Processing projection data

Abstract

A mechanism is disclosed for aiding in the automated identification of a CT image that can be produced from projection data. Projection data, representing an anatomical cycle of a subject, is split into parts, each of which is reconstructed into a CT image. The reconstructed CT images are processed to quantify predicted motion within each CT image. The measures of predicted motion are processed to produce an indicator of a temporal phase of the anatomical cycle of the subject.

Description

[0001] PROCESSING PROJECTION DATA

[0002] FIELD OF THE INVENTION

[0003] The present invention relates to the field of Computed Tomography (CT) scanning and, in particular, to the processing of projection data produced by a CT scanning system.

[0004] BACKGROUND OF THE INVENTION

[0005] There is an ongoing interest in the performance of non-invasive imaging of a subject or patient. In modem medicine, images produced using such techniques are important for aiding in the performance of diagnosis and analysis of a condition of the subject or patient. A CT scan provides one approach for performing imaging of the subject.

[0006] In particular, CT scanners are well-established medical imaging devices that use a detector to detect an interaction between X-ray radiation and irradiated material in order to generate medical imaging data. Typically, a CT scanner will produce projection data of an imaged area. The projection data is gradually acquired overtime, such that different parts of the projection data are captured or represent different points in time. If the imaged area includes (e.g., part of) an anatomical structure that is undergoing an anatomical cycle (e.g., the heart or lungs), then any images produced using such projection data may suffer from motion artefacts due to motion of the anatomical structure.

[0007] There is an ongoing desire to reduce the impact or presence of motion. In particular, automatic phase selection or identification is useful to facilitate identification of data that represents an interval or phase of minimal or low movement within the imaged area.

[0008] SUMMARY OF THE INVENTION

[0009] According to examples in accordance with an aspect of the invention, there is provided a computer-implemented method for processing projection data. The computer-implemented method comprises: obtaining the projection data of a portion of an anatomical cycle of a subject; identifying, from the projection data, at least two first parts of the projection data, each first part representing a different temporal phase of the anatomical cycle of the subject; processing each identified first part of the projection data to reconstruct a CT image for each identified first part of the projection; processing a portion of each reconstructed CT image using a motion measuring algorithm to produce a measure of motion for each reconstructed CT image; and generating an indicator that indicates the temporal phase of the anatomical cycle of the subject associated with one of the reconstructed CT images responsive to the measure of motion of each reconstructed CT image.

[0010] The indicator may indicate the temporal phase of the anatomical cycle of the subject associated with the reconstructed CT image having the measure of motion indicating a lowest motion amongst the reconstructed CT images. The present disclosure proposes a technique that generates a measure of motion for each of a plurality of different reconstructed CT images. In particular, each CT image is processed using a motion measuring algorithm independently of any other CT image.

[0011] Existing methods for automatic phase selection analyze temporally neighboring volumes. However, this approach is not only sensitive to different motion states in the different volumes, but it also measures differences in cone-beam artifacts (since different parts of the raw data are used) and different distribution of contrast agent, e.g., in the left ventricle. Thus, motion detection is less accurate.

[0012] The proposed approach, therefore, provides a mechanism for identifying the temporal phase of the anatomical cycle (or the part of the projection data associated with such a cycle) that is more sensitive to motion artifacts, and less sensitive to other forms of artifacts, such as cone-beam artifacts or changes in contrast agent distribution.

[0013] In the context of the present disclosure, the indicator may represent an indication of: the relevant temporal phase (directly, e.g., as a measure of relative time); the part of the projection data that represents the relevant temporal phase; and / or the reconstructed CT image having the measure of motion indicating a lowest motion amongst all reconstructed CT images. Other forms of indicator will be apparent to the appropriately skilled person.

[0014] The processing of a portion of each reconstructed CT image may comprise, for each reconstructed CT image: processing the reconstructed CT image to identify at least one region of interest within the reconstructed CT image; and processing only the at least one region of interest within the reconstructed CT image using the motion measuring algorithm to produce a measure of motion for each reconstructed CT image.

[0015] In some examples, for each reconstructed CT image, the at least one region of interest covers only a part of the reconstructed CT image.

[0016] The motion measuring algorithm may comprise a machine-learning algorithm. In some examples, the motion measuring algorithm is configured to estimate a motion vector for the reconstructed CT image.

[0017] The processing of a portion of each reconstructed CT image may comprise, for each reconstructed CT image, processing only the portion of the reconstructed CT image independently of any portion of any other reconstructed CT image.

