Spread field imaging collimator for radiation-based imaging and method of using the radiation-based imaging

Abstract

To provide an image processing method that improves imaging resolution without sacrificing signal sensitivity.SOLUTION: A radiation-based imaging system includes a collimator configured to filter radiation emitted from a target object, the collimator including a plurality of apertures non-uniformly distributed on the collimator. A largest acceptance angle of the plurality of apertures is not larger than 15°. The radiation-based imaging system further includes a detector for detecting the radiation that has passed through the collimator. The collimator is spaced from the detector such that one point on a top surface of the detector that faces the collimator is simultaneously illuminated by two or more of the plurality of apertures.SELECTED DRAWING: Figure 1A

Description

[Technical Field]

[0001] Priority This application claims priority to and the benefit of U.S. Provisional Patent Application No. 63 / 168,778, filed March 31, 2021, and U.S. Provisional Patent Application No. 63 / 071,540, filed August 28, 2020, which claims priority to U.S. Patent Application No. 17 / 399,768, filed August 11, 2021, the entire contents of which are incorporated herein by reference. [Background technology]

[0002] In radiation imaging, such as molecular medicine imaging (also known as nuclear medicine imaging), images depicting the distribution of radiopharmaceuticals are produced for medical diagnosis. Prior to imaging, a subject, such as a patient, is injected with the radiopharmaceutical. The radiopharmaceutical emits radioactive photons that pass through the body and are detected by a photon detector. Based on information from the received photons, the photon detector can then determine the distribution of the radiopharmaceutical within the patient. Because this distribution represents the patient's physiology, distribution images provide valuable clinical information for diagnosing various diseases and conditions in cardiology, oncology, neurology, and other fields.

[0003] Images are generated by a collimator and a photon detector working in tandem. A collimator is a device that guides photons along a specific path. Unlike X-rays and CT scans, where photons are emitted from a known source, radiation imaging uses photons from unknown locations within the subject's body. Without a collimator, photons from all directions would be recorded by the photon detector, potentially making image reconstruction difficult. Therefore, a collimator, like a lens in a photographic camera, is used to guide photons along possible paths and reconstruct an image. While existing radiation imaging systems are generally suitable for their intended purposes, they are not completely satisfactory in all respects. For example, existing collimators and detectors often lack excellent imaging resolution and signal sensitivity, and tradeoffs between the two must be made. Therefore, improvements to radiation imaging systems are desirable. Summary of the Invention [Means for solving the problem]

[0004] According to various embodiments, the present disclosure provides a radiation imaging system. The radiation imaging system includes a collimator configured to filter radiation emitted from a target object, the collimator including a plurality of apertures non-uniformly distributed on the collimator, the plurality of apertures having a maximum acceptance angle of 15° or less; and a detector for detecting radiation passing through the collimator. The collimator is spaced from the detector such that a point on an upper surface of the detector facing the collimator is simultaneously illuminated by two or more of the plurality of apertures. In some embodiments, a point on the upper surface of the detector is illuminated by a predetermined percentage of the plurality of apertures, the percentage being less than about 25%. In some embodiments, the number of the plurality of apertures exceeds 1,000. In some embodiments, the plurality of apertures differ in at least one of aperture size, aperture shape, acceptance angle, aperture length, and aperture pitch. In some embodiments, the distance between the collimator and the detector is about 0.5 to about 10 times (e.g., about 1.5 to about 10 times) the thickness of the collimator. In some embodiments, the plurality of apertures form a predetermined aperture pattern including a repeating base pattern. In some embodiments, the repeating base pattern is a coded aperture pattern. In some embodiments, an illumination area of ​​one of the plurality of apertures is substantially equal to an area of ​​the base pattern. In some embodiments, the collimator includes a first portion and a second portion, the first portion including through-holes uniformly distributed on the first portion, and the second portion including through-holes non-uniformly distributed on the second portion and corresponding to the plurality of apertures, such that the through-holes of the first portion are blocked by the second portion. In some embodiments, the thickness of the first portion is about 2 to about 10 times that of the second portion.In some embodiments, the collimator is a first collimator, the detector is a first detector, and the system further includes a second collimator and a second detector coupled to the second collimator, and the second collimator is attached to the second detector.

[0005] According to various embodiments, the present disclosure provides a radiation imaging system including: a first collimator configured to filter radiation emitted from a target object, the first collimator including a first plurality of apertures forming a first aperture pattern; a first detector associated with the first collimator and configured to detect radiation passing through the first collimator; a second collimator configured to filter radiation emitted from the target object, the second collimator including a second plurality of apertures forming a second aperture pattern; and a second detector associated with the second collimator and configured to detect radiation passing through the second collimator, the target object being positioned between the first collimator and the second collimator; and the first aperture pattern being different from the second aperture pattern. In some embodiments, the first collimator contacts the first detector, and the second collimator is spaced apart from the second detector. In some embodiments, the distance between the second collimator and the second detector is about 0.5 to about 7 times (e.g., about 1.5 to about 7 times) the thickness of the second collimator. In some embodiments, the first plurality of apertures are uniformly distributed on the first collimator, and the second plurality of apertures are non-uniformly distributed on the second collimator. In some embodiments, the second aperture pattern includes a repeating coding pattern. In some embodiments, the repeating coding pattern is one of a uniformly redundant array (URA) pattern, a modified uniformly redundant array (MURA) pattern, and a perfect binary array (PBA) pattern. In some embodiments, the facing directions of the first and second detectors have an offset angle.

[0006] According to various embodiments, the present disclosure provides a collimator configured to filter radiation emitted from a target object. The collimator includes a first portion including a uniformly distributed first plurality of through-holes and a second portion including a non-uniformly distributed second plurality of through-holes, the first and second portions being spaced apart and aligned such that photons emitted by the object are blocked by the second portion upon passing through some of the first plurality of through-holes. In some embodiments, the second plurality of through-holes form a coding pattern. In some embodiments, the second portion is operable to slide relative to a surface of the first portion facing the second portion, thereby changing the alignment of the first and second plurality of through-holes.

[0007] According to various embodiments, the present disclosure provides a method for correcting misalignment in an imaging system. The method includes providing a collimator and a detector having a digitized pixel grid, illuminating the collimator with a flood light source, and recording a signal intensity at each pixel of the digitized pixel grid, the signal intensity being generated by illuminating the collimator. The recording step includes deriving a descriptor from the recorded signal intensities, generating a set of offsets introduced into the digitized pixel grid, determining a selected offset from the set of offsets that optimizes the descriptor, and reconstructing the digitized pixel grid based on the actual optimal offset derived from the selected offsets. In some embodiments, the selected offset includes a pair of offset values ​​in two orthogonal directions. In some embodiments, the descriptor is one of a contrast value, a peak value, and a valley value of the recorded signal intensities. In some embodiments, the contrast value is determined by a ratio of a peak value to a valley value of a signal intensity line corresponding to the signal intensities recorded in the digitized pixel grid. In some embodiments, determining the selected offset comprises sweeping through the range of offsets and selecting a selected offset that maximizes or minimizes the descriptor, In some embodiments, reconstructing the digitized pixel grid comprises adjusting the selected offset by a fixed offset to generate the actual optimum offset, and reconstructing the digitized pixel grid by the actual optimum offset. [Brief explanation of the drawings]

