Switching between neural networks based on scout scan analysis
Patent Information
- Application Number
- JP2023532492
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2020-11-30
- Filing Date
- 2021-11-22
- Publication Date
- 2024-11-29
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Conventional neural networks trained for specific anatomical structures struggle when processing images containing multiple anatomical regions, leading to suboptimal performance due to limited network capacity and execution time, especially in CT scans.
A method that applies separate processing routines to different anatomical structures within a single image, using scout scans to determine where to switch between these routines, and employs machine learning algorithms like CNNs trained on specific anatomical regions for each segment.
Improves image quality by effectively processing distinct anatomical regions in a single image, enhancing clarity and reducing noise through tailored processing routines for each anatomical structure.
Smart Images

Figure 00000000_0000_ABST
Abstract
Description
Technical Field
[0001]
[0001] This disclosure generally relates to systems and methods for processing images, such as medical images, using trained neural networks.
Background Art
[0002]
[0002] Conventionally, acquiring images through standard imaging modalities such as computer tomography (CT) scans results in image artifacts and noise incorporated in such images. Thus, the images are generally processed using noise removal algorithms. Such noise removal algorithms are typically associated with specific anatomical structures, designed to enhance specific image features, and, in the case of learning algorithms, trained to enhance specific image features.
[0003]
[0003] Thus, when processing a CT scan, a reconstruction filter is used to enhance specific image features such as, for example, sharpness or soft tissue contrast. As an example, when reconstructing a head image, the system uses a filter designed to enhance soft tissue contrast in brain tissue, and a different filter is used to reconstruct a body image. In contrast, the filter used to reconstruct a body image is sharper.
[0004]
[0004] Reconstructed images generally contain noise and various artifacts of the reconstruction process, and various denoising algorithms in particular are applied to improve image quality. Therefore, in the context of training algorithms for denoising, such as convolutional neural networks (CNNs) for processing CT scans, different filters have been used during image reconstruction to enhance image features, such as sharpness, or to suppress different types of noise. Denoising must be performed by techniques that preserve the image features enhanced by the filters used for reconstruction. Therefore, training algorithms such as CNNs are trained for specific types of anatomical structures and specific reconstruction filters.
[0005]
[0005] Generally due to limited network capacity and the resulting execution time, when training a CNN to denoise low-dose CT images, for example, the conventional technique trains different networks for different anatomical regions and reconstruction filters designed for such regions. This is particularly beneficial if the filters are designed to capture highly characteristic image features in specific anatomical structures. [Overview of the project] [Problems that the invention aims to solve]
[0006]
[0006] However, a single scan, such as a head and neck scan, covers multiple different anatomical regions and corresponding anatomical structures. When such images are reconstructed using a head filter, the natural choice is to use a network trained on the same filter, but such a network typically has not observed any anatomical structures other than the head during training. Especially when the filter is designed for a specific anatomical structure, training one network on one filter using various anatomical structures typically results in suboptimal performance in other regions.
[0007]
[0007] Furthermore, due to limited available network capacity and the resulting execution time, it is generally not feasible to train a single network based on various anatomical structures reconstructed using the same filter. [Means for solving the problem]
[0008]
[0008] Systems and methods for processing medical images are described in which multiple anatomical structures appear in a single image, and separate processing routines are applied to each of such anatomical structures. For example, separate machine learning methods are used to process different parts of a single image containing multiple such anatomical structures.
[0009]
[0009] In addition to applying separate processing routines to different anatomical structures appearing in a single image, the systems and methods described further determine where to switch between such processing techniques in such images. Thus, such determination is made based on scout scans available independently of the primary medical image being processed. Thus, multiple anatomical structures are defined in preliminary images, such as scout scans, using machine learning or prior art. Such definitions are then applied to primary images that are different from the scout scans, the primary images of which are of the same object.
[0010]
[0010] Accordingly, a method for processing medical images is provided, wherein the method receives a first image, receives a second image different from the first image, and the second image includes the same object as the first image. The method then identifies a plurality of different anatomical structures in the first image, then defines a plurality of image segments in the second image based on the locations of the different anatomical structures identified in the first image, wherein the first image segment of the plurality of image segments includes a first anatomical structure of the plurality of different anatomical structures, and the second image segment of the plurality of image segments includes a second anatomical structure of the plurality of different anatomical structures.