[0018] The motion measuring algorithm may comprise an image feature measuring algorithm for measuring one or more image features of the reconstructed CT image.

[0019] In some examples, the motion measuring algorithm is configured to generate a measure of a sharpness and / or a blurring.

[0020] The computer-implemented method may further comprise identifying, from the projection data, at least two second parts of the projection data responsive to the identified temporal phase of the anatomical cycle, each second part representing a different temporal phase of the anatomical cycle of the subject; processing each identified second part of the projection data to reconstruct a further CT image for each identified second part of the projection; processing a portion of each reconstructed further CT image using a motion measuring algorithm to produce a further measure of motion for each reconstructed CT further image; and generating an indicator that indicates the temporal phase of the anatomical cycle of the subject associated with one of the reconstructed further CT images responsive to the further measure of motion of each reconstructed CT image.

[0021] In some examples, for each second part of the projection data, a temporal difference between the temporal phase represented by the second part and a nearest temporal phase represented by any other second part is less than the smallest temporal difference between the temporal phases represented by any first part and any other first part.

[0022] The step of obtaining projection data may comprise controlling a CT scanning system to acquire the projection data.

[0023] The step of obtaining projection data may comprise receiving the projection data from a memory or storage unit.

[0024] There is also proposed a computer program product comprising computer program code means which, when executed on a computing device having a processing system, cause the processing system to perform all of the steps of any herein disclosed method.

[0025] There is also proposed a device configured to process projection data, the device comprising: processing circuitry; and a memory containing instructions that, when executed by the processing circuitry, configure the processing circuitry to: obtain the projection data of a portion of an anatomical cycle of a subject; identify, from the projection data, at least two first parts of the projection data, each first part representing a different temporal phase of the anatomical cycle of the subject; process each identified first part of the projection data to reconstruct a CT image for each identified first part of the projection; process a portion of each reconstructed CT image using a motion measuring algorithm to produce a measure of motion for each reconstructed CT image; and generate an indicator that indicates the temporal phase of the anatomical cycle of the subject associated with one of the reconstructed CT images responsive to the measure of motion of each reconstructed CT image.

[0026] These and other aspects of the invention will be apparent from and elucidated with reference to the embodiment(s) described hereinafter.

[0027] BRIEF DESCRIPTION OF THE DRAWINGS

[0028] For a better understanding of the invention and to show more clearly how it may be carried into effect, reference will now be made, by way of example only, to the accompanying drawings, in which:

[0029] Figure 1 illustrates a system in which embodiments may be employed;

[0030] Figure 2 illustrates a proposed method;

[0031] Figure 3 illustrates a feature for use in the proposed method; Figure 4 illustrates identifying parts of projection data; and Figure 5 illustrates a variant to the proposed method.

[0032] DETAILED DESCRIPTION OF THE EMBODIMENTS

[0033] The invention will be described with reference to the Figures.

[0034] It should be understood that the detailed description and specific examples, while indicating exemplary embodiments of the apparatus, systems and methods, are intended for purposes of illustration only and are not intended to limit the scope of the invention. These and other features, aspects, and advantages of the apparatus, systems and methods of the present invention will become better understood from the following description, appended claims, and accompanying drawings. It should be understood that the Figures are merely schematic and are not drawn to scale. It should also be understood that the same reference numerals are used throughout the Figures to indicate the same or similar parts.

[0035] The invention provides a mechanism for aiding in the automated identification of a CT image that can be produced from projection data. Projection data, representing an anatomical cycle of a subject, is split into parts, each of which is reconstructed into a CT image. The reconstructed CT images are processed to quantify predicted motion within each CT image. The measures of predicted motion are processed to produce an indicator of a temporal phase of the anatomical cycle of the subject.

[0036] The present disclosure provides a mechanism for identifying a temporal phase of an anatomical cycle of the subject with low motion, which mechanism is less sensitive to cone beam artifacts and changes in contrast agent distribution.

[0037] Figure 1 illustrates a system 100 in which embodiments may be employed, for improved contextual understanding. The system 100 comprises a CT scanning system 110 and a device 120.

[0038] The CT scanning system 110 is configured to capture and / or generate CT projection data 150 of a subject 190. In this context, the projection data may comprise the raw data captured by the CT scanning system 110 during an imaging procedure of the subject 190 before reconstruction has been performed. The projection data may undergo initial filtering and / or processing.