[0008] The present disclosure is best understood from the following detailed description when taken in conjunction with the accompanying drawings. It should be noted that, in accordance with common practice in the art, the various components are not drawn to scale and are used for illustrative purposes only. Rather, the dimensions of the various components have been arbitrarily expanded or reduced for clarity of illustration. [Figure 1A] 1A, 1B, and 1C are schematic diagrams of exemplary radiological imaging systems according to various aspects of the present disclosure. [Figure 1B-1C] 1A, 1B, and 1C are schematic diagrams of exemplary radiological imaging systems according to various aspects of the present disclosure. [Figure 2A-2B] 2A, 2B, 2C, 2D, 2E, and 2F are partial perspective, cross-sectional, and plan views of a collimator according to some embodiments of the present disclosure. [Figure 2C-2D] 2A, 2B, 2C, 2D, 2E, and 2F are partial perspective, cross-sectional, and plan views of a collimator according to some embodiments of the present disclosure. [Figures 2E-2F] 2A, 2B, 2C, 2D, 2E, and 2F are partial perspective, cross-sectional, and plan views of a collimator according to some embodiments of the present disclosure. [Figure 3] FIG. 3 is a partial cross-sectional view of a radiological imaging system according to some embodiments of the present disclosure. [Figure 4] FIG. 4 is a partial cross-sectional view of a radiation imaging system according to some alternative embodiments of the present disclosure. [Figure 5A-5B] 5A and 5B are partial plan views of a collimator with non-uniformly distributed openings according to various embodiments of the present disclosure. [Figures 6A-6C] 6A, 6B, and 6C are plan views of collimators with repeating base patterns according to some embodiments of the present disclosure. [Figure 7] FIG. 7 illustrates the decomposition of an original image into sub-images in a fast image reconstruction algorithm according to various aspects of the present disclosure. [Figures 8A-8C] 8A, 8B, 8C, 9A, 9B, 9C, 10A, and 10B show cross-sectional views of some embodiments of lightweight collimators according to some embodiments of the present disclosure. [Figures 9A-9C]8A, 8B, 8C, 9A, 9B, 9C, 10A, and 10B show cross-sectional views of some embodiments of lightweight collimators according to some embodiments of the present disclosure. [Figures 10A-10B] 8A, 8B, 8C, 9A, 9B, 9C, 10A, and 10B show cross-sectional views of some embodiments of lightweight collimators according to some embodiments of the present disclosure. [Figures 11A-11B] 11A, 11B, 12A, 12B, and 13 show plan and cross-sectional views of several embodiments of radiological imaging systems having alignment adjustments. [Figures 12A-12B] 11A, 11B, 12A, 12B, and 13 show plan and cross-sectional views of several embodiments of radiological imaging systems having alignment adjustments. [Figure 13] 11A, 11B, 12A, 12B, and 13 show plan and cross-sectional views of several embodiments of radiological imaging systems having alignment adjustments. DETAILED DESCRIPTION OF THE INVENTION

[0009] The following disclosure provides many different embodiments, or examples, for implementing different features of the present disclosure. To simplify the disclosure, specific examples of components and arrangements are described below. These are, of course, merely examples and are not intended to be limiting. Any changes and further modifications to the described devices, systems, and methods, and any further applications of the principles of the present disclosure, as would normally occur to one skilled in the art to which the present disclosure pertains, are naturally contemplated. For example, features, components, and / or steps described in connection with one embodiment can be combined with features, components, and / or steps described in connection with another embodiment of the present disclosure, even if the combination is not explicitly set forth, to thereby form yet another embodiment of a device, system, or method according to the present disclosure. Furthermore, although reference numerals and / or letters may be repeated in various examples in the present disclosure, such repetition is for the purposes of brevity and clarity and does not in itself dictate a relationship between the various embodiments and / or configurations described.

[0010] Furthermore, in the following disclosure, "an element on, connected to, and / or coupled with another element" may include embodiments in which the element directly contacts the other element, as well as embodiments in which additional elements are interposed between the element and the other element, and the element does not directly contact the other element. Furthermore, spatial relative terms, such as "lower," "upper," "horizontal," "vertical," "over," "across," "below," "below," "above," "downward," "top," "bottom," and the like, and derivatives thereof (e.g., "horizontally," "downward," "upward," etc.), are used in this disclosure to facilitate the relationship of one element to another. Such spatial relative terms are intended to encompass different orientations of a device including the element. Furthermore, when a numerical value or range of values ​​is described with the terms "about," "approximately," etc., the term is intended to encompass values ​​within a reasonable range including the stated value, such as values ​​within ±10% of the stated value, or other values ​​understood by one of ordinary skill in the art. For example, the term "about 5 cm" encompasses a dimensional range of 4.5 cm to 5.5 cm.

[0011] The present disclosure relates to collimators used in radiation imaging, and more particularly to collimators with non-uniformly distributed apertures used in nuclear medicine (molecular) imaging systems. The term "non-uniformly distributed apertures" refers to collimator apertures (through-holes) that vary in at least one of the aperture profiles, including, but not limited to, aperture size, aperture shape (including cross-sectional shape and longitudinal shape), aperture length, aperture pitch, and aperture direction.

[0012] In radiation imaging, such as nuclear medicine (molecular) imaging systems, a collimator and detector work in conjunction to generate an image depicting the distribution of a radiopharmaceutical within a subject. Many nuclear medicine imaging systems, such as single-photon emission computed tomography (SPECT) and positron emission tomography (PET) imaging systems, use one or more detectors to acquire imaging data, such as gamma ray or photon imaging data. Prior to image acquisition, a radiopharmaceutical is typically administered orally or injected into a subject, such as a patient. Nuclear decay occurs in the radiopharmaceutical, resulting in the emission of gamma photons with specific velocities and characteristic energies through direct or indirect annihilation. One or more detector units are positioned around the subject to record or monitor the emissions. For manufacturing and data processing convenience, detectors are often configured in a planar geometry, so data is acquired in a 2D matrix format, often referred to as a projection. Based on the recorded information, including the location, energy, and number of detected events, an image depicting the distribution of the radiopharmaceutical can be reconstructed to observe the function of specific parts of the subject (e.g., a patient's body part).

[0013] However, existing collimator and detector designs have various problems. For example, collimators with parallel holes are usually closely coupled (or attached) to the detector to reduce crosstalk between detector pixels within the detector, resulting in only one possible photon path to any detector pixel. Typically, a longer collimator aperture improves the imaging system's imaging resolution but degrades signal sensitivity. Therefore, the length of the collimator aperture must be determined by considering the trade-off between imaging resolution and signal sensitivity.

[0014] The present disclosure presents collimator design embodiments in which a collimator with non-uniformly distributed apertures is positioned at a distance from a detector, allowing a specific area of ​​the detector to be simultaneously illuminated by photons passing through multiple apertures. In other words, at least some of the collimator apertures have overlapping illumination areas on the detector. By spacing the collimator away from the detector and utilizing a larger aperture size (or wider acceptance angle), the effective length of the collimator aperture is increased, improving imaging resolution without sacrificing signal sensitivity. Additionally, in some embodiments of the present disclosure, the collimator apertures have a repeating pattern. In further embodiments, this pattern may be one of a uniformly redundant array (URA) pattern, a modified uniformly redundant array (MURA) pattern, a perfect binary array (PBA) pattern, a random pattern, a pseudorandom pattern, and other suitable patterns. This repeating pattern allows radiation imaging systems to employ simplified image reconstruction algorithms, reducing computational complexity and improving system performance. In various embodiments, the collimator may also employ a lightweight design by including thin plates defining apertures that are not completely filled with closed-holes. This novel collimator design therefore improves the performance of radiological imaging systems in many aspects.

[0015] 1A-1C illustrate an exemplary radiological imaging system 100 (particularly a nuclear medicine imaging system) incorporating features of the present disclosure according to some embodiments. FIG. 1A is a side cross-sectional view of a portion of the system 100 in the XZ plane. FIGS. 1B and 1C are axial views of the system 100 in the YZ plane according to two embodiments. The imaging system 100 can be used to medically examine or treat a target object, such as a patient. The imaging system 100 includes an integrated gantry 102 that further includes a rotation mechanism (e.g., a rotor oriented about the gantry central bore) configured to support and rotate one or more detectors 108 (two detectors 108 shown in opposing positions) about an axial axis (e.g., the X direction as shown). In one embodiment illustrated in FIG. 1B, the two detectors 108 are operable to rotate along a circle 152 in opposing positions, where normal directions 154A and 154B to the top surfaces of the two detectors 108 both point toward the center of the circle 152 and in opposite directions. In an alternative embodiment shown in FIG. 1C , two detectors 108 are similarly operable to rotate along the circle 152 in opposing positions, with normal directions 154A and 154B to the top surfaces of the two detectors 108 pointing toward the center of the circle 152. The difference between the two is that the normal directions 154A and 154B in FIG. 1C are not completely opposite, but form a small angle Υ. In some embodiments, the angle Υ may range from approximately 0.5° to approximately 20°, depending on the system performance needs. Details of the setup in FIGS. 1B and 1C are described below. Each detector 108 operates in conjunction with a collimator 110. The collimator 110 is a device that guides the photon path. Unlike X-rays and CT, where photons are emitted from a known source location, in molecular imaging, photons are emitted from an unknown location within the subject's body. Without the collimator 110, photons from all directions would be recorded by the detectors 108, potentially making image reconstruction difficult. Therefore, the collimator 110 is used to direct the photon path to reconstruct the image.The imaging system 100 further includes a patient table 112 coupled to a table support system 114, which may be coupled directly to the floor or to the gantry 102 via a base. The patient table 112 is configured to slide relative to the table support system 114, facilitating movement of the patient 150 to and from an examination position substantially aligned with the axial axis. A control console 120 provides operation and control of the imaging system 100 in any manner known in the art. For example, the control console 120 can be used by an operator or technician to control mechanical movements such as rotation of the rotor 104, movement, rotation, or tilt of the detector 108 and collimator 110, and sliding of the patient table 112. The imaging system 100 further includes computer components (not shown), such as a data storage device, an image processor, an image storage device, and a display, for acquiring data and reconstructing nuclear medicine images. In some embodiments, one or more computer components may be partially or wholly located remotely (e.g., using cloud computing). In some embodiments, one or more of these components are located locally or remotely.