[0011]
[0011] The method then applies a first processing routine associated with a first anatomical structure to the first image segment in order to obtain the processed first image segment, applies a second processing routine associated with a second anatomical structure to the second image segment in order to obtain the processed second image segment, and then outputs a processed second image that includes both the processed first image segment and the processed second image segment.
[0012]
[0012] In some embodiments, the first or second image is selected from imaging modalities including computed tomography (CT), magnetic resonance imaging (MRI), positron emission tomography (PET), single-photon emission computed tomography (SPECT), digital X-ray radiogrammetry (DXR), or imaging modalities including fluorescence fluoroscopy sequences in image-assisted therapy (IGT) imaging.
[0013]
[0013] For example, the second image is a CT image or a primary image for a different imaging modality, and the first image is a scout scan relative to the primary image. Therefore, if the second image is a CT image, the scout scan is acquired using a lower radiation dose than the second image.
[0014]
[0014] In some embodiments, multiple different anatomical structures in the first image are identified before the second image is received, and therefore their identification information is available when the second image is received. In some embodiments, multiple different anatomical structures in the first image are identified when the second image is received.
[0015]
[0015] In some embodiments, the first image segment excludes the second anatomical structure, and the second image segment excludes the first anatomical structure. In some embodiments, the first and second image segments are linearly perspectived, so that the first image segment includes the entire top width of the second image, and the second image segment includes the entire bottom width of the second image.
[0016]
[0016] The first or second anatomical structure is selected from the head, neck, upper body, abdomen, pelvic region, lower body, and legs.
[0017]
[0017] In some embodiments, the first processing routine is a first machine learning algorithm associated with a first anatomical structure, and the second processing routine is a second machine learning algorithm associated with a second anatomical structure.
[0018]
[0018] Further provided is an imaging system comprising a memory for storing multiple instructions, an imaging unit, and a processor circuit coupled to the memory and configured to execute instructions for acquiring an image and carrying out the method described above. A second image is received from the imaging unit in such a system.
[0019]
[0019] A non-temporary computer-readable medium is further provided which stores a program for processing medical images, including instructions for performing the method described above. Such a method is implemented, for example, in the context of the system described. [Brief explanation of the drawing]
[0020] [Figure 1]
[0020] Schematic diagram of a system according to one embodiment of the present disclosure. [Figure 2]
[0021] It is a figure which shows the imaging device by one Embodiment of this indication. [Figure 3A]
[0022] FIGS. 3A and 3B are diagrams showing the same image processed using two separate processing routines. [Figure 3B] FIGS. 3A and 3B are diagrams showing the same image processed using two separate processing routines. [Figure 4]
[0023] It is a flowchart which shows the method for processing a medical image according to this indication. [[ID=第十九]] [Figure 5]
[0024] It is a figure which shows the parsing of the image by this indication.
Mode for Carrying Out the Invention
[0021]
[0025] The description of exemplary embodiments in accordance with the principles of the present invention is intended to be read with reference to the accompanying drawings, which are considered to be a part of the overall description. In the description of the embodiments of the present invention disclosed herein, all references to direction or orientation are for convenience of explanation only and are not intended to limit the scope of the invention in any way. For example, relative terms such as "lower," "upper," "horizontal," "vertical," "above," "below," "on," "under," "top," and "bottom," and their derivatives (e.g., "horizontally," "downwardly," "upwardly," etc.) should be construed as representing the orientation when so described or as shown in the drawings being described. These relative terms are for convenience of explanation only and do not require the device to be configured or operated in a particular orientation unless so explicitly stated. Terms such as "mounted," "attached," "connected," "coupled," "interconnected," and the like, unless otherwise specified, represent a relationship where structures are fixed or attached to each other directly, indirectly through intervening structures, and in both movable and immovable mounting or relationships. Further, the features and advantages of the present invention are shown by reference to exemplary embodiments. Accordingly, the present invention should not be explicitly limited to such exemplary embodiments showing any possible non-limiting combinations of features that exist alone or in other combinations of features, and the scope of the present invention is defined by the claims appended hereto.