[0039] The device 120 comprises processing circuitry and a memory. The memory contains instructions that, when executed by the processing circuitry, configure the processing circuitry to perform one or more tasks or functions. The device 120 may, for instance, be replaced by any other form of processing system.

[0040] The device 120 may be communicatively coupled to the CT scanning system 110 so as to receive at least projection data from the CT scanning system. The communicative coupling may be wired or wireless, and other approaches that are known in the art. In other embodiments, the CT scanning system 110 may store the projection data in a memory or storage unit 130, which may form part of the system 100. The device 120 may be communicatively coupled to the memory or storage unit, e.g., to receive or obtain projection from the memory storage unit 130.

[0041] The processing circuitry may include, but is not limited to, one or more of the following: conventional microprocessors, application specific integrated circuits (ASICs), and / or field- programmable gate arrays (FPGAs). The memory may comprise any volatile and / or non-volatile computer memory such as RAM, PROM, EPROM, and EEPROM. The instructions contained in the memory may effectively define one or more programs that, when executed on the processing circuitry, cause the processing circuitry to perform encoded functions.

[0042] The present disclosure recognizes that there is a strong desire to achieve high-quality and automated phase selection in CT imaging of an anatomical object that undergoes an anatomical cycle (e.g., cardiac CT imaging of a heat). This is of particular importance, as motion can significantly impair image quality, e.g., due to motion, and affect operator or clinician understanding of the CT images.

[0043] One approach for assessing an amount of motion of CT images is to apply a similarity measure to CT images of neighboring temporal phases, i.e., temporally neighboring CT images, to determine the amount of change. Alternatively, a motion vector field can be estimated between temporally neighboring CT images. Thus, more generally, analysis is made of temporally neighboring reconstructed CT images.

[0044] It is herein recognized that this approach is not only sensitive to different motion states represented by the different CT images, but it also influenced by differences in cone-beam artifacts as well as different distributions of contrast agent, if present. In brief, cone-beam artifacts originate from the natural inexactness of the axial scan trajectory. The cone-beam artifacts hamper correct identification of a CT image with the least motion, i.e., noise, since differences between temporally neighboring CT images may be dominated by differences in cone-beam artifacts rather than differences in motion state.

[0045] The present disclosure provides a technique that does not rely upon the need for neighboring CT images to facilitate identification of at least the temporal phases and possibly the corresponding CT images, associated with the least motion or a motion below an acceptable threshold.

[0046] Figure 2 is a flowchart illustrating a method 200 for processing projection data. The method 200 may be performed by the processing circuitry, e.g., when executing instructions stored in the memory. Thus, in practice, the method 200 may be a computer-implemented method.

[0047] The method 200 comprises a step 210 of obtaining the projection data of a portion of an anatomical cycle of a subject. The portion of the anatomical cycle may comprise the entire anatomical cycle or only a part of the anatomical cycle. The method 200 also comprises a step 220 of identifying, from the projection data, at least two first parts of the projection data. Each first part represents a different temporal phase of the anatomical cycle of the subject. It will be appreciated that different parts of the projection data are captured at or otherwise represent different windows in time.

[0048] Step 220 may, for instance, comprise dividing or segmenting the projection data into a plurality of different parts of the projection data. The different parts may partially overlap with one another.

[0049] In some examples, each part of the projection data is a short scan part of the projection data. In this context, a short scan part of the projection data is a piece of projection data that provides the minimum amount of data needed to reconstruct a complete CT image therefrom. The size of such projection data is predefined, and may depend, for instance, on the reconstruction algorithms used and / or the nature of the CT scanning system.

[0050] The method 200 also comprises a step 230 of processing each identified first part of the projection data to reconstruct a CT image for each identified first part of the projection. Any appropriate techniques for performing CT image reconstruction may be performed here, which are well known and established in the art.

[0051] The method 200 also comprises a step 240 of processing a portion of each reconstructed CT image using a motion measuring algorithm to produce a measure of motion for each reconstructed CT image. In this way, an image-specific measure of motion is produced for each reconstructed CT image. More particularly, each CT image is processed using a motion measuring algorithm independently of any other CT image, i.e., using a blind motion measuring algorithm.

[0052] In the context of the present disclosure, a motion measuring algorithm is therefore an algorithm that processes a portion of a CT image to produce a value or measure that changes responsive to changes in motion of the subject during capturing of the data used to produce the CT image. In other words, a motion measuring algorithm aims to quantify motion or motion artifacts within a CT image.