[0016] In some embodiments, detector 108 is a semiconductor detector, such as made of cadmium telluride (CdTe), cadmium zinc telluride (CZT), or high-purity germanium (HPGe). In some embodiments, detector 108 is a scintillator detector (such as made of sodium iodide (NaI) or cesium iodide (CsI)). In some other embodiments, detector 108 may also be a scintillator coupled with a miniature photomultiplier tube (PMT), a silicon photomultiplier tube (SiPMT), or an avalanche photodiode. One or more radiopharmaceuticals administered orally or injected into patient 150 undergo nuclear decay, emitting radiation (e.g., gamma photons) with a specific speed and characteristic energy through direct or indirect annihilation. Detector 108 is positioned near patient 150 to record or monitor the radiation emissions. Based on the recorded information, including the location, energy, and number of detections of the detected events, an image showing the distribution of the radiopharmaceutical can be reconstructed to observe the condition or function of the body part of the patient 150.

[0017] The collimator 110 includes multiple walls (also known as septa) that define one or more apertures (also called through-holes). In various embodiments, the septa are made of a heavy metal such as lead or tungsten. The thickness of the septa is large enough to block a large portion of the radiation, depending on the energy of the photons, so that the photons pass primarily through the small apertures in the plate. A larger thickness is required to image higher-energy gamma rays. The collimator 110 is positioned between the detector 108 and an imaging target, such as a patient 150. The collimator apertures determine the direction and angular range (acceptance angle) through which radiation can pass and reach a particular location on the detector. The collimator 110 may be a single-aperture collimator, a multi-aperture collimator, a coded aperture collimator, or any other suitable type of collimator, depending on the number and geometric arrangement of the apertures.

[0018] Imaging system 100 may include other components necessary for an imaging gantry, such as connectors for coupling components to one another (e.g., coupling detector 108 and collimator 110 to one another), motors for moving components, photon-shielding components, housing components for housing other components, etc. For example, coupling and shielding components 116 may couple detector 108 and collimator 110 such that they move (e.g., rotate) together and block radiation (photons) from reaching detector 108 through paths other than collimator 110. In another embodiment, detector 108 and collimator 110 may move independently relative to one another.

[0019] FIG. 2A shows a perspective view of an exemplary multi-aperture (or multi-port) collimator 202, and FIG. 2B shows a cross-sectional view of the collimator 202 (along section line BB in FIG. 2A ). The collimator 202 has a number of holes 208, each having a predetermined height H and a diameter (or width) W. The height H is also the thickness of the collimator 202. The holes 208 are also referred to as apertures or through-holes. The acceptance angle Θ of the aperture 208 is defined by the angle between two rays R0 and R1 traveling from an end point on the upper opening of the aperture 208 to an opposite end point on the lower opening. Only incident photons (or radiation) propagating within the acceptance angle Θ can pass through the aperture 208 (not considering the negligible fraction of rays that may be transmitted through the septum 212). As mentioned above, the septum 212 is made of a radiation-absorbing heavy metal(s) or alloy, such as lead or tungsten. The septum 212 absorbs most of the radiation that is not emitted from (or propagating toward) the target direction. Collimators for high-energy radiation have much thicker septums than collimators for low-energy radiation. The septums are typically designed so that the transmittance of unwanted photons through the septum does not exceed 5%, and in some cases, 1%. Note that the radiation or photons blocked or absorbed by the collimator do not require 100% blocking, since a very small percentage of photons (e.g., 5% or less) may still penetrate the thickness of the radiation-absorbing material. In other words, blocking (or other similar terms) means that a large percentage of photons (e.g., 95% or more, or 99% or more) are absorbed by the radiation-absorbing material.

[0020] The collimator 202 provides positional information of detected photons by limiting the acceptance angle Θ of incident photons. Typically, the smaller the acceptance angle Θ, the higher the imaging resolution provided by the collimator 202. The collimator 202 can have multiple substantially identical elongated apertures 208 arranged side by side and parallel to one another. Elongated apertures result in a small acceptance angle Θ, which correspondingly provides high imaging resolution. The collimator 202 has a thickness H greater than 15 mm, e.g., 20 mm to 70 mm, and an aperture diameter (or width) W of approximately 1 mm to 5 mm, thereby providing a large aspect ratio (H / W) and a small acceptance angle Θ. The acceptance angle Θ of the parallel-hole collimator 202 is limited to a small angle, e.g., less than 15°. If the acceptance angle Θ is too large (e.g., greater than 15°), the imaging resolution provided by a particular parallel-hole collimator will be low and considered undesirable. Unless otherwise specified, the acceptance angles of exemplary collimators of the present disclosure are less than 15°. In addition to the holes of Figure 2B, which have a rectangular "xz" cross section, the above acceptance angle concept can be extended to holes of other shapes, such as those of convergent or divergent collimators, where the holes may be trapezoidal in shape and have a slope that represents the maximum angular difference of light rays that can pass through the hole.

[0021] The collimator 202 shown in FIG. 2A is a parallel-hole collimator with uniformly distributed apertures. In the illustrated embodiment, the apertures 208 are substantially identical, each having the same height H and the same diameter or width W, except for variations caused by limitations in fabrication accuracy. Therefore, each aperture has the same small acceptance angle Θ (e.g., less than 15°, such as about 5° in a particular example). The aperture pitch P is also substantially the same across the collimator 202. That is, the aperture spacing S (thickness of the partition walls) is uniform across the collimator 202. Apertures with uniform aperture shape, aperture size, aperture length, and aperture pitch are referred to herein as "uniformly distributed apertures."

[0022] While FIGS. 2C and 2D show plan views of two arrangements of parallel-hole collimators (such as collimator 202) with uniformly distributed apertures, other arrangements are possible and do not depart from the spirit and scope of the present disclosure. In FIG. 2C, the apertures 208 are arranged in orthogonal rows and columns, forming an orthogonal two-dimensional grid arrangement. Each aperture 208 can be considered to be aligned with a square unit grid 240 in the grid network (represented by the dashed lines in FIG. 2C). Two adjacent apertures 208 have the same pitch P, where P = W + S, where W is the width (or diameter) of the aperture 208 and S is the minimum spacing (or partition thickness) between two adjacent apertures 208. In FIG. 2D, the apertures 208 are arranged in a staggered pattern with adjacent rows in succession, and the two adjacent rows are offset from each other by a predetermined distance (or offset). This offset may be W in one embodiment and may be any other suitable value in alternative embodiments. Each aperture 208 can be considered to be aligned with a hexagonal unit grid 240 in the grid network (represented by dashed lines in FIG. 2D ). Two adjacent apertures 208 have the same pitch P, where P = W + S, where W is the width (or diameter) of the aperture 208 and S is the minimum spacing (or partition thickness) between two adjacent apertures 208. In some alternative embodiments (not shown), the apertures 208 can be aligned with honeycomb-shaped unit grids 240 in the grid network. FIGS. 2E and 2F show variations of FIGS. 2C and 2D , respectively. One or more of the unit grids 240 are designed without any aperture(s) extending through the grid, which corresponds to having apertures filled (or closed) by partition material (represented by dashed circles in FIGS. 2E and 2F ). Such an aperture arrangement is considered to have a variable aperture pitch or to have closed holes, and therefore does not fall under the category of a parallel-hole collimator with uniformly distributed apertures. That is, the openings in Figures 2E and 2F are unevenly distributed.

[0023] 3 is a schematic cross-sectional view of the collimator 202 shown in FIG. 2B and detector 204 operating in conjunction to generate an image representing the distribution of a radiopharmaceutical within a target object 206 (e.g., a patient). The collimator 202 is disposed between the target object 206 and the detector 204 and is configured to filter radiation by blocking certain photons and passing others. In the illustrated embodiment, the components that connect and shield the collimator 202 and the detector 204 have been omitted for simplicity. In this embodiment, the collimator 202 is a parallel-hole collimator in which the openings or holes therein have rectangular cross-sections and are arranged parallel to one another. In alternative embodiments, the openings or holes in the collimator 202 may have diverging cross-sections, converging cross-sections, or other cross-sectional shapes.