[0022]
[0026] This disclosure describes one or more best modes of implementing the present invention as presently contemplated. This description is not intended to be understood in a limiting sense, but rather provides examples of the present invention presented for illustrative purposes only by reference to the accompanying drawings to advise those skilled in the art of the advantages and configurations of the present invention. In the various figures of the drawings, like reference numerals refer to like or similar parts.
[0023]
[0027] It is important to note that the embodiments disclosed are merely examples of many useful applications of the innovative teachings herein. In general, the descriptions in this specification do not necessarily limit any of the disclosures contained in the various claims. Furthermore, some descriptions may apply to some features of the invention but not to others. In general, unless otherwise stated, one element may be made up of multiple elements and vice versa, so as not to lose generality.
[0024]
[0028] Generally, one image is reconstructed using one filter, and therefore the filter is selected to suit the subject of the image being reconstructed. Thus, when reconstructing a head image, the system uses a filter designed to enhance soft tissue contrast in brain tissue. However, when reconstructing a body image, a different filter, and sometimes a sharper filter, is used.
[0025]
[0029] Therefore, in the context of computed tomography (CT)-based medical imaging, for example, different image processors, such as machine learning algorithms in the form of convolutional neural networks (CNNs), are used to process images. These image processors, in the case of machine learning algorithms, are trained based on corresponding different anatomical regions and structures in the context of reconstruction filters.
[0026]
[0030] For example, in medical imaging other than CT, such as magnetic resonance imaging (MRI) or positron emission tomography (PET), methods other than the use of such reconstruction filters are used to reconstruct and process images. Rather, different reconstruction algorithms are used depending on the type of scan or data acquisition and the desired image features for a particular scan. The reconstruction algorithm is then adjusted to acquire specific image features. For example, in common iterative MRI or PET reconstructions, the number of iterations is carefully selected and a regularization term is added.
[0027]
[0031] However, a single scan, such as a head and neck scan, covers different anatomical regions. Therefore, a system and method are provided for effectively handling distinct anatomical regions in a single image.
[0028]
[0032] Figure 1 is a schematic diagram of system 100 according to one embodiment of the present disclosure. As shown, system 100 typically includes a processing device 110 and an imaging device 120.
[0029]
[0033] The processing device 110 applies a processing routine to the received image. The processing device 110 includes a memory 113 and a processor circuit 111. The memory 113 stores multiple instructions. The processor circuit 111 is coupled to the memory 113 and configured to execute the instructions. The instructions stored in the memory 113 include data related to processing routines and multiple machine learning algorithms, such as various convolutional neural networks for processing images.
[0030]
[0034] The processing device 110 further includes an input 115 and an output 117. The input 115 receives information, such as an image, from the imaging device 120. The output 117 outputs information to a user or user interface device. The output includes a monitor or display.
[0031]
[0035] In some embodiments, the processing device 110 is directly related to the imaging device 120. In alternative embodiments, the processing device 110 is independent of the imaging device 120, and therefore the processing device 110 receives images for processing via a network or other interface at input 115.
[0032]
[0036] In some embodiments, the imaging device 120 includes an image data processing device and a spectral or conventional CT scanning unit for generating CT projection data when scanning an object (e.g., a patient).
[0033]
[0037] Figure 2 shows an exemplary imaging device according to one embodiment of the present disclosure. A CT imaging device is shown, and the following description is in the context of CT images, but it will be understood that similar methods are applicable in the context of other imaging devices, and that the images to which these methods are applied are obtained by a wide variety of techniques.
[0034]
[0038] In the imaging device according to embodiments of the present disclosure, the CT scanning unit is adapted to perform multiple axial and / or helical scans of an object to generate CT projection data. In the imaging device according to embodiments of the present disclosure, the CT scanning unit comprises an energy-resolved photon count image detector. The CT scanning unit includes a radiation source that emits radiation to traverse the object when acquiring projection data.
[0035]
[0039] Furthermore, in the imaging device according to the embodiments of the present disclosure, the CT scanning unit performs a scout scan distinct from the primary scan, thereby generating separate images related to the scout scan and the primary scan, which are different but include the same object.
