(19)
(11)EP 2 633 494 B1

(12)EUROPEAN PATENT SPECIFICATION

(45)Mention of the grant of the patent:
11.12.2019 Bulletin 2019/50

(21)Application number: 11784781.4

(22)Date of filing:  17.10.2011
(51)International Patent Classification (IPC): 
G06T 5/00(2006.01)
G06T 11/00(2006.01)
(86)International application number:
PCT/IB2011/054588
(87)International publication number:
WO 2012/056364 (03.05.2012 Gazette  2012/18)

(54)

LOW DOSE CT DENOISING

ENTRAUSCHEN VON NIEDRIGDOSIERTER CT

DÉBRUITAGE DE DONNÉES DE TOMODENSITOMÉTRIE PAR ORDINATEUR À FAIBLE DOSE


(84)Designated Contracting States:
AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

(30)Priority: 27.10.2010 US 407040 P

(43)Date of publication of application:
04.09.2013 Bulletin 2013/36

(73)Proprietor: Koninklijke Philips N.V.
5656 AE Eindhoven (NL)

(72)Inventors:
  • BROWN, Kevin, M.
    NL-5656 AE Eindhoven (NL)
  • ZABIC, Stanislav
    NL-5656 AE Eindhoven (NL)

(74)Representative: Philips Intellectual Property & Standards 
High Tech Campus 5
5656 AE Eindhoven
5656 AE Eindhoven (NL)


(56)References cited: : 
WO-A2-2010/011676
US-B1- 6 493 416
US-A1- 2003 076 988
  
  • HSIEH JIANG: "Adaptive streak artifact reduction in computed tomography resulting from excessive x-ray photon noise", MEDICAL PHYSICS, AIP, MELVILLE, NY, US, vol. 25, no. 11, 1 November 1998 (1998-11-01), pages 2139-2147, XP012010348, ISSN: 0094-2405, DOI: 10.1118/1.598410
  • WANG J ET AL: "Penalized Weighted Least-Squares Approach to Sinogram Noise Reduction and Image Reconstruction for Low-Dose X-Ray Computed Tomography", IEEE TRANSACTIONS ON MEDICAL IMAGING, IEEE SERVICE CENTER, PISCATAWAY, NJ, US, vol. 25, no. 10, 1 October 2006 (2006-10-01), pages 1272-1283, XP001545734, ISSN: 0278-0062, DOI: 10.1109/TMI.2006.882141
  
Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention).


Description


[0001] The following generally relates to de-noising data and finds particular application to computed tomography (CT) and is also amenable to other imaging modalities such as a hybrid PET/CT system, a digital x-ray system, and/or other imaging modality.

[0002] A multi-slice computed tomography (CT) scanner includes an x-ray tube mounted on a rotatable gantry that rotates around an examination region about a longitudinal or z-axis. The x-ray tube emits radiation that traverses the examination region and a subject or object therein. A two-dimensional detector array subtends an angular arc opposite the examination region from the x-ray tube. The detector array includes a plurality of rows of detectors that are aligned with respect to each other and that extend along the z-axis. The detectors detect radiation that traverses the examination region and the subject or object therein and generate projection data indicative thereof. A reconstructor processes the projection data and reconstructs three-dimensional (3D) volumetric image data indicative thereof. The volumetric image data is processed to generate one or more images of the examination region, including the portion of the subject or object disposed therein.

[0003] Unfortunately, CT scanners emit ionizing radiation and thus expose the patient to ionizing radiation, which may increase risk of cancer. Generally, the radiation dose deposited in the patient depends on multiple factors, including, but not limited to, tube current (mAs), tube voltage (kVp), pitch/exposure time (for helical scans), slice thickness and spacing (for axial scans), the number of scans in a study, and patient build (e.g., thicker or thinner). The deposited dose can be reduced by decreasing tube current, tube voltage and/or the number of scans, and/or increasing the pitch, slice thickness and/or slice spacing. However, image noise is inversely proportional to radiation dose, and thus reducing radiation dose not only reduces the dose deposited in the patient but also increases image noise in the acquired data, which is propagated to the images during reconstruction, reducing image quality (i.e., noisier images), which may degrade the diagnostic value of the procedure.