[0024] The detector 204 includes an array of sensing pixels 214, such as a rectangular array, a square array, or other suitable array of pixels 214. In operation, each sensing pixel 214 independently records or monitors the amount of radiation incident on that sensing pixel and generates a signal associated with the amount of radiation (e.g., a voltage or current). The sensing pixels 214 may be substantially the same size and shape (e.g., circular, rectangular, or square). In various embodiments, the sensing pixels 214 are approximately 1x1 mm in size. 2 ~Approx. 5x5mm 2The pitch P' of the detector pixel array can range from less than 1 mm to approximately 6 mm in various embodiments. In one example, a CZT or silicon photomultiplier (SiPM)-based detector 204 can be fabricated to a size of 4 cm x 4 cm, where the detector 204 comprises a 16 x 16 detector pixel array with a unit pixel size of 2.5 mm x 2.5 mm. In some embodiments, the detector 204 includes appropriate electronic circuitry (e.g., an ASIC) for collecting and processing signals generated from the detector pixels 214. In some other embodiments, the detector pixels 214 are not physically distinct from one another, as is the case in most PMT-based detector systems, but are simply the product of digitizing a continuous detector surface. For example, the detector surface can be digitized as a grid network identical to the grid network of the collimator 202, with each unit grid on the detector surface corresponding to one opening in the collimator 202.

[0025] 3, for simplicity, it is assumed that the sensing pixel pitch P' is substantially equal to the aperture pitch P, and each sensing pixel 214 is directly below a corresponding aperture 208. Furthermore, the collimator 202 is closely coupled (or attached) to the detector 204 with a minimal spacing between them so that photons can propagate only along the aperture 208 and reach the corresponding sensing pixel 214 directly below it, although there may be minimal or negligible crosstalk (e.g., photons passing through a partition wall or photons passing through one aperture and impinging on a sensing pixel directly below another aperture). In this disclosure, the terms "closely coupled" and "attached" both refer to a spacing between the collimator and the detector that is less than half the thickness of the collimator, including cases where the collimator 202 and the detector 204 are in physical contact.

[0026] In a parallel-hole collimator, the imaging resolution rapidly decreases as the imaging source moves away from the collimator. The aperture size W and length (or effective length) H eFor a target object 206 located at a vertical distance Z above the top surface 210 of the collimator, the imaging resolution R c is given by the following equation (1).

[0027]

number

[0028] Effective opening length H e is H e =H - 2 / μ, where H is the aperture length and μ is the linear attenuation coefficient of the collimator material. For lead, μ = 22.43 cm at 150 keV. -1 ,

[0029]

number

[0030]

number

[0031] As can be seen from equation (1), the imaging resolution R c increases linearly with distance Z, and at the same distance Z, H e If is greater than R c becomes smaller (i.e., the imaging resolution becomes better). In other words, the longer the aperture (i.e., the larger H), the better the imaging resolution. However, a longer aperture has an adverse effect on signal sensitivity in an inverse square relationship. If the collimator efficiency, defined as the ratio of the amount of radiation (e.g., gamma rays) that passes through the collimator to the amount of radiation emitted from the source, is G, then G is expressed as He -2 is directly proportional to

[0032]

number

[0033] FIG. 4 shows an alternative collimator design, referred to as a "spread field imaging (SFI) collimator," in which the aperture length is uniformly increased, resulting in higher imaging resolution without sacrificing signal sensitivity. Compared to the parallel-hole collimator 202 shown in FIG. 3, the SFI collimator 203 shown in FIG. 4 has at least two notable features. First, the spacing D between the SFI collimator 203 and the detector 204 is significantly larger than in the configuration of FIG. 3. Second, the apertures 208 of the SFI collimator 203 are not all identical but differ in at least one respect, as will be further described below.

[0034] By increasing the spacing D between the SFI collimator 203 and the detector 204, photons passing through one aperture 208 can impinge on the detector 204 over an extended area beyond the aperture 208. That is, radiation from the target object 206 passes through one aperture 208 and illuminates an extended area on the detector 204 that is substantially larger than the size of the aperture 208, extending to the area directly beneath the other adjacent aperture. In contrast, when the collimator 202 is closely coupled to the detector 204, as shown in FIG. 3, radiation from the target object 206 propagating through one aperture 208 illuminates only an area on the detector 204 that is substantially equal to the size of the aperture 208. As shown in FIG. 4, incident photons incident between lines A1 and A2, which define the acceptance angle Θ, pass through the aperture 208 to reach and "illuminate" the detector 204. The illuminating area of ​​each aperture 208 is the surface area of ​​the detector 204 bounded by the lines defining the acceptance angle of the aperture 208. For example, the illuminating area of ​​the central aperture 208 is defined by the surface area on the detector 204 between lines A1 and A2 that intersect the detector 204. The size and shape of the illuminating area depend on the profile of the corresponding aperture and the spacing D between the collimator 202 and the detector 204. In some embodiments, the illuminating area of ​​an aperture 208 on the detector 204 is at least twice the size of the corresponding unit grid, which includes the corresponding aperture 208 and half of the partition wall surrounding the corresponding aperture 208 (note that the boundary of a unit grid is at the center of the partition wall between that unit grid and an adjacent unit grid). Furthermore, the illuminating area of ​​one aperture 208 may overlap the illuminating area of ​​an adjacent aperture. For example, in various embodiments, more than 10%, 20%, or 50% of the illuminated area may overlap with other illuminated areas (in this case, only radiation passing through opening 208 is considered, excluding radiation transmitted through a partition surrounding opening 208). Figure 4 shows the overlap area between an illuminated area defined by lines A1 and A2 below opening 208 and another illuminated area defined by lines A1' and A2' below an adjacent opening.That is, a point (or sensing pixel 214) within the overlapping area of ​​the detector 204 can receive radiation from one or more apertures 208. Furthermore, in some embodiments, any one sensing pixel 214 in the detector 204 is simultaneously illuminated by no more than a predetermined percentage of the total number of apertures 208, where this percentage is less than 25%. In some embodiments, this percentage may be less than 10%, 5%, or 2%. In some other embodiments, the illuminated area of ​​any aperture is no more than a predetermined percentage of the total illuminated area (through all apertures) of the concatenated detectors 204, where this percentage may be less than 10%, 5%, or 2%. In contrast, the apertures in coded aperture collimators or multi-pinhole collimators typically illuminate a larger percentage of the total illuminated area than the embodiments of the present disclosure. For example, the illuminated area of ​​one aperture in these collimators may be greater than 25% or greater than 40% of the total illuminated area of ​​all apertures in each collimator, thereby introducing a significant amount of crosstalk into the imaging system. By limiting the illumination area of ​​any one aperture to 10%, 5%, or even 2% or less of the total illumination area (through all apertures) in the coupled detector 204 (depending on the needs of system performance), crosstalk from different apertures is effectively reduced and good performance in imaging resolution and signal sensitivity can be achieved.

[0035] Shielding 244 (typically a heavy metal material) is provided along the peripheral region to cover (or seal) the space between collimator 203 and detector 204, preventing radiation from passing through the space and reaching detector 204. While gaps are shown in FIG. 4 between shielding 244 and SFI collimator 203 and between shielding 244 and detector 204, this is for illustrative purposes only. In various embodiments, shielding 244 is deployed to provide a closed (or sealed) region where only radiation passing through collimator 203 can reach detector 204.