[0036]
[0040] In the example shown in Figure 2, a CT scanning unit 200, such as a computed tomography (CT) scanner, includes a fixed gantry 202 and a rotating gantry 204 rotatably supported by the fixed gantry 202. The rotating gantry 204 rotates around its long axis with respect to the examination area 206 relative to the object when acquiring projection data. The CT scanning unit 200 includes a support 207 configured for supporting a patient within the examination area 206 and for passing the patient through the examination area during the imaging process.
[0037]
[0041] The CT scanning unit 200 includes a radiation source 208, such as an X-ray tube, which is supported by a rotating gantry 204 and configured to rotate with the rotating gantry 204. The radiation source 208 includes an anode and a cathode. A source voltage applied between the anode and the cathode accelerates electrons from the cathode to the anode. The flow of electrons provides a flow of current from the cathode to the anode to generate radiation, for example, across the inspection area 206.
[0038]
[0042] The CT scanning unit 200 includes a detector 210. The detector 210 is positioned in an arc at an angle to the radiation source 208, opposite the inspection area 206. The detector 210 includes a one-dimensional or two-dimensional array of pixels, such as direct-convert detector pixels. The detector 210 is adapted to detect radiation traversing the inspection area and to generate a signal indicating its energy.
[0039]
[0043] The CT scanning unit 200 further includes generators 211 and 213. Generator 211 generates tomography projection data 209 based on signals from detector 210. Generator 213 receives the tomography projection data 209 and generates a raw image 311 of the object based on the tomography projection data 209. The initial image 311 is a scout scan or primary scan and is input to input 115 of processing device 100.
[0040]
[0044] Figures 3A and 3B show the same image processed using two separate processing routines that result in differently processed images. The raw image 311 is typically received at input 115 of the processing device 110. Image 311 is then processed according to the processing routine. The processing routine includes, for example, the reconstruction of image 311 using filters that implement machine learning algorithms, such as a convolutional neural network (CNN). Such machine learning algorithms are typically trained on sample images similar to those to which such machine learning algorithms are ultimately applied. Thus, the processing routine is specific to a particular anatomical region or the corresponding reconstruction filter.
[0041]
[0045] Therefore, the processing routine involves applying reconstruction filters designed for specific anatomical regions, and corresponding CNNs trained on those specific anatomical regions, such as a head network trained on a head filter. Head filters are specifically designed to enhance soft tissue contrast in brain tissue and are then trained using head images, while body filters are alternatively designed to provide sharper results and are paired with CNNs trained on body images.
[0042]
[0046] However, when used, scans acquired using the imaging system 120 include multiple anatomical regions. Considering this, Figure 3A shows an image acquired in a head and neck scan after it has been reconstructed using a head filter and then denoised using a network trained on the anatomical structures of the head with that head filter. In contrast, Figure 3B shows the same image as in Figure 3A, after it has been reconstructed using a body filter and then denoised using a network trained on the anatomical structures of the body.
[0043]
[0047] Despite the fact that each image was denoised using a CNN trained based on the corresponding filter, the resulting processed image shown in Figure 3A is highly noisy and of lower quality than that shown in Figure 3B. This is because the head network typically does not observe any anatomical structures other than the head during training. A similar phenomenon is observed when the body network and filters are applied to head region scans.
[0044]
[0048] Therefore, when an image containing multiple anatomical structures is provided, existing systems and methods must determine which of several possible processing routines should be applied. According to this disclosure, a method is proposed in which separate image segments within a single image are processed using separate processing routines. Thus, separate filters using different CNNs trained with such filters are applied to different image segments associated with corresponding anatomical structures.
[0045]
[0049] A method according to one embodiment of the present invention further determines where it is best to switch between such different processing routines during processing by evaluating a first image, such as a scout scan, which contains the same target object as the primary image to be processed.
[0046]
[0050] Figure 4 is a flowchart showing a method for processing medical images in accordance with the present disclosure. As shown, the method involves applying separate networks trained on different anatomical structures to different parts of an image.