[0004] Image-based de-noising algorithms have been applied. However, they have difficulty in dealing with "streaky" images, where the noise is strongly correlated between neighboring voxels of the image data. Generally, very low levels of photon flux in projection measurements generates streaks in the reconstructed images. Also, when the mean number of detected photons is very low (e.g., <10), the logarithm operation introduces a bias, which can show up in images as a shifted mean CT number. Iterative reconstructions such as the Maximum Likelihood (ML) based reconstructions have the potential to yield improved images in these cases. However, such ML based reconstructions are extremely computationally expensive, which currently hinders their use in routine practice.

[0005] US 2003/076988 A1 discloses a method for treating noise in low-dose computed tomography projections and reconstructed images. The method comprises acquiring raw data at a low mA value, applying a domain specific filter in a sinogram domain of the raw data, and applying an edge preserving smoothing filter in an image domain of the raw data after filtering in the sinogram domain.

[0006] Aspects of the present application address the above-referenced matters and others.

[0007] According to one aspect, a system according to claim 1 is provided.

[0008] According to another aspect, a method according to claim 9 is provided.

[0009] The invention may take form in various components and arrangements of components, and in various steps and arrangements of steps. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention.

FIGURE 1 illustrates an example imaging system in connection with a projection data de-noiser.

FIGURE 2 illustrates an example projection data de-noiser.

FIGURE 3 illustrates an example method for de-noising projection data with the projection data de-noiser.



[0010] FIGURE 1 illustrates an imaging system 100 such as a computed tomography (CT) scanner. The imaging system 100 includes a generally stationary gantry 102 and a rotating gantry 104. The rotating gantry 104 is rotatably supported by the stationary gantry 102 and rotates around an examination region 106 about a longitudinal or z-axis 108.

[0011] A radiation source 110, such as an x-ray tube, is rotatably supported by the rotating gantry 104. The radiation source 110 rotates with the rotating gantry 104 and emits radiation that traverses the examination region 106. A source collimator includes collimation members that collimate the radiation to form a generally cone, wedge, fan or other shaped radiation beam.

[0012] A two-dimensional radiation sensitive detector array 112 subtends an angular arc opposite the radiation source 110 across the examination region 106. The detector array 112 includes a plurality of rows of detectors that extend along the z-axis 108 direction. The detector array 112 detects radiation traversing the examination region 106 and generates projection data indicative thereof.

[0013] A projection data de-noiser 114 de-noises projection data. As descried in greater detail below, in one instance, the de-noiser 114 employs an algorithm in which projections corresponding to a lower number of photons are de-noised more aggressively than projections corresponding to a higher number of photons. Such an algorithm allows for reducing streaks and/or bias in the reconstructed images due to a lower number of photons (e.g., due to patient size, a low dose scan, etc.) while preserving strong gradients (i.e., edges) in the data and thus image resolution.

[0014] A reconstructor 116 reconstructs the de-noised projection data and generates three-dimensional (3D) volumetric image data indicative thereof. The reconstructor 116 may employ a conventional 3D filtered-backprojection reconstruction, a cone beam algorithm, an iterative algorithm and/or other algorithm.

[0015] A patient support 118, such as a couch, supports an object or subject such as a human patient in the examination region 106.

[0016] A general-purpose computing system or computer serves as an operator console 120. A processor of the console 120 executes computer readable instructions on the console 126, which allows an operator to control operation of the system 100 such as selecting a full dose or low dose scan protocol, activating projection data de-noising, initiating scanning, etc.

[0017] In the illustrated embodiment, the projection data de-noiser 114 is shown as a separate component. In another embodiment, the projection data de-noiser 114 is part of the console 120 and/or other computing device.

[0018] FIGURE 2 illustrates an example of the projection data de-noiser 114. The illustrated projection data de-noiser 114 includes a data unlogger 200 that unlogs the projection data, which converts the attenuation line integrals into detected photons.