[0036] In this embodiment, the apertures 208 of the SFI collimator 203 are not all identical. The apertures 208 may also differ in aperture shape, aperture size, or aperture length. While different aperture sizes may result in different acceptance angles Θ, in this embodiment, the maximum acceptance angle Θ is less than 15°. As described above, an acceptance angle Θ less than 15° provides desirable imaging resolution. Furthermore, for a given collimator thickness H, a smaller acceptance angle Θ results in a smaller aperture size, and therefore a smaller collimator dimension. An acceptance angle Θ less than 15° provides a good compromise between designing a compact collimator and manufacturing difficulties (e.g., mechanical tolerances). Furthermore, an acceptance angle Θ greater than 15° increases crosstalk from adjacent apertures, resulting in reduced resolution and more complex post-imaging processing. In some applications, depending on system performance needs, the acceptance angle Θ can be less than approximately 10°, or even less than approximately 5°, to achieve higher imaging resolution and reduce crosstalk. In some embodiments, some of the apertures 208 of the SFI collimator 203 are closed, essentially blocking photons from passing through. Therefore, the aperture pitch also varies across the collimator 203. For example, in FIG. 5B, the pitch and partitions around 208′ are relatively different from the other partitions. Apertures with non-uniformity in at least one parameter of the aperture profile (e.g., aperture shape, aperture size, aperture length, aperture pitch, etc.) are referred to as “non-uniformly distributed apertures.” FIG. 5A shows one embodiment of an SFI collimator 203 with different aperture shapes (e.g., circular and square), aperture sizes, and different aperture pitches (caused by blocked apertures represented by dashed circles). In FIG. 5A, each aperture 208 is aligned (e.g., center-to-center) with a respective unit grid 240. FIG. 5B shows yet another embodiment of an SFI collimator 203, in which at least one aperture 208′ is offset from the unit grid 240. An SFI collimator 203 with such an array of apertures can be considered a parallel-hole collimator with non-uniformly distributed apertures.5A and 5B show only a portion of the SFI collimator 203. In one embodiment, the total number of apertures 208 on the collimator 203 is greater than one thousand (1,000), such as approximately two thousand (2,000), five thousand (5,000), ten thousand (10,000), thirty thousand (30,000), or even one hundred thousand (100,000). Because each aperture illuminates only a small portion of the detector 204 surface area, and there is significant crosstalk or multiplexing between the apertures 208, a large number of apertures is beneficial for higher sensitivity, a better range of field of view (FOV), and efficient utilization of the detector 204 surface area.

[0037] In some embodiments, some or all of the openings 208 in the collimator 203 can be arranged in a repeating pattern, such as repeating patterns of size 3x3, 4x4, 3x5, 5x5, etc. The repeating pattern is referred to as a base pattern. In some embodiments, the base pattern has an odd number of rows and / or columns such that there is one central hole in the pattern; such a pattern may be advantageous for data processing. FIG. 6A shows a plan view (or a portion of a plan view) of one embodiment of the SFI collimator 203. At least a portion of the SFI collimator 203 includes a repeating base pattern of a 5x5 unit cell. In one base pattern illustrated in FIG. 6B, white pixels represent open openings and gray pixels represent closed openings. The open openings are surrounded by partitions. The dotted lines in FIG. 6A represent boundaries between base patterns. In the illustrated embodiment, the openings are aligned with a square unit grid and can expand to the size of the entire unit grid. The base pattern can be a coded aperture pattern, such as a URA, MURA, PBA, random, or pseudorandom pattern. In some embodiments, the number of repeating patterns is greater than 2x2, such as 3x3, 3x5, or 5x5. The basic pattern illustrated in FIG. 6A is a MURA 5 pattern, which is repeated three times in the X direction and three times in the Y direction to form a 15x15 size collimator (or portion of a collimator). In various embodiments, the basic pattern can be repeated any number of times in each of the X and Y directions, as needed for system requirements. FIG. 6C shows a plan view of another embodiment of an SFI collimator 203 (or portion thereof), in which the MURA 5 basic pattern is repeated 25 times in the X direction and 25 times in the Y direction to form a 125x125 size collimator (or portion of a collimator). In the illustrated embodiment, the openings have the same shape and size. In some embodiments, the openings may vary in shape, size, or both.

[0038] Returning to FIG. 4 , the spacing D between the collimator 203 and the detector 204 can be selected so that the illumination area of ​​each opening 208 on the detector 204 is equal to or less than the repeating basic pattern, but is at least twice the spacing (pitch) of the openings. The size of the illumination area of ​​the openings 208 on the detector 204 is substantially equal to the size of the basic pattern of the collimator 203, as illustrated by dotted lines B1 and B2 in FIG. 4 . As described in more detail below, making the size of the illumination area substantially equal to the basic pattern (or an integer multiple of the basic pattern) provides a fast image reconstruction algorithm that significantly improves system efficiency and accuracy. As a result, the spacing D between the collimator 203 and the detector 204 is optimized when it is greater than half the aperture length (collimator thickness) H. In various embodiments, D can be less than 10 times H, for example, between 1.5 and 10 times H, between 1 and 7 times H, between 1.5 and 7 times H, or any other suitable range. If D is too small, such as less than 1.5 times H, there is no significant effect on improving resolution. On the other hand, if D is greater than 10 times H, crosstalk from neighboring apertures becomes high, increasing complexity during image reconstruction and degrading resolution. In some examples, D is approximately 1, 1.5, 2, 2.5, 3, or 3.5 times H.

[0039] Further, referring to FIG. 4, for a sensing pixel 214 of the detector 204, the area that can be seen through the aperture directly above it is equal to the imaging resolution R n It is also called and is given by the following equation (2).

[0040]

number

[0041] When a photon passes through one aperture and enters a detection region directly beneath another aperture, crosstalk or multiplexing exists. In contrast, for any given sensing pixel on the surface of a detector equipped with a parallel-hole collimator (e.g., collimator 202 shown in FIG. 3), photons can only pass through one aperture (excluding septum transmission, an event that parallel-hole collimator designs must mitigate). Because there are multiple possible paths a detected photon can take at a given location on the detector, and because for a given source, there are multiple apertures through which radiation can pass to reach the detector, this crosstalk effect typically reduces resolution. This is true even when all apertures are essentially identical, as in parallel-hole collimators.

[0042] On the other hand, for an SFI collimator 203 as shown in Figure 4, the apertures are not all the same. Specifically, the apertures within the visible region of a sensing pixel on the detector surface are not all the same (the visible region here means the collection of all apertures through which photons can pass to reach a sensing pixel on the detector surface). Thus, due to the different patterns of apertures below, light sources that are close to each other cast different shadows on the detector. These different shadows can cause the reconstruction algorithm to reconstruct the original light source distribution, i.e., the object image, as R. n This helps to recover the image with high resolution approaching 1. The detector surface is typically digitized as a network of grids, each grid representing the boundary of a sensing pixel. In some embodiments, the aperture patterns observed by adjacent sensing pixels are different.

[0043] Similar to single photon emission computed tomography (SPECT) imaging, images (also called projections) can be acquired from multiple angles by rotating a coupled pair of collimator (such as collimator 203) and detector (such as detector 204) around the target object. In this acquisition mode, the coupled collimator and detector move together without relative motion to each other. The forward projection p at a particular angle α can be expressed as:

[0044]

number

[0045]

number

[0046]

number

[0047] The algorithm in equation (5) requires a large number of voxels and pixels to be multiplied by the number of angles, which can be time-consuming and requires a large amount of storage space for the K matrix. Projections at a single angle can be digitized from 64x64 up to 512x512 pixels, so typically 60 or 120 angles are required. Furthermore, object images often have 128x128x128 or 256x256x256 voxels, so the matrix Kα(i,j) can be very large.

[0048] If the apertures are constructed by repeating a basic pattern and corresponding through-holes in all repeating patterns have the same shape and dimensions (even if the holes in the basic pattern are different), the matrix K and algorithm can be simplified. Typically, for a repeating basic pattern with size XxY (X and Y can be different), the number of possible patterns that a point source can project onto the detector is equal to a multiple of the product of XxY. Figures 6A and 6B show an example where the repeating pattern is a 5x5 (X=Y=5) MURA 5 pattern. Let m be the index (grid unit including open and closed holes) of the holes in the basic pattern, and m = 1,...,25. This example assumes a plane located at a distance Z from the collimator 203 (see Figure 4). A point source at any point on this plane projects a predetermined pattern onto the detector 204. Here, the size of the projected pattern is much smaller than the size of the detector, so the pattern that the point source can project onto the detector is only XxY=25. Depending on the index of the aperture directly under this point source (excluding the boundary area), psf m,Z The detector pixel size is denoted as (x,y) where x=-M,...M and y=-N,...N. Here, the size of the detector pixel is the same as or a fraction of the collimator pitch (hole spacing). Therefore, the 25 pattern can be reduced to a basic 5x5 pattern by indexing by aperture. On the other hand, there is a f pattern where the pixel (also called voxel) size is the same as the collimator pitch (hole spacing) and the pixels are aligned with the collimator pitch. ZThe image of an object at a distance Z expressed as f Z The image can be divided into 25 sub-images of the same size as the aperture, where in each sub-image, the pixels on the aperture with the same index as the aperture are kept the same, and the remaining pixels are set to 0. If the pattern index is m, then the slice f of the object image at distance Z is Z is the sub-image f Z It can be expressed as (j,m).

[0049]

number

[0050]

number

[0051]

number

[0052] Equation (7) also gives the projection pattern psf when the object image is composed of slices parallel to the collimator plane, denoted by a distance Z. m,zis the same for different angles α, which further simplifies the algorithm.