[0053] In some examples, the motion measuring algorithm comprises a machine-learning algorithm. For instance, the motion measuring algorithm may be configured to estimate a motion vector or motion field for the reconstructed CT image. Approaches for subsequently generating a measure of motion from a motion vector / field are well established in the art and may be employed.

[0054] In other examples, the motion measuring algorithm comprises an image feature measuring algorithm for measuring one or more image features of the reconstructed CT image. In the context of the present disclosure, an image feature is any feature or property of the CT image that changes responsive to noise-causing motion of the subject during capturing of the data used to produce the image. By way of example, the motion measuring algorithm may be configured to generate a measure of a sharpness and / or a blurring. Thus, sharpness and blurring are suitable examples of image features. Suitable image feature measuring algorithms include those suggested by De, Kanjar, and V. Masilamani. "Image sharpness measure for blurred images in frequency domain." Procedia Engineering 64 (2013): 149-158; Campanella, Gabriele, et al. "Towards machine learned quality control: A benchmark for sharpness quantification in digital pathology." Computerized Medical Imaging and Graphics 65 (2018): 142-151; and Jaya, V. L., and R. Gopikakumari. "IEM: a new image enhancement metric for contrast and sharpness measurements." International Journal of Computer Applications 79.9 (2013). These provide only a small selection of appropriate techniques, and the skilled person would readily identify further approaches for performing image feature measuring.

[0055] The method 200 also comprises a step 250 of generating an indicator that indicates the temporal phase of the anatomical cycle of the subject associated with one of the reconstructed CT images responsive to the measure of motion of each reconstructed CT image.

[0056] The indicator may, for instance, identify one of the reconstructed CT images. Each CT image will be associated with a different temporal phase of the anatomical cycle due to being reconstructed from different first parts of the projection data. In other examples, the indicator may identify one of the first parts of the projection data.

[0057] Step 250 may, for instance, be configured such that the indicator indicates the temporal phase of the anatomical cycle of the subject associated with the reconstructed CT image having the measure of motion indicating a lowest motion amongst the reconstructed CT images.

[0058] Thus, step 250 may effectively generate an indicator that facilitates identification of the reconstructed CT image with the lowest level of motion blur, e.g., lowest motion.

[0059] As another example, step 250 may be configured such that the indicator indicates the temporal phase of the anatomical cycle associated with a reconstructed CT image having a measure of motion that meets one or more predetermined requirements or conditions, such as a measure of motion that indicates a level of motion below a predetermined motion threshold.

[0060] In some examples, the method 200 may further comprise a step 260 of outputting the part of the projection data and / or the CT image associated with the temporal phase of the anatomical cycle indicated by the indicator, e.g., generated in step 250).

[0061] Step 260 may, for instance, comprise controlling a user interface to provide a visual representation of the output CT image. This aids an operator or clinician in assessing the condition of the subject, e.g., performing diagnosis, as the selection is less likely to be influenced by motion.

[0062] In one example, step 260 comprises performing reconstruction on the part of the projection data and / or the CT image associated with the temporal phase of the anatomical cycle indicated by the indicator.

[0063] However, other steps may be performed in addition to or instead of the illustrated step 260. For instance, in some examples, the method 200 may further comprise storing the indicator for later reference. In some examples, the method may further comprise marking the projection data to identify the part of the projection data. Figure 3 illustrates one approach for performing step 240 of processing a portion of each reconstructed CT image to produce a respective measure of motion.

[0064] In this approach, each measure of motion is produced using only a region of interest of each reconstructed CT image. The region of interest may, for instance, be a region of each reconstructed CT image that contains a desired anatomical feature or structure, e.g., a heart.

[0065] Thus, the step 240 may comprise a sub-step 310 of processing the reconstructed CT image to identify at least one region of interest within the reconstructed CT image. It will be appreciated that, in practice, the at least one region of interest for each reconstructed CT image covers only a part of the reconstructed CT image.

[0066] Sub-step 310 may, for instance, comprise any suitable object detection algorithm for identifying a desired anatomical object within the CT image.

[0067] Suitable examples of object detection algorithms are well known in the art, including the YOLO algorithm, first set out by Redmon, Joseph, et al. "You only look once: Unified, real-time object detection." Proceedings of the IEEE conference on computer vision and pattern recognition. 2016 or anatomical segmentation techniques, such as those proposed by Zhu, Wentao, et al. "AnatomyNet: deep learning for fast and fully automated whole-volume segmentation of head and neck anatomy." Medical physics 46.2 (2019): 576-589.