[0019] A photon estimator 202 estimates a number of detected photons for each projection and generates a signal indicative thereof. The photon estimator 202 can employ various approaches to estimate the number of photons. By way of example, the detected number of photons can be estimated as the mean number of detected photons for each projection. Additionally, this mean may be smoothed, for example, using a moving average, which may facilitate mitigating large deviations in the Poisson random variable for a very small number of detected photons. Other techniques including but not limited generally known techniques for estimating the number of detected photons are also contemplated herein.

[0020] A de-noiser 204 de-noises the detected photon signal based on the estimated noise. The de-noiser 204 can de-noise the detected photons for a two dimensional projection based solely on the two dimensional projection or on the two dimensional projection and one or more neighboring two dimensional projections. As noted above, the de-noiser 204 may employ an algorithm which de-noises projections corresponding to a lower number of photons more aggressively than projections corresponding to a higher number of photons, which facilitates mitigating streaks while preserving edges.

[0021] A data logger 206 logs the de-noised detected photons, which converts the de-noised detected photons back into attenuation line integrals, which can be reconstructed by the reconstructor 116.

[0022] An evaluator 208 can be utilized to determine whether a given projection should be de-noised based on a predetermined photon number threshold. In this case, where the estimated number of photons for a projection indicates that there is a sufficient number of photons for the projection, the projection is not de-noised. Otherwise, the projection is de-noised. De-noising only those projections that are deemed to not have enough photons may increase processing speed relative to de-noising every projection. In another embodiment, the evaluator 208 can be omitted.

[0023] The following provides a non-limiting example of a suitable de-noising algorithm, which is based on a total variation-minimization algorithm, treating each 2D projection as an image. A similar 3D method with appropriate redesign can also be used.

[0024] Computed tomography (CT) projection data can be represented as shown in Equation 1:

wherein I represents the measured photons, I0 represents the input photons, µx represents the attenuation function, and l represents the acquisition line.

[0025] Given the noise characteristics of photon measurements, / represents the mean photon measurements, and a single actual measurement f is a realization of Poisson random variables with mean I and probability



[0026] The mean photon measurement I can be estimated from f by minimizing a cost function expressed as a sum of a total variation term and a weighted least-squares term as shown in Equation 2:

wherein v represents a general statistical weighting that gives a preference to the original projection when the noise is small and a preference to projections with small total variation when the noise is large.

[0027] Generally, the total variation term dominates when the estimated number of detected photons is lower and the weighted least-squares term dominates when the estimated number of detected photons is higher.

[0028] In Equation 2, β is an optional variable that represents a tuning parameter which controls the aggressiveness of the algorithm smoothing in which smaller values of β lead to more aggressive smoothing overall.

[0029] Equation 2 can be solved by discretizing the Euler-Lagrange partial differential equation (PDE) for as shown in Equation 3:



[0030] Equation 3 can be expressed in terms of an estimated mean number of detected photons for each projection as shown in Equation 4:

wherein ρsm(m,r) represents the estimated mean number of detected photons, m represents the number of detectors, and r represents the number of rows. As described in greater detail below, ρsm(m,r) in Equation 4 is derived by taking a smoothed version of the original measurement f; however, the estimated mean number of photons need not be smoothed. It is reasonable to replace v in Equation 3 with ρsm(m,r) as the noise variance of the logged measurements is proportional to 1/ ρsm(m,r).

[0031] By scaling as shown in Equation 4, small total variation in regions of high noise (low photons) and closeness to the original image in regions of high photon counts are preferred. Note that other forms of scaling the detected photons by β are also possible. For example, in another embodiment, β can vary with view angle.

[0032] The mean number of detected photons can be estimated as follows.

[0033] Starting from an input 2-D projection of the form of ρ(m,r), where the data is in the logged attenuation domain, the detected number of photons associated with this measurement can be estimated as the mean number of detected photons ρsm(m,r) for each projection.