[0053] One way to implement equation (5) independently for each angle is to exclude α from the summation and calculate the estimate f along the distance Z. Z (j) is calculated as the sum of angle p'. α This is a method to obtain the secondary projection of the object image f from p'. α Since the projection geometry for deriving p' is similar to that of parallel-hole collimation, image reconstruction algorithms for parallel-hole collimation, such as filtered back projection (FBP), algebraic reconstruction (ART), and OSEM, can be used to obtain p'. α f can be reconstructed from

[0054] One alternative is to interpolate fz to f α,z where f α,z is constructed from slices parallel to the collimator plane (denoted by Z) at the above angle. Then, a forward projection and a back projection process are performed before applying the interpolation.

[0055]

number

[0056] Returning to FIG. 4 , with respect to collimator 203, the height H of the opening (i.e., the thickness of collimator 203) may be much greater than its opening width or diameter W. For example, the opening width W may be less than 2 mm, and the opening height H may be greater than 20 mm. However, for closed-type openings, the opening does not need to be completely filled; rather, it only needs to be filled thick enough to block radiation from passing through the opening. For example, if the radioisotope Tc-99m is used as the septum material, a 3 mm thick lead can block more than 99% of incident 140 keV gamma radiation. Therefore, the collimator can be manufactured with a lightweight design. Referring to FIGS. 8A and 8B , collimator 203 is a collimator with open and closed holes and includes two sections. One section is similar to a parallel-hole collimator with uniformly distributed through-holes (also referred to as parallel-hole section 262). The other section is a substantially thinner plate 264 (but thick enough to substantially block radiation) with unevenly distributed through-holes, with fewer through-holes (e.g., about half as many) than those in parallel-bore section 262. Collimator 203 shown in FIGS. 8A and 8B is also referred to as a multi-section collimator. The through-holes in thin plate 264 align with the through-holes in parallel-bore section 262 to together define the aperture pattern of collimator 203, while other through-holes in parallel-bore section 262 are blocked by the opaque material of thin plate 264. Thus, the through-holes in thin plate 264 define the final opening locations of collimator 203. Thin plate 264 is also referred to as a patterned aperture section 264. In some embodiments, all of the through-holes in parallel-bore section 262 are substantially identical, while the through-holes in patterned aperture section 264 are different and may include closed apertures. Additionally, the patterned aperture section 264 may be movable in some embodiments, such as operable to slide horizontally relative to the surface of the parallel bore section 262 .Thus, by moving the patterned aperture section 264 relative to the parallel hole section 262, multiple projections can be acquired at one angle.

[0057] The patterned aperture section 264 has a thickness H1 sufficient to block the target radiation. In various embodiments, the thickness H2 of the parallel-bore section 262 may be approximately two to ten times the thickness H1 of the coded aperture section 264. For example, a 25 mm aperture length can be constructed with a parallel-bore section 262 having a thickness of 20 mm and a patterned aperture section 264 having a thickness of 5 mm. In this way, the overall weight of the collimator can be significantly reduced compared to a completely filled closed aperture. The patterned aperture section 264 may be located on the side of the parallel-bore section facing the detector (FIG. 8A) or the side facing the target object (FIG. 8B), or in the middle of two sub-parallel-bore sections (e.g., sub-parallel-bore sections 262-a and 262-b in FIG. 8C, where H2-a + H2-b = H2). In fact, a closed aperture requires only a thin layer of heavy metal material to block radiation. This thin layer can be divided into multiple segments as needed. Thus, this new design starts by fabricating regular parallel through-holes, and then a thin layer can be placed anywhere within the open opening to create a closed opening.

[0058] While the apertures shown in FIGS. 8A-8C have the same cross-section and aperture width, many different designs exist that allow for different aperture profiles. For example, lightweight designs can also be applied to collimators with the same aperture pitch but different aperture sizes and partitions. Some such examples are shown in FIGS. 9A-9C. In FIG. 9A, collimator 203 has some apertures with a diameter (or width) W and some apertures with a smaller diameter (or width) W'. Instead of using partitions of constant width throughout the thickness of collimator 203 to create narrower aperture sidewalls, the partitions may have thicker sections only at the side edges. The thicknesses H1-a and H1-b at the side edges are sufficient to block target radiation and maintain a smaller acceptance angle Θ, similar to the narrower aperture design. The multi-section design concepts discussed above with reference to FIGS. 8A-8C can also be applied to the example collimator of FIG. 9A. FIG. 9B illustrates a collimator 203 having one parallel-bore section 262 and two patterned aperture sections 264-a / 264-b that define the aperture size. In some cases, only one of 264-a / 264-b is needed to adjust the acceptance angle Θ. The apertures may also have different shapes in both the vertical and horizontal cross sections. In some embodiments, the apertures may have different orientations. The patterned aperture sections 264-a / 264-b may define the aperture size by a pattern of open and closed apertures, as shown in FIG. 9C. Collectively referring to FIGS. 8A-9C, a single parallel-bore section 262 may be paired with different patterned aperture sections 264a and / or 264b to create different collimator configurations. Therefore, the multi-section collimator according to the present disclosure provides a lightweight design and a low-cost solution. In particular, it should be noted that Figures 8A-8C and 9A-9C illustrate embodiments in which adjacent sections of the multi-section collimator are in physical contact with one another.In an alternative embodiment, adjacent sections are closely coupled to each other without physical contact (to prevent photons from entering through one hole and exiting through another), as long as collimator performance is guaranteed.

[0059] 10A. Unlike the exemplary collimators shown in FIGS. 8A-9C, in which the parallel-hole section 262 and the patterned aperture section 264 are attached to one another, the exemplary collimator shown in FIG. 10A has the parallel-hole section 262 spaced apart from the patterned aperture section 264. Furthermore, the patterned aperture section 264 may be closely coupled to the detector 204. In the illustrated embodiment, the sensing pixels 214 of the detector 204 have corresponding openings in the patterned aperture section 264 to receive illumination from the parallel holes in the parallel-hole section 262 above. The openings in the patterned aperture section 264 are intentionally sized differently, resulting in a different number of parallel holes in the parallel-hole section 262 that can illuminate the pixel 214. For example, sensing pixel 214a has a corresponding opening in the patterned aperture section 264 that is larger than the adjacent sensing pixel 214b. As a result, sensing pixel 214a can be illuminated by multiple parallel holes in parallel hole section 262, bounded by lines C1 and C2 that intersect the bottom surface of parallel hole section 262. In contrast, pixel 214b, because the corresponding opening in patterned aperture section 264 is smaller, is illuminated only by a single parallel hole directly above it in parallel hole section 262, bounded by lines D1 and D2 that intersect the bottom surface of parallel hole section 262. The number of parallel holes in parallel hole section 262 that can illuminate the same sensing pixel 214 depends on the acceptance angle of the corresponding opening in patterned aperture section 264. The number of parallel holes in parallel hole section 262 that can illuminate the same sensing pixel 214 also depends on the distance H3 between parallel hole section 262 and patterned aperture section 264. Essentially, distance H3 allows some of the pixels 214 of detector 204 to be illuminated by at least two or more of the parallel holes 208. Note that in the alternative embodiment shown in FIG. 10B, it is the parallel-hole section 262 that is closely coupled to the detector 204, while the patterned aperture section 264 is spaced apart from 262.

[0060] The design presented in this invention can be adapted to other variants with multiple elongated holes, such as collimators with convergent or divergent holes, modified by the two features presented above.