[0068] The identity of the desired anatomical object may, for instance, depend upon the use-case scenario. For instance, in the context of cardiac CT imaging, the anatomical object may comprise a heart. For liver imaging, the anatomical object may comprise a fetus. For respiratory imaging, the anatomical object may comprise the lungs.

[0069] In some examples, the desired object is a predefined object. In other examples, the desired object is indicated in an object indicator. Thus, sub-step 310 may comprise receiving an object indicator (e.g., from a user input or central control system) indicating a desired object and using an object detection algorithm to identify any regions of interest predicted to contain the desired object.

[0070] The step 240 may also comprise a sub-step 320 of processing only the at least one region of interest within the reconstructed CT image using the motion measuring algorithm to produce a measure of motion for each reconstructed CT image. Thus, in some approaches, only a part, i.e., not all, of each reconstructed CT image may be processed in order to produce the measure of motion.

[0071] Referring back to Figure 2, it has been previously described, with reference to step 220, how each first part represents a different temporal phase of the anatomical cycle of the subject.

[0072] Figure 4 conceptually illustrates an instance of projection data 400. The projection data comprises a sequence of projection data elements, each projection data element being captured by the CT scanning system at a different point in time. Of course, due to practical considerations and constraints of capturing data, each point in time may actually represent a small period of time.

[0073] In this way, the projection data 400 may comprise a sequence of projection data elements captured from an initial point in time tiN to an end point in time IEND. In this way, the projection data elements cover a portion, i.e., some or all, of an anatomical cycle of the subject. The difference between the end point in time and the initial point in time, i.e., IEND - tiN, may be labelled the total time of the projection data.

[0074] Correspondingly, each first part of the projection data comprises a subset, i.e., not all, of the projection data elements of the projection data. Each first part of the projection data therefore represents a different window of time, and therefore a different temporal phase of an anatomical cycle of the subject.

[0075] Figure 4 also conceptually illustrates an example performance of step 220. The projection data 400 is processed to identify a plurality of different first parts of projection data.

[0076] As previously mentioned, each first part of the projection data represents a different window of time. Thus, an exemplary first part 410 of the projection data can be associated with a start time ts, a center or midpoint time tcand an end time te. Any of these times could be labelled or used as the “time” of the first part of the projection data or its corresponding temporal phase. In effect, a temporal phase may have a “time”, being one of these previously mentioned times associated with the part of the projection data associated with the temporal phase.

[0077] In this way, a part of the projection data is associated with, or has, both a time window, from a start time tsto an end time te, and a time, e.g., one of the start time ts, the center or midpoint time tcand the end time te. A temporal phase associated with a part of the projection data is similarly said to be associated with or have both a time window from a start time tsto an end time teand a time, e.g., one of the start time ts, the center or midpoint time tcand the end time te.

[0078] A difference between the two times of two different parts of the projection data, e.g., between the midpoint times tc, of the first parts can be labelled a temporal difference. Thus, it is possible to define a temporal difference between two temporal phases, being the difference between the times associated with the two temporal phases.

[0079] For a part of the projection data or temporal phase, any time associated therewith may be represented as an absolute time, e.g., using a timestamp, as a relative time, e.g., measured in units of time from the initial point in time tiN or units of time until the end point in time IEND, as a percentage or as a fraction. Here, the percentage / fraction may represent a proportion of the total time, between the initial and end points in time, which has elapsed from the initial point in time. Thus, for instance, if a center time tcoccurs 5 seconds after the initial point in time (and IEND - tiN = 10 seconds), then the center time tccan be alternatively expressed as 50% or 'A.

[0080] Referring back to Figure 2, the step 250 of generating an indicator may comprise generating an indicator identifying the time of one of the parts of the projection data.

[0081] Figure 5 illustrates a variant to the previously described method 200 for processing projection data. In this variant, the projection data obtained in step 210 is iteratively processed one or more times. Each iteration produces an updated indicator that indicates a temporal phase of the anatomical cycle of the subject. In other words, steps 220, 230, 240 and 250 are performed in each iteration.

[0082] Thus, the method 200 may further comprise a step 510 of determining whether or not to perform a further iteration of processing the projection data. Responsive to a positive communication, the projection data is reprocessed in a further iteration. Otherwise, the method may end or perform the optional step 260 of outputting data. Step 260 is, of course, appropriately adapted to make use of the indicator generated in the most recent iteration of step 250.