[0047]
[0051] The method first involves receiving a first image (at 400) from an imaging device 120 at input 115 of a processing device 110. The first image is, for example, a scout scan performed using the same imaging device 120 as the later primary image. The received scout scan is, for example, a planning image used to plan the later primary image, and in the case of a CT scan, is acquired using a lower radiation dose than the later primary image. It is understood that the image is received in a wide variety of formats, including the image format directly output by the imaging device 120 and partially processed data associated with such an image. Furthermore, the first image received (at 400), and any further images described below, may take different forms and be of different file types. Thus, the first image is received in a raw data format that can be processed to identify anatomical structures as described below, regardless of whether such data can be fully reconstructed to generate the corresponding first image.
[0048]
[0052] The method then identifies several different anatomical structures in the first image (in 410). These anatomical structures include, for example, the head, neck, upper body, abdomen, pelvic region, lower body, and legs. The first image may contain two such anatomical structures, or the first image may contain several such structures, for example, in the context of a whole-body scan.
[0049]
[0053] After receiving a first image (at 400), the method receives a second image (at 420) that is different from the first image. The second image is a primary image, and although the second image is different from the first image, the second image typically contains the same object as the first image. Therefore, if the first image acquired at 400 contains multiple anatomical structures identified at 410, the second image acquired at 420 typically contains the same multiple anatomical structures.
[0050]
[0054] The first or second image is acquired using various imaging modalities. Such images are then acquired, for example, by computed tomography (CT) via the imaging device 120 described above, or by X-ray imaging, including magnetic resonance imaging (MRI), positron emission tomography (PET), single-photon emission computed tomography (SPECT), digital X-ray radiogrammetry (DXR), or by fluorescence fluoroscopy sequences in image-assisted therapeutic (IGT) imaging. When any of these imaging modalities are used, both the first and second images may be acquired using the same imaging modality, or the first image may be acquired using a different imaging modality than that of the second image.
[0051]
[0055] For example, the first image is a scout scan relative to a CT scan and is acquired using CT. However, the first image is then acquired using a lower radiation dose than the second image and is then used as a planning image relative to the CT scan, which is then acquired using a higher radiation dose. In other embodiments, the scout scan is acquired using a different imaging modality, and therefore the scout scan is faster, simpler, less invasive, lower cost, or otherwise more convenient. For example, the scout scan is acquired using an X-ray process, followed by a primary scan acquired using CT. Similarly, a CT scan is used as a scout scan, followed by a PET or SPECT as the primary image. In some such embodiments, the CT scan is used to improve PET or SPECT reconstruction in addition to the use of such CT images to define different anatomical structures in the image. Either the first or second image is a three-dimensional image, or the first or second image is a two-dimensional image. Furthermore, the first image is acquired with lower resolution, or lower contrast or color density, compared to the second image.
[0052]
[0056] Furthermore, in some embodiments, the identification of multiple different anatomical structures in the first image (in 410) is performed before the reception of the second image (in 420) from the imaging unit 120. In such embodiments, the method then ensures that the anatomical structures are defined before the reception of the second image, enabling more efficient processing of the corresponding second image. Similarly, in some embodiments, multiple different anatomical structures in the first image are identified when the second image (in 420) is received. These approaches allow the system 100 to implement the described methods to improve the system's processing efficiency by completing all processing related to identifying different anatomical structures in the first image before or while the second image is being received. Thus, after the second image is partially or completely received, the second image is immediately processed based on the identification information obtained from the first image.
[0053]
[0057] In some embodiments, the processing of the first image does not need to be completed before the processing of the second image begins. Thus, after anatomical structures are identified in the first image and the corresponding portion of the second image is received, the relevant portion of the second image is processed according to the following steps.
[0054]
[0058] The method then defines multiple image segments in a second image based on the locations of different anatomical structures identified in the first image (in 430). Therefore, if the first image is specified to include a first anatomical structure, such as the head, and a second anatomical structure, such as the neck (in 410), the method defines a first image segment containing the first anatomical structure and a second image segment containing the second anatomical structure.