[0061] 1A-1C, radiation imaging system 100 can use at least two different sets of coupled collimators and detectors. In one such embodiment, one set includes an SFI (extended-field imaging) collimator 203 (e.g., as shown in FIGS. 4, 6A-6C, 8A-8C, 9A-9C, or 10A-10B) spaced apart from the coupled detector, and the other set includes a parallel-bore collimator 202 attached to the coupled detector (e.g., as shown in FIG. 3). Typically, in this configuration, the aperture (and / or acceptance angle) of parallel-bore collimator 202 is smaller than that of SFI collimator 203. As shown in equation (1), parallel-bore collimators have good resolution at close ranges (small Z) due to their smaller apertures, but imaging resolution rapidly degrades as the target object moves away from the collimator. In contrast, the SFI collimator 203 provides superior imaging resolution when the target object moves away from the collimator. During multi-angle image acquisition, two sets of coupled collimators and detector modules rotate around the target object to acquire images (projections). Naturally, a particular portion of the target object will be closer to one collimator at a given angle and farther from the target object at the opposite angle, except for the portion close to the center of rotation (COR). By combining these two collimators, a greater weight can be assigned to the update coefficients estimated from the projections of the parallel-bore collimator 202 for positions closer to the parallel-bore collimator 202, and a greater weight can be assigned to the update coefficients estimated from the projections of the SFI collimator 203 for positions farther from the parallel-bore collimator 202. In other words, the weighting between the estimates of the two sets of coupled collimators and detector modules depends on the distance. Another advantage of this configuration is that a quick estimate of the object image can be obtained by reconstruction using only the parallel-hole collimator, which can be used as an initial estimate for iterative reconstruction using projections from both collimators.Methods for reconstructing projections from multi-angle parallel-hole sections can include filtered back projection (FBP), algebraic reconstruction (ART), and ordered subset expectation maximization (OSEM). To mitigate the low scan count due to the parallel-hole collimator of the present invention having only half the acquisition angle compared to conventional scanning in which both cameras are equipped with parallel-hole collimators, a step of performing image reconstruction at lower resolution (by applying low-pass filtering (LPF) and / or reconstructing larger pixels) may be performed. In another embodiment, at least two sets of coupled collimator and detector modules are used in the system, both of which include SFI collimators 203 (e.g., as shown in Figures 4, 6A-6C, 8A-8C, 9A-9C, or 10A-10B) positioned away from the coupled detector. However, the elongated through-holes of the two collimators have different acceptance angles. For example, the maximum acceptance angle of one SFI collimator 203 is at least twice that of the other SFI collimator 203 (e.g., 10° for one and 5° for the other). In another embodiment, at least two sets of coupled collimator and detector modules are used in the system, where at least one detector is coupled to an SFI collimator 203 having a long narrow hole (small acceptance angle) and positioned at a distance from the coupled detector (e.g., as shown in Figures 4, 6A-6C, 8A-8C, 9A-9C, or 10A-10B), and another detector is coupled to a multi-pinhole collimator, including a coded aperture collimator, where the maximum acceptance angle is significantly larger than that of the SFI collimator 203, such as greater than 20°.

[0062] Referring to FIG. 1B , when two identical sets of coupled collimators and detectors in opposing positions are used in the radiological imaging system 100, each set of coupled collimators and detectors only needs to rotate half a circle in equal sweeping steps. A full circle of images is acquired by combining data from the first and second semicircles. For example, in a 3° sweeping step, the first set sequentially acquires images at angles of 0°, 3°, 6°, ... 177°, and the second set sequentially acquires images at angles of 180°, 183°, 186°, ... 357°. When two different sets of coupled collimators and detectors are used in the radiological imaging system 100, each set "observes" differently and therefore must rotate one circle. To maintain the same operating time, each set sweeps one circle with twice the sweeping steps. For example, in a 6° sweep step, the first set sequentially acquires images at angles of 0°, 6°, 12°, ... 180°, 186°, ... 354°, while the second set sequentially acquires images at angles of 180°, 186°, 192°, ... 354°, 0°, ... 174°. For each given angle swept by the system, two images are acquired: one image from the first set and another image from the second set. In other words, the two sets "observe" differently, but each set repeatedly acquires an image from the same angle as the other set. In contrast, referring to FIG. 1C , the two different sets of coupled collimators and detectors are not perfectly opposite each other, but are offset by half the angle of the sweep step. For example, in a 6° sweep step, the directions 154A and 154B of the two sets form an angle Υ of approximately 3°. By offsetting the orientation of the two sets, system 100 can capture each image from a different angle.For example, with a sweep step of 6° and an offset of 3°, the first set would sequentially acquire images at angles of 0°, 6°, 12°, ... 180°, 186°, ... 354°, and the second set would sequentially acquire images at angles of 183°, 189°, 195°, ... 357°, 3°, ... 177°.

[0063] Reference is now made to FIGS. 11A and 11B. In a radiation imaging system, performance is optimized when the collimator apertures are aligned with the detector sensing pixels. FIG. 11A shows a plan view in the XY plane of a portion of an exemplary SFI collimator 203 (basic pattern shown in FIG. 6B) superimposed on a detector 204. FIG. 11B shows a schematic cross-sectional view in the XZ plane of a portion of the collimator 203 and detector 204 along section line AA in FIG. 11A. The detector 204 shown in FIGS. 10A and 10B includes an array of individual sensing pixels 214. The individual sensing pixels 214 may be based on a CZT or SiPM structure. Typically, the sensing pixels 214 are arranged in a grid network (also referred to as a pixel grid) formed by rows and columns, such as rows i=1, 2, 3, ... and columns j=1, 2, 3, ... in the illustrated embodiment. Any detector pixel 214 located at a position (x, y) in the XY plane can be represented by an index (i, j) in the grid network (i and j are integers) and has a corresponding aperture (or a coding pattern or a closed aperture in a partition) in the grid network (also called the aperture grid) of the collimator 203. Because the detector pixels 214 in Figures 11A and 11B are individually and physically distinct from one another, alignment between the detector and the collimator can be easily achieved by placing the detector pixel 214 directly below the corresponding aperture in the collimator 203 so that the pixel grid and the aperture grid overlap.

[0064] Reference is now made to FIGS. 12A and 12B. As described in connection with FIGS. 3 and 4, in some embodiments, the sensing pixels of a detector are not physically distinct from one another, as is the case in most PMT-based detector systems, but are simply the product of digitizing a continuous detector surface. FIG. 12A shows a plan view in the XY plane of a portion of an exemplary SFI collimator 203 (the basic pattern shown in FIG. 6B) superimposed on such a detector 204. FIG. 12B shows a schematic cross-sectional view in the XZ plane of a portion of the collimator 203 and detector 204 along section line AA in FIG. 12A. The detector surface is digitized as a pixel grid that is the same as the corresponding aperture grid of the collimator 203, i.e., has the same spacing between grid points. That is, each pixel unit grid on the detector surface has the closest corresponding aperture (or closed aperture in the coding pattern or septum) in the collimator 203. Each pixel unit grid on the detector surface is also referred to as a pixel of the detector. The detector pixels are arranged in a grid network with rows i = 1, 2, 3, ... and columns j = 1, 2, 3, ... and can be labeled as pixel (i, j) (i and j are integers). Ideally, in a perfectly aligned state, the detector pixel grid and the collimator aperture grid should overlap. However, as shown in Figures 12A and 12B, misalignment between the pixel grid and the aperture grid, such as during device assembly, can occur, resulting in a significant misalignment in the X direction (denoted by Δx) and / or the Y direction (denoted by Δy). Such misalignment can lead to system degradation. For example, as shown in Figure 12B, in a perfectly aligned state, an incident photon passing through the aperture of the collimator 203 that should be recorded as an event (e.g., a count) occurring at pixel (3,1) is instead recorded as an event occurring at pixel (3,2), thereby reducing the contrast of the aperture pattern in the recorded signal.

[0065] Taking photomultiplier tube (PMT)-based detection systems as an example (and other detection systems using other architectures as well), PMTs function by converting incident photons into photoelectrons at a photocathode. These electrons generate a large number of secondary electrons from a series of charged cathodes, generating a measurable current pulse at the anode. Based on the measurable current pulses at adjacent PMTs, a location (x, y) on a continuous XY plane can be determined as the location where the incident photon event occurred. Based on the mapping between the continuous XY plane and the index (i, j), the event occurring at location (x, y) is then digitized as an event occurring at pixel (i, j). However, as mentioned above, misalignment of the initial mapping between the continuous XY plane and the index (i, j) can degrade performance. Image contrast provides a way to compensate for misalignment due to offset. For the human eye and mind to recognize an image, contrast must be present in the acquired image array. Any given part of the subject requires different signal intensities to be recorded at different pixels. The simplest example is transmission contrast caused by partially or fully opaque structures, such as collimator openings and closed apertures (or septa). The amount of contrast present in the image determines the accuracy required to adjust for misalignment. FIG. 13 illustrates a setup that algorithmically compensates for misalignment without physically moving the collimator. In FIG. 13, a flood light source 242, such as a Co-57 or Tc-99m rectangular flood source, provides a uniform illumination field above the collimator 203. The flood light source 242 can be positioned closely adjacent to the collimator 203, such as less than half the thickness of the collimator 203, even when the collimator 203 and the flood light source 242 are in physical contact. When illuminated by the flood light source 242, the image acquired by the detector 204 is an image of the flood light source 242, including the shadow cast by the collimator 203. Because the illumination field is uniform, the contrast in the acquired image is due to the transparent and opaque characteristics of the collimator 203.That is, pixels under a closed aperture (or partition) or a small aperture will have a lower image signal strength compared to pixels under an open aperture or a large aperture.