[0083] Various termination criteria may be used to perform step 510. For instance, step 510 may comprise determining whether or not a predetermined number of iterations have been performed with a failure to perform the predetermined number causing a further iteration to be performed. Other examples will be provided later.

[0084] For the sake of descriptive ease, each N-th iteration of step 220 may comprise identifying at least two N-th parts of the projection data, e.g., a first part for a first iteration, a second part for a second iteration and so on. The N-th iterations of step 230 may similarly comprise processing each identified N-th part.

[0085] Thus, the term “first” previously used in the context of steps 220, 230, 240, 250 may be replaced, where appropriate, by the term “N-th” or by another appropriate ordinal number, e.g., “second” for a second iteration or “third” for a third iteration. Similarly, other elements of steps 220, 230, 240, 250, e.g., CT image, may be further labeled by the term “N-th”, e.g., N-th CT image, or by an appropriate ordinal number.

[0086] In the context of this description, the term (N-l)th refers to an iteration immediately previous to the N-th iteration. Thus, during an N-th iteration, the (N-l)th iteration will be the immediately previous iteration. Thus, an (N-l)th part of the projection data is a part identified in an immediately previous iteration, e.g., a first part for the second iteration.

[0087] Thus, for a second iteration of steps 220 - 250, the method may comprise: identifying, from the projection data, at least two second parts of the projection data responsive to the identified temporal phase of the anatomical cycle, each second part representing a different temporal phase of the anatomical cycle of the subject; processing each identified second part of the projection data to reconstruct a further CT image for each identified second part of the projection; processing a portion of each reconstructed further CT image using a motion measuring algorithm to produce a further measure of motion for each reconstructed CT further image; and generating an indicator that indicates the temporal phase of the anatomical cycle of the subject associated with one of the reconstructed further CT images responsive to the further measure of motion of each reconstructed CT image.

[0088] For instance, the indicator may identify the temporal phase of the anatomical cycle of the subject associated with the reconstructed further CT image having the measure of motion indicating a lowest motion amongst all reconstructed further CT images. Preferably, in each iteration of step 220, for each N-th part of the projection data, a temporal difference between the temporal phase represented by the N-th part and a nearest temporal phase represented by any other N-th part is less than the smallest temporal difference between the temporal phases represented by any (N-l)th part and any (N-l)th part.

[0089] By way of example, for each second part of the projection data, a temporal difference between the temporal phase represented by the second part and a nearest temporal phase represented by any other second part may be less than the smallest temporal difference between the temporal phases represented by any first part and any other first part.

[0090] For instance, for each N-th part of the projection data, a temporal difference between the temporal phase represented by the N-th part and a nearest temporal phase represented by any other N- th part may be half the smallest temporal difference between the temporal phases represented by any (N-l)th part and any (N-l)th part.

[0091] Preferably, in each iteration of step 220, the proportion of the total time of the projection cumulatively represented by all the N-th parts of the projection data may reduce.

[0092] Preferably, in each iteration of step 220, the N-th parts of the projection data include, e.g., are centered around, the time associated with the temporal phase indicated in the indicator generated in the (N-l)th iteration of step 250, i.e., the most recent iteration of step 250.

[0093] Thus, consider a scenario in which the indicator identifies a time associated with the temporal phase of the anatomical cycle associated with the reconstructed CT image indicating a lowest motion amongst the reconstructed CT images. In this scenario, during an Nth iteration, excluding the first iteration, the N-th parts of the projection data include at least one N-th part representing a time window that includes the time identifies in the indicator.

[0094] This approach helps to more precisely narrow down or track the projection data with the lowest motion, i.e., by performing an iterative refinement procedure.

[0095] Preferably, for the first iteration of step 220, the proportion of the total time of the projection cumulatively represented by all the first parts of the projection data is greater than 95%, e.g., 100%.

[0096] Another termination criterion for use in step 510 is that the measure of motion for the reconstructed CT image associated with the temporal phase of the anatomical cycle indicated by the indicator meets a predetermined condition, e.g., indicates a motion below a predetermined threshold motion.

[0097] At least one previously disclosed approach makes use of a machine-learning algorithm to function as the motion measuring algorithm. A machine-learning algorithm is any self-training algorithm that processes input data in order to produce or predict output data. Here, the input data comprises a CT image and the output data comprises a measure of motion (or data from which a measure of motion can be derived).