[0055]
[0059] In embodiments where anatomical structures are defined (in 410) before the reception of the second image (in 420), the image segment (in 430) is defined immediately after the reception of the second image. In some embodiments, such anatomical structures are identified and corresponding image segments are defined when the image is acquired by the imaging device 120 (in 420), when the image is received by the processing unit 110, before the entire second image is received, or while the entire second image is being received. For example, if a whole-body scan is performed after a scout scan, the image segments corresponding to the first anatomical structures being scanned are obtained and processed as described below before the completion of the scan that generates the second image.
[0056]
[0060] After multiple image segments are defined (in 430), different processing routines are applied to each image segment. Thus, a first processing routine is applied to a first image segment (in 440), and the first processing routine is related to a first anatomical structure contained in the first image segment.
[0057]
[0061] Separately, a second processing routine is applied to the second image segment (at 450), and the second processing routine is related to a second anatomical structure contained in the second image segment. The first and second anatomical structures are different anatomical structures, and the processing routines applied to each are different from each other.
[0058]
[0062] As described above, the embodiments shown describe the initial definition of multiple image segments (in 430) and the subsequent application of processing routines (in 440, 450), however, in some embodiments, one of the multiple image segments is defined before the complete second image is received. In such scenarios, the first image segment is defined (in 430), and processing begins (in 440) before the complete second image is received.
[0059]
[0063] The processing routines described are generally software routines designed to improve the quality or clarity of a second image. Such routines may be, for example, denoising routines or routines that remove artifacts of the imaging modality used. The processing routines are machine learning algorithms, each of which is trained on a corresponding anatomical structure. Thus, the first processing routine applied in 440 is a first machine learning algorithm, such as a CNN trained on a first anatomical structure. Next, the second processing routine applied in 450 is a second machine learning algorithm, such as a CNN trained on a second anatomical structure.
[0060]
[0064] The processing routine further includes reconstructing the relevant image segment using appropriate filters designed for the corresponding anatomical structure. Thus, if the first anatomical structure is the patient's head, the first processing routine applied in 440 is the reconstruction of the corresponding first image segment using a head filter, followed by denoising using a corresponding head CNN trained on the head filter. Similarly, if the second anatomical structure is the patient's neck or torso, the second processing routine applied in 450 is the reconstruction of the corresponding second image segment using a body filter, followed by denoising using a neck or torso-related CNN trained on the same body filter.
[0061]
[0065] In some embodiments, the same reconstruction filter is used for multiple image segments, but different CNNs trained on corresponding anatomical structures are applied to different image segments. While CNNs are described, it is understood that alternative algorithmic architectures, including various machine learning algorithms, may be used similarly.
[0062]
[0066] After the defined first and second image segments are processed (at 440 and 450, respectively), the processed second image is output (at 460). The processed second image includes the processed first image segment and the processed second image segment. If the first and second image segments have been parsed into separate images for processing purposes, they are typically recombined into a single image before being output.
[0063]
[0067] In the embodiments shown, the method identifies multiple anatomical structures in a first image (in 410). This is done by using a processing routine, such as applying a machine learning algorithm to identify anatomical structures. Anatomical structures are identified, for example, by identifying labels in the image associated with such structures. Alternatively, such anatomical structures are manually identified before image parsing to corresponding image segments. Such manual identification is done by an engineer identifying structures in a scout scan via a user interface before, during, or after acquiring a primary scan.
[0064]
[0068] In some embodiments, the identification of anatomical structures is performed in the same image being processed. In such embodiments, rather than identifying anatomical structures in the first image and then processing the second image, the first analysis is performed in the image in which the anatomical structures and corresponding image segments are identified, followed by a second analysis in which the image segments are processed in a different manner.
[0065]
[0069] Figure 5 shows image parsing 500 according to the present disclosure. As shown, image 500 includes the head 510, neck 520, and upper torso 530 of a patient being scanned. Typically, the method then identifies distinct anatomical structures associated with each of the head 510, neck 520, and upper torso 530, and then defines a first image segment 540 associated with a first anatomical structure, such as the head 510, and a second image segment 550 associated with a second anatomical structure, such as the upper torso 530.
[0066]
[0070] Two image segments 540 and 550 are shown, but a third separate image segment is similarly defined to enclose the neck 520. Alternatively, in the shown embodiment, the second image segment 550 is defined to include both the neck 520 and the upper torso 530, and to process both using a single processing routine suitable for both.