[0066] Contrast can be defined as a metric of variation (e.g., ratio or absolute difference) between a single intensity peak and a valley in an image. As an example, the signal intensity line 244 in FIG. 13 shows V p and V v One way to maximize contrast is to p / V v the ratio of, or

[0067]

number

[0068] Because the detector surface is digitized into a pixel grid, misalignment can be compensated for by an algorithm, i.e., contrast maximization can be achieved without physically moving the collimator. In one exemplary method, each signal emerging from the detector surface (e.g., a current pulse from the PMT in the case of a PMT-based detection system) is recorded as an event with coordinates (x, y). Thus, each entry in the data list (also called list mode) is an attribute vector containing information about a single detected photon. For example, a series of events can be recorded as (x n ,y n ) (n=1,2,3,...), where each pair represents the X- and Y-coordinates of the nth detected event. Based on the mapping between the continuous XY plane and the discrete index (i,j), the total number of records for pixel (i,j), denoted as C(i,j), is

[0069]

number

[0070]

number

[0071] Next, to "offset" the pixel grid, an offset pair (δx, δy) is introduced into the digitization function and C(i,j) is recalculated. Thus, C(i,j) is

[0072]

number

[0073]

number

[0074] Alternatively, instead of sweeping δx and δy together, the method may first perform a sweep in the X direction only (by fixing δy) to obtain an optimized offset δx in the X direction, and then perform a sweep in the Y direction only (using the optimized δx) to obtain an optimized offset δy in the Y direction. Alternatively, the method may first perform a sweep in the Y direction, followed by a sweep in the X direction. Furthermore, the method may use a V sweep from the entire pixel grid to calculate contrast. p and V v Instead of selecting offsets (δx, δy) two-dimensionally, the method may select a row (by fixing i in the set of C(i,j), such as signal intensity lines 244 in FIG. 13), sweep (δx, δy) to obtain the maximum contrast in the row, and apply the selected offset pair (δx, δy) to the entire pixel grid. Similarly, the method may select a column (by fixing j in the set of C(i,j)), sweep (δx, δy) to obtain the maximum contrast in the column, and apply the selected offset pair (δx, δy) to the entire pixel grid.

[0075] Note that different hole patterns may require different optimization methods. In some patterns, the maximum or minimum recorded signal intensity descriptor (e.g., contrast, signal intensity at peaks, signal intensity at valleys, etc.) is used to maximize / minimize the maximum contrast, maximum peak value, or minimum valley value. When a repeating hole pattern is used, corresponding pixels (corresponding to the same hole in the base pattern) can be added together to reduce randomness (i.e., quantum noise) in the pixel values. Also, in some embodiments, when a particular hole pattern is used, the selected offset pair (δx, δy) from the optimization routine is located at a fixed offset away from the actual optimal offset, which may be determined by the hole pattern. For example, in some hole patterns, the maximum pixel value in the flood light image is located at the center of an adjacent 2x2 open opening. In that case, the fixed offset is half the hole pitch, and the actual optimal offset used to align pixels to the holes is corrected by the fixed offset. That is, after determining the selected offset pair (δx, δy), a further step may optionally be performed of adding or subtracting a fixed offset to or from the selected offset pair (δx, δy) to obtain an actual optimum offset to use for misregistration adjustment.

[0076] Without intending to be limiting, one or more embodiments of the present disclosure provide many advantages for imaging radiation in a target object, such as a patient. For example, a collimator design having a repeating coding pattern spaced apart from an associated detector provides superior imaging resolution without sacrificing signal sensitivity, thus improving system performance.

[0077] The foregoing description outlines features of several embodiments so that those skilled in the art may better understand the aspects of the present disclosure. Those skilled in the art will appreciate that they may readily use this disclosure as a basis for designing or modifying other processes and structures which carry out the same purposes and / or achieve the same advantages as the embodiments presented herein. Those skilled in the art should also recognize that such equivalent structures do not depart from the spirit and scope of the present disclosure, and that various changes, substitutions, and alterations can be made herein without departing from the spirit and scope of the present disclosure. It is therefore appropriate that the scope of the appended claims be accorded the broadest interpretation consistent with the present disclosure.

Claims

1. 1. A radiation imaging system, comprising: a collimator configured to filter radiation emitted from a target object, a plurality of openings non-uniformly distributed on the collimator; a maximum acceptance angle of the plurality of openings is 15° or less; the plurality of openings form a predetermined opening pattern including a base pattern, the base pattern being repeated more than four times within the opening pattern; the collimator; a detector for detecting radiation that has passed through the collimator; and the collimator is positioned at a distance from the detector such that a point on an upper surface of the detector facing the collimator is simultaneously illuminated by two or more of the plurality of openings; Radiation imaging system.

2. 2. The radiation imaging system of claim 1, wherein a point on the top surface of the detector is illuminated by a predetermined percentage of the plurality of apertures, the percentage being less than about 25%.

3. 2. The radiation imaging system of claim 1, wherein the number of the plurality of apertures exceeds 1,000.

4. 2. The radiation imaging system according to claim 1, wherein the plurality of apertures are different in at least one of aperture size, aperture shape, light acceptance angle, aperture length, and aperture pitch.

5. 2. The radiation imaging system of claim 1, wherein the distance between the collimator and the detector is about 1.5 to about 10 times the thickness of the collimator.

6. 2. A radiation imaging system according to claim 1, wherein said basic pattern is a coded aperture pattern.

7. 2. The radiation imaging system according to claim 1, wherein an illumination area of ​​one of said plurality of apertures is substantially equal to an area of ​​said basic pattern.

8. 2. The radiation imaging system of claim 1, wherein the collimator includes a first portion and a second portion, the first portion including through holes uniformly distributed on the first portion, and the second portion including through holes non-uniformly distributed on the second portion and corresponding to the plurality of openings, whereby the through holes of the first portion are blocked by the second portion.

9. 9. The radiation imaging system of claim 8, wherein the thickness of the first portion is about 2 to about 10 times that of the second portion.

10. 10. The radiation imaging system of claim 1, wherein the collimator is a first collimator and the detector is a first detector, the system further comprising: a second collimator; and a second detector coupled to the second collimator; and the second collimator is attached to the second detector; Radiation imaging system.

11. 1. A radiation imaging system, comprising: a first collimator configured to filter radiation emitted from a target object, the first collimator including a first plurality of apertures forming a first aperture pattern; a first detector associated with the first collimator to detect radiation that has passed through the first collimator; a second collimator configured to filter radiation emitted from the target object, a second plurality of openings forming a second opening pattern; the second plurality of openings are non-uniformly distributed on the second collimator; a maximum acceptance angle of the second plurality of openings is less than or equal to 15°; the second aperture pattern is repeated more than four times within the second aperture pattern. the second collimator; a second detector associated with the second collimator, the second detector detecting radiation passing through the second collimator; and the target object is disposed between the first collimator and the second collimator, and the first aperture pattern is different from the second aperture pattern. Radiation imaging system.

12. 12. The radiation imaging system of claim 11, wherein the first collimator is in contact with the first detector.

13. 13. The radiation imaging system according to claim 12, wherein the distance between the second collimator and the second detector is about 1.5 to about 7 times the thickness of the second collimator.

14. 12. The radiological imaging system of claim 11, wherein the first plurality of apertures are uniformly distributed on the first collimator.

15. 12. The radiation imaging system of claim 11, wherein the basic pattern comprises a coding pattern.

16. 16. The radiation imaging system of claim 15, wherein the coding pattern is one of a uniformly redundant array (URA) pattern, a modified uniformly redundant array (MURA) pattern, and a perfect binary array (PBA) pattern.

17. 16. A radiation imaging system according to claim 15, wherein a normal direction to a top surface of the first detector and a normal direction to a top surface of the second detector form a non-zero angle.

18. 1. A method for correcting misalignment in an imaging system, comprising: providing a collimator and a detector having a digitized pixel grid; illuminating the collimator with a flood light source; recording a signal intensity at each pixel of the digitized pixel grid, the signal intensity being generated by illuminating the collimator; deriving descriptors from said recorded signal intensities; generating a series of offsets introduced into the digitized pixel grid; determining, from said set of offsets, a selected offset that optimizes said descriptor; reconstructing the digitized pixel grid based on an actual offset derived from the selected offset; A method comprising:

19. 20. The method of claim 18, wherein the selected offsets include a pair of offset values ​​in two respective orthogonal directions.

20. 20. The method of claim 18, wherein the descriptor is one of a contrast value, a peak value, and a valley value of the recorded signal intensity.

Images