[0098] Suitable machine-learning algorithms for being employed in the present invention will be apparent to the skilled person. Examples of suitable machine-learning algorithms include decision tree algorithms and artificial neural networks. Other machine-learning algorithms such as logistic regression, support vector machines or Naive Bayesian models are suitable alternatives.

[0099] The structure of an artificial neural network (or, simply, neural network) is inspired by the human brain. Neural networks are comprised of layers, each layer comprising a plurality of neurons. Each neuron comprises a mathematical operation. In particular, each neuron may comprise a different weighted combination of a single type of transformation (e.g., the same type of transformation, sigmoid etc. but with different weightings). In the process of processing input data, the mathematical operation of each neuron is performed on the input data to produce a numerical output, and the outputs of each layer in the neural network are fed into the next layer sequentially. The final layer provides the output.

[0100] Methods of training a machine-learning algorithm are well known. Typically, such methods comprise obtaining a training dataset, comprising training input data entries and corresponding training output data entries.

[0101] For some machine-learning algorithms, such as a neural network, training is performed by applying an initialized machine-learning algorithm to each input data entry to generate predicted output data entries. An error between the predicted output data entries and corresponding training output data entries is used to modify the machine-learning algorithm. This process can be repeated until the error converges, and the predicted output data entries are sufficiently similar (e.g., ±1%) to the training output data entries. This is commonly known as a supervised learning technique.

[0102] For example, where the machine-learning algorithm is formed from a neural network, (weightings of) the mathematical operation of each neuron may be modified until the error converges. Known methods of modifying a neural network include gradient descent, backpropagation algorithms and so on.

[0103] Consider a scenario in which the machine-learning algorithm is to be trained to process a CT image to produce a motion vector. In this scenario, the training input data entries correspond to example CT images. The training output data entries correspond to example motion fields. One approach for generating pairs of example CT images and corresponding example motion fields is to apply a respective motion field to each of a plurality of different sample CT images, e.g., using a motion compensated reconstruction algorithm. This can effectively produce motion corrupted CT images that function as example CT examples for the purposes of training the machine-learning algorithm.

[0104] The skilled person would be readily capable of developing a device for carrying out any herein described method. Thus, each step of the flow chart may represent a different action performed by processing circuitry of a device, e.g., carrying out instructions stored by a memory of the device.

[0105] Embodiments may therefore make use of processing circuitry. A processor is one example of processing circuitry which employs one or more microprocessors that may be programmed using software, e.g., microcode, stored by the memory of the device, to perform the required functions. Processing circuitry may however be implemented with or without employing a processor, and also may be implemented as a combination of dedicated hardware to perform some functions and a processor, e.g., one or more programmed microprocessors and associated circuitry, to perform other functions.

[0106] Examples of processing circuitry components that may be employed in various embodiments of the present disclosure include, but are not limited to, conventional microprocessors, application specific integrated circuits (ASICs), and field-programmable gate arrays (FPGAs).

[0107] Processing circuitry is associated with memory, e.g., one or more storage media such as volatile and non-volatile computer memory such as RAM, PROM, EPROM, and EEPROM. The memory is encoded with one or more programs that, when executed on the processing circuitry, perform the required functions. One or more various memories may be fixed within the processing circuitry or may be transportable, such that the one or more programs stored thereon can be loaded into the processing circuitry.

[0108] It will be understood that disclosed methods are preferably computer-implemented methods. As such, there is also proposed the concept of a computer program comprising code means for implementing any described method when said program is run on a processing system, such as a computer or the processing circuitry of the previously mentioned device. Thus, different portions, lines or blocks of code of a computer program according to an embodiment may be executed by a processing system, computer or processing circuitry to perform any herein described method.

[0109] There is also proposed a non-transitory storage medium that stores or carries a computer program or computer code that, when executed by a processing system or processing circuitry, causes the processing system or processing circuitry to carry out any herein described method.

[0110] In some alternative implementations, the functions noted in the block diagrams or flow charts may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved.

[0111] Variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure and the appended claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

[0112] In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. If the term "adapted to" is used in the claims or description, it is noted the term "adapted to" is intended to be equivalent to the term "configured to". If the term "arrangement" is used in the claims or description, it is noted the term "arrangement" is intended to be equivalent to the term "system", and vice versa.