[0067]
[0071] To facilitate understanding of the method according to one embodiment of the present invention, it should be noted that Figure 5 is provided as an example of a ScoutScan. While image segments 540 and 550 are shown in ScoutScan image 500, it will be understood in use that such image segments are defined in the primary image received after the ScoutScan has been received.
[0068]
[0072] As shown, the first image segment 540 and the second image segment 550 are linearly oriented, so that the first image segment 540 includes the entire top width of the image 500 and the second image segment 550 includes the entire bottom width of the image.
[0069]
[0073] Furthermore, as shown, the first image segment 540 is specified to exclude a second anatomical structure, namely the upper torso 530, and the second image segment 550 is specified to exclude the first anatomical structure, namely the head 510. This approach prevents one part of the image from being processed separately using the first and second processing routines.
[0070]
[0074] Alternatively, a portion of the image is defined as part of both the first image segment 540 and the second image segment 550. In such an embodiment, the results of the first and second processing routines are merged (in 460) before outputting the processed second image. Such merging is done, for example, by averaging the resulting overlapping image segments.
[0071]
[0075] Although the methods described herein are primarily described in the context of CT scan images, a wide range of imaging techniques, including various medical imaging techniques, are envisioned, and it is understood that images generated using a broad range of imaging techniques can be effectively denoised or processed in other ways using the methods described herein.
[0072]
[0076] The methods described herein may be implemented in a computer, in special hardware, or a combination of both, as a computer-implemented method. Executable code for the methods described herein may be stored in a computer program product. Examples of computer program products include memory devices, optical storage devices, integrated circuits, servers, online software, and the like. Preferably, the computer program product includes non-temporary program code stored on a computer-readable medium for performing the methods described herein when the program product is executed on a computer. In embodiments, the computer program includes computer program code adapted to perform all the steps of the methods described herein when the computer program is executed on a computer. The computer program is embodied on a computer-readable medium.
[0073]
[0077] While this disclosure is described in some length and detail in relation to several described embodiments, it is not intended that this disclosure be limited to any of such details or embodiments or any particular embodiment. Rather, this disclosure should be interpreted in light of the appended claims to provide the broadest possible interpretation of the appended claims in consideration of the prior art, and thereby effectively encompass the scope intended by this disclosure.
[0074]
[0078] All examples and conditions expressed herein are intended for educational purposes to assist the reader in understanding the principles of this disclosure and the concepts to which the inventors contribute to the advancement of the art, and are construed as not being limited to such explicitly stated examples and conditions. Furthermore, all descriptions herein describing the principles, aspects, and embodiments of this disclosure, as well as specific examples thereof, are intended to encompass both structural and functional equivalents thereof. Moreover, such equivalents are intended to include both currently known equivalents and equivalents to be developed in the future, i.e., any elements developed that perform the same function regardless of their structure.
Claims
1. receiving a first image; receiving a second image different from the first image, the second image including the same target object as the first image; identifying a plurality of different anatomical structures in the first image; defining a plurality of image segments in the second image based on locations of the distinct anatomical structures identified in the first image, such that a first image segment of the plurality of image segments includes a first anatomical structure of the plurality of distinct anatomical structures and such that a second image segment of the plurality of image segments includes a second anatomical structure of the plurality of distinct anatomical structures; applying a first processing routine associated with the first anatomical structure to the first image segment to obtain a processed first image segment; applying a second processing routine, different from the first processing routine, associated with the second anatomical structure to the second image segment to obtain a processed second image segment; outputting a processed second image including both the processed first image segment and the processed second image segment; 23. A method for processing medical images comprising:
2. the first image or the second image is selected from an imaging modality including at least one of computed tomography (CT), magnetic resonance imaging (MRI), positron emission tomography (PET), single photon emission computed tomography (SPECT), x-ray imaging including digital x-ray radiogrammetry (DXR), and fluoroscopy sequences in image-assisted therapy (IGT) imaging; The method of claim 1.
3. the first image corresponded to a scout scan acquired using a lower radiation dose than the second image; The method of claim 1.