[0034] This mean or a smoothed mean can alternately be employed. A smoothed mean may facilitate mitigating large deviations in the Poisson random variable for a very small number of detected photons. With the latter case, the estimated number of photons is smoothed, for example, using a moving average over the estimated number of detected photons, as shown in EQUATION 5:

wherein j ∈ {m-n:m+n}, k ∈ {r-n:r+n}, nKernel represents a smoothing kernel, and n =

Other smoothing approaches are also contemplated herein. For detectors at the edges of the projection, values can be extrapolated or otherwise determined to fill the necessary buffer.

[0035] Equation 4 can be solved by using Equation 5 based on an iterative fixed point approach in which each update i+1 is determined from a previous update image i according to EQUATION 6:

wherein ρeff(m, r) = β · ρsm(m, r) and Wp represents weights, which can be variously computed, for example, as known in the art.

[0036] The de-noised detected photons are logged as shown in EQUATION 7:

wherein ρout(m,r) represents the output logged de-noised detected photons, ρdn(m,r) represents the de-noised detected photons, Nr(m,r) represents the number of photons, and s represents a scaling factor. EQUATION 7 converts the de-noised detected photons back into attenuation line integrals for reconstruction.

[0037] The foregoing de-noising algorithm is computationally efficient, for example, because the de-noising operation is applied once to the projection data at acquisition, followed by a single filtered-backprojection reconstruction, and does not required the heavy computational burden of multiple forward- and backprojections in a Maximum Likelihood iterative approach.

[0038] Furthermore, the foregoing de-noising algorithm provides significant streak reduction to the reconstructed images, while not affecting the projections with good statistics, relative to a configuration in which the de-noising algorithm is not applied the projections. By applying the de-noising algorithm in the unlogged projection domain, bias from a low photon count can also be mitigated.

[0039] FIGURE 3 illustrates an example method for de-noising projection data.

[0040] At 302, a plurality of two-dimensional projection signals (projection data) is acquired. The projection can be generated by the system 100 and/or other imaging system.

[0041] At 304, a mean number of detected photons is estimated for each of the projections;
At 306, at least a sub-set of the projections are de-noised using an algorithm which gives preference to the original projections when the noise is small and preference to projections with small total variation when the noise is large as described herein. As noted above, projection having a sufficient number of photons need not be de-noised.

[0042] At 308, the de-noised projections are converted into de-noised attenuation data.

[0043] At 310, the de-noised attenuation data are reconstructed to generate one or more images.

[0044] The above may be implemented by way of computer readable instructions, which when executed by a computer processor(s), cause the processor(s) to carry out the described acts. In such a case, the instructions are stored in a computer readable storage medium associated with or otherwise accessible to the relevant computer.

[0045] It is to be appreciated that the projection data de-noiser 114 can be implemented through one or more processors that execute one or more computer readable and/or executable instructions stored or encoded on computer readable storage medium such as physical memory. Additionally or alternatively, the instructions can be stored on transitory medium such as signal medium or the like.

[0046] The invention has been described herein with reference to the various embodiments. Modifications and alterations may occur to others upon reading the description herein. It is intended that the invention be construed as including all such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.


Claims

1. A system, comprising:

a source (110) configured to rotate about an examination region and to emit radiation that traverses the examination region;

a radiation sensitive detector array (112) configured to detect radiation traversing the examination region and to generate projection data indicative of the detected radiation;

an evaluator (208) configured to determine whether a projection is de-noised based on a number of detected photons for the projection; and

a projection data de-noiser (114) configured to de-noise the projection databased on the number of detected photons for the projection,
wherein the projection data de-noiser further comprises a de-noiser (204) configured to de-noise projections based on minimizing a cost function, including at least two terms, a first total variation term and a second general statistically weighted least-square term that gives a preference to the original projection when the noise is small and a preference to projections with small total variation when the noise is large.


 
2. The system of claim 1, the projection data de-noiser further comprising:
a photon estimator (202) configured to estimate a number of detected photons for each projection, wherein the projection data de-noiser is configured to de-noise the projection data based on the estimated number of detected photons.
 