[0113] A single processor or other unit may fulfill the functions of several items recited in the claims. If a computer program is discussed above, it may be stored / distributed on a suitable medium, such as an optical storage medium or a solid-state medium supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems.

[0114] Any reference signs in the claims should not be construed as limiting the scope.

Claims

CLAIMS:

1. A computer-implemented method for processing projection data in Computed Tomography (CT) imaging, the computer-implemented method comprising: obtaining the projection data of a portion of an anatomical cycle of a subject; identifying, from the projection data, at least two first parts of the projection data, each first part representing a different temporal phase of the anatomical cycle of the subject; processing each identified first part of the projection data to reconstruct a CT image for each identified first part of the projection; processing a portion of each reconstructed CT image using a motion measuring algorithm to produce a measure of motion for each reconstructed CT image; and generating an indicator that indicates the temporal phase of the anatomical cycle of the subject associated with one of the reconstructed CT images responsive to the measure of motion of each reconstructed CT image.

2. The computer-implemented method of claim 1, wherein the indicator indicates the temporal phase of the anatomical cycle of the subject associated with the reconstructed CT image having the measure of motion indicating a lowest motion amongst the reconstructed CT images.

3. The computer-implemented method of claim 1 or 2, wherein the processing of a portion of each reconstructed CT image comprises, for each reconstructed CT image: processing the reconstructed CT image to identify at least one region of interest within the reconstructed CT image; and processing only the at least one region of interest within the reconstructed CT image using the motion measuring algorithm to produce a measure of motion for each reconstructed CT image.

4. The computer-implemented method of claim 3, wherein, for each reconstructed CT image, the at least one region of interest covers only a part of the reconstructed CT image.

5. The computer-implemented method of any of claims 1 to 4, wherein the motion measuring algorithm comprises a machine-learning algorithm.

6. The computer-implemented method of claim 5, wherein the motion measuring algorithm is configured to estimate a motion vector for the reconstructed CT image.

7. The computer-implemented method of any of claims 1 to 6, wherein the processing of a portion of each reconstructed CT image comprises, for each reconstructed CT image, processing only theportion of the reconstructed CT image independently of any portion of any other reconstructed CT image.

8. The computer-implemented method of any of claims 1 to 7, wherein the motion measuring algorithm comprises an image feature measuring algorithm for measuring one or more image features of the reconstructed CT image.

9. The computer-implemented method of claim 8, wherein the motion measuring algorithm is configured to generate a measure of a sharpness and / or a blurring.

10. The computer-implemented method of any of claims 1 to 9, further comprising: identifying, from the projection data, at least two second parts of the projection data responsive to the identified temporal phase of the anatomical cycle, each second part representing a different temporal phase of the anatomical cycle of the subject; processing each identified second part of the projection data to reconstruct a further CT image for each identified second part of the projection; processing a portion of each reconstructed further CT image using a motion measuring algorithm to produce a further measure of motion for each reconstructed CT further image; and generating an indicator that indicates the temporal phase of the anatomical cycle of the subject associated with one of the reconstructed further CT images responsive to the further measure of motion of each reconstructed CT image.

11. The computer-implemented method of claim 10, wherein, for each second part of the projection data, a temporal difference between the temporal phase represented by the second part and a nearest temporal phase represented by any other second part is less than the smallest temporal difference between the temporal phases represented by any first part and any other first part.

12. The computer-implemented method of any of claims 1 to 11, wherein obtaining projection data comprises controlling a CT scanning system to acquire the projection data.

13. The computer-implemented method of any of claims 1 to 11, wherein obtaining projection data comprises receiving the projection data from a memory or storage unit.

14. A computer program product comprising computer program code means which, when executed on a computing device having a processing system, cause the processing system to perform all of the steps of the method according to any of claims 1 to 13.

15. A device for processing projection data in Computed Tomography (CT) imaging, the device comprising: processing circuitry; and a memory containing instructions that, when executed by the processing circuitry, configure the processing circuitry to: obtain the projection data of a portion of an anatomical cycle of a subject; identify, from the projection data, at least two first parts of the projection data, each first part representing a different temporal phase of the anatomical cycle of the subject; process each identified first part of the projection data to reconstruct a CT image for each identified first part of the projection; process a portion of each reconstructed CT image using a motion measuring algorithm to produce a measure of motion for each reconstructed CT image; and generate an indicator that indicates the temporal phase of the anatomical cycle of the subject associated with one of the reconstructed CT images responsive to the measure of motion of each reconstructed CT image.

Images