4. the plurality of distinct anatomical structures in the first image are identified prior to receiving the second image. The method of claim 1.
5. the plurality of distinct anatomical structures in the first image are identified upon receiving the second image. The method of claim 1.
6. the first image segment excludes the second anatomical structure, and the second image segment excludes the first anatomical structure. The method of claim 1.
7. the first processing routine is a first machine learning algorithm associated with the first anatomical structure, and the second processing routine is a second machine learning algorithm associated with the second anatomical structure. The method of claim 1.
8. the first image segment and the second image segment are linearly parsed such that the first image segment encompasses the entire width of the top of the second image and the second image segment encompasses the entire width of the bottom of the second image; The method of claim 1.
9. the first anatomical structure or the second anatomical structure is selected from the head, the neck, the upper body, the abdomen, the pelvic region, the lower body, and the legs; The method of claim 1.
10. A memory for storing a plurality of instructions; An imaging unit; a processor circuit coupled to the memory, the processor circuit comprising: Obtaining a first image; acquiring a second image from the imaging unit different from the first image, the second image including the same target object as the first image; identifying a plurality of distinct anatomical structures in the first image; defining a plurality of image segments in the second image based on locations of the distinct anatomical structures identified in the first image, such that a first image segment of the plurality of image segments includes a first anatomical structure of the plurality of distinct anatomical structures and such that a second image segment of the plurality of image segments includes a second anatomical structure of the plurality of distinct anatomical structures; applying a first processing routine associated with the first anatomical structure to the first image segment to obtain a processed first image segment; applying a second processing routine, different from the first processing routine, associated with the second anatomical structure to the second image segment to obtain a processed second image segment; outputting a processed second image including both the processed first image segment and the processed second image segment; and executing said instructions to perform said imaging system.
11. the imaging unit performs imaging modalities including at least one of computed tomography (CT), magnetic resonance imaging (MRI), positron emission tomography (PET), single photon emission computed tomography (SPECT), x-ray imaging including digital x-ray radiogrammetry (DXR), and fluoroscopy sequences in image-assisted therapy (IGT) imaging; 11. The imaging system of claim 10.
12. the first image corresponded to a scout scan acquired using a lower radiation dose than the second image; 11. The imaging system of claim 10.
13. the plurality of distinct anatomical structures in the first image are identified prior to receiving the second image.
11. The imaging system of claim 10.
14. the plurality of distinct anatomical structures in the first image are identified upon receiving the second image.
11. The imaging system of claim 10.
15. the first image segment excludes the second anatomical structure, and the second image segment excludes the first anatomical structure.
11. The imaging system of claim 10.
16. the first processing routine is a first machine learning algorithm associated with the first anatomical structure, and the second processing routine is a second machine learning algorithm associated with the second anatomical structure.
11. The imaging system of claim 10.
17. the first image segment and the second image segment are linearly parsed such that the first image segment encompasses the entire width of the top of the second image and the second image segment encompasses the entire width of the bottom of the second image; 11. The imaging system of claim 10.
18. the first anatomical structure or the second anatomical structure is selected from the head, the neck, the upper body, the abdomen, the pelvic region, the lower body, and the legs; 11. The imaging system of claim 10.
19. Obtaining a first image; acquiring a second image different from the first image, the second image including the same target object as the first image; identifying a plurality of distinct anatomical structures in the first image; defining a plurality of image segments in the second image based on locations of the distinct anatomical structures identified in the first image, such that a first image segment of the plurality of image segments includes a first anatomical structure of the plurality of distinct anatomical structures and such that a second image segment of the plurality of image segments includes a second anatomical structure of the plurality of distinct anatomical structures; applying a first processing routine associated with the first anatomical structure to the first image segment to obtain a processed first image segment; applying a second processing routine, different from the first processing routine, associated with the second anatomical structure to the second image segment to obtain a processed second image segment; outputting a processed second image including both the processed first image segment and the processed second image segment; A non-transitory computer readable medium having stored thereon a program for processing medical images, the program including instructions for:
20. the instructions providing for identifying the plurality of different anatomical structures in the first image prior to receiving the second image or while receiving the second image.
20. The non-transitory computer readable medium of claim 19.