3. The system of claim 2, wherein the estimated number of detected photons represents a smoothed moving average over the estimated number of detected photons.
 
4. The system of claim 1, wherein the de-noiser (204) is configured to employ an iterative algorithm to minimize the cost function.
 
5. The system of claim 1, further comprising:
a reconstructor (116) configured to reconstruct the de-noised projection data to generate volumetric image data.
 
6. The system of claim 1, wherein the projection data de-noiser (114) is configured to de-noise projections having a lower number of photons to a greater degree relative to projections having a higher number of photons.
 
7. The system of claim 1, wherein at least one projection includes a number of detected photons that satisfies a predetermined number of photons threshold and is not de-noised, and at least one projection includes a number of detected photons that does not satisfy the predetermined number of photons threshold and is de-noised.
 
8. The system of claim 1, wherein the projection data de-noiser (114) is configured to reduce noise while maintaining a given image resolution.
 
9. A method, comprising:

obtaining projection data generated by an imaging system;

determining whether a projection is de-noised based on a number of detected photons for the projection; and

de-noise the projection databased on the number of detected photons for the projection,
wherein projections are de-noised based on minimizing a cost function, including at least two terms, a first total variation term and a second general statistically weighted least-square term that gives a preference to the original projection when the noise is small and a preference to projections with small total variation when the noise is large.


 
10. The method of claim 9, wherein a degree of the de-nosing is proportional to the estimated number of detected photons for the projection, and at least two projections are de-noised to two different degrees.
 
11. The method of claim 9, further comprising:
de-noising only those projections having an estimated number of photons that do not satisfy a predetermined threshold and de-noising remaining projections to a degree based on a respective estimated number of photons.
 
12. The method of claim 9, further comprising:
smoothing the estimated number of detected photons by using a moving average over the estimated number of detected photons, wherein the projection is de-noised based on the smoothed estimated number of detected photons.
 


Ansprüche

1. System, umfassend:

eine Quelle (110), die konfiguriert ist, um sich um einen Untersuchungsbereich zu drehen und Strahlung zu emittieren, die den Untersuchungsbereich durchläuft;

eine strahlungsempfindliche Detektoranordnung (112), die konfiguriert ist, um Strahlung zu erfassen, die den Untersuchungsbereich durchläuft, und um Projektionsdaten zu erzeugen, die die erfasste Strahlung anzeigen;

einen Auswerter (208), der konfiguriert ist, um zu bestimmen, ob eine Projektion basierend auf einer Anzahl von erfassten Photonen für die Projektion entrauscht ist; und

einen Projektionsdaten-Entrauscher (114), das konfiguriert ist, um die Projektionsdaten basierend auf der Anzahl der erfassten Photonen für die Projektion zu entrauschen,

wobei der Projektionsdaten-Entrauscher weiter einen Entrauscher (204) umfasst, der konfiguriert ist, um Projektionen basierend auf der Minimierung einer Kostenfunktion zu entrauschen, die mindestens zwei Terme, einen ersten Gesamtvariationsterm und einen zweiten allgemeinen statistisch gewichteten kleinsten quadratischen Term beinhalten, der eine Präferenz für die ursprüngliche Projektion gibt, wenn das Rauschen klein ist, und eine Präferenz für Projektionen mit geringer Gesamtvariation, wenn das Rauschen groß ist.


 
2. System nach Anspruch 1, wobei der Projektionsdaten-Entrauscher ferner umfasst: einen Photonenschätzer (202), der konfiguriert ist, um eine Anzahl von erfassten Photonen für jede Projektion zu schätzen, wobei der Projektionsdaten-Entrauscher konfiguriert ist, um die Projektionsdaten basierend auf der geschätzten Anzahl von erfassten Photonen zu entrauschen.
 
3. System nach Anspruch 2, wobei die geschätzte Anzahl der erfassten Photonen einen geglätteten gleitenden Durchschnitt über die geschätzte Anzahl der erfassten Photonen darstellt.
 
