of background quadrants is determined.

7. The method of claim 3, wherein determining the intensity distribution of the segmented reference RoI comprises: Generate a histogram of pixels in the segmented reference RoI; as well as One or more intensity ranges of the segmented reference RoI are determined based on the histogram and a predetermined threshold, wherein the one or more intensity ranges represent the intensity distribution of the segmented reference RoI.

8. The method of claim 1, wherein automatically selecting the relevant seed in the current image comprises: If the intensity of one or more translation points does not fall within the background intensity distribution of the segmented reference RoI, then one or more translation points are selected.

9. The method of claim 8, wherein automatically selecting the relevant seed in the current image comprises: If the intensity of at least a predetermined number of neighboring pixels of the one or more translation points falls within the intensity distribution of the segmented reference RoI, then the one or more translation points are selected.

10. The method of claim 9, wherein automatically selecting the relevant seed in the current image comprises: If the intensity of at least a predetermined number of adjacent pixels of the one or more translation points does not fall within the background intensity distribution range of the background, then the one or more translation points are selected.

11. The method of claim 8, wherein automatically selecting the relevant seed in the current image comprises: Determine whether the intensity of the one or more translation points is less than the intensity distribution range of the segmented reference RoI and the background intensity distribution range of the background.

12. The method of claim 8, wherein automatically selecting the relevant seed in the current image comprises: Determine whether the intensity of the one or more translation points is greater than the highest intensity value of the intensity distribution of the segmented reference RoI and the highest intensity value of the background intensity distribution range of the background.

13. An image segmentation system for automatically propagating segmentation in medical images, the image segmentation system comprising: processor; as well as A memory, communicatively coupled to the processor, wherein the memory stores processor-executable instructions that, when executed, cause the processor to: Based on the analysis of segmented regions of interest (RoI) in a reference image, one or more segmentation parameters are determined, wherein determining the one or more segmentation parameters includes determining one or more morphological characteristics of the segmented reference RoI, and wherein determining the one or more morphological characteristics of the segmented reference RoI includes determining the minor axis and the longest diameter of the segmented reference RoI. Based on one or more morphological characteristics of the segmented reference RoI, determine multiple reference points corresponding to the segmented reference RoI; Multiple translation points are generated on the current image by translating each of the plurality of reference points onto the current image, where there is a target RoI that needs to be segmented; Based on the one or more segmentation parameters, a relevant seed in the current image is automatically selected from the plurality of translation points, wherein the selection includes selecting the one or more translation points from the plurality of translation points if the intensity of one or more translation points falls within the intensity distribution of the segmented reference RoI. as well as Perform multiple sub-segments of the selected relevant seed to estimate and segment the target RoI in the current image, wherein the target RoI is the propagation segmentation of the segmented RoI in the reference image, and wherein the pixel intensity of the current image is quantitatively comparable to or can be normalized to the pixel intensity of the reference image.

14. The image segmentation system of claim 13, wherein the reference image and the current image are images of a single entity captured at different time periods.

Images