4. System nach Anspruch 1, wobei der Entrauscher (204) konfiguriert ist, um einen iterativen Algorithmus zum Minimieren der Kostenfunktion zu verwenden.
 
5. System nach Anspruch 1, weiter umfassend:
einen Rekonstruktor (116), der konfiguriert ist, um die entrauschten Projektionsdaten zu rekonstruieren, um volumetrische Bilddaten zu erzeugen.
 
6. System nach Anspruch 1, wobei der Projektionsdaten-Entrauscher (114) konfiguriert ist, um Projektionen mit einer geringeren Anzahl von Photonen in größerem Maße im Vergleich zu Projektionen mit einer höheren Anzahl von Photonen zu entrauschen.
 
7. System nach Anspruch 1, wobei mindestens ein Vorsprung eine Anzahl von erfassten Photonen beinhaltet, die eine vorbestimmte Anzahl von Photonenschwellen erfüllt und nicht entrauscht ist, und mindestens ein Vorsprung eine Anzahl von erfassten Photonen beinhaltet, die nicht die vorbestimmte Anzahl von Photonenschwellen erfüllt und entrauscht ist.
 
8. System nach Anspruch 1, wobei der Projektionsdaten-Entrauscher (114) konfiguriert ist, um Rauschen zu reduzieren und gleichzeitig eine gegebene Bildauflösung beizubehalten.
 
9. Verfahren, umfassend:

Erhalten von Projektionsdaten, die von einem Bildgebungssystem erzeugt werden;

Bestimmen, ob eine Projektion basierend auf einer Anzahl von erfassten Photonen für die Projektion entrauscht ist; und

Entrauschen der Projektionsdaten basierend auf der Anzahl der erfassten Photonen für die Projektion,

wobei Projektionen entrauscht werden, basierend auf der Minimierung einer Kostenfunktion, die mindestens zwei Terme, einen ersten Gesamtvariationsterm und einen zweiten allgemeinen statistisch gewichteten kleinsten quadratischen Term beinhaltet, der eine Präferenz für die ursprüngliche Projektion gibt, wenn das Rauschen klein ist, und eine Präferenz für Projektionen mit geringer Gesamtvariation, wenn das Rauschen groß ist.


 
10. Verfahren nach Anspruch 9, wobei ein Grad der Entriegelung proportional zur geschätzten Anzahl der erfassten Photonen für die Projektion ist und mindestens zwei Projektionen in zwei verschiedenen Graden entrauscht werden.
 
11. Verfahren nach Anspruch 9, weiter Folgendes umfassend:
Entrauschen nur derjenigen Projektionen, die eine geschätzte Anzahl von Photonen aufweisen, die einen vorbestimmten Schwellenwert nicht erfüllen, und Entrauschen der verbleibenden Projektionen in einem Maße, das auf einer entsprechenden geschätzten Anzahl von Photonen basiert.
 
12. Verfahren nach Anspruch 9, weiter umfassend:
Glätten der geschätzten Anzahl von erfassten Photonen unter Verwendung eines gleitenden Mittelwerts über die geschätzte Anzahl von erfassten Photonen, worin die Projektion basierend auf der geglätteten geschätzten Anzahl von erfassten Photonen geräuschlos ist.
 


Revendications

1. Système comprenant :

une source (110) configurée pour tourner autour d'une région d'examen et émettre un rayonnement qui traverse la région d'examen ;

un réseau de détecteurs sensibles au rayonnement (112) configuré pour détecter le rayonnement traversant la région d'examen et générer des données de projection indicatives du rayonnement détecté ;

un évaluateur (208) configuré pour déterminer si une projection est débruitée sur la base d'un nombre de photons détectés pour la projection ; et

un débruiteur de données de projection (114) configuré pour débruiter les données de projection sur la base du nombre de photos détectés pour la projection,

dans lequel le débruiteur de données de projection comprend en outre un débruiteur (204) configuré pour débruiter des projections sur la base de la minimisation d'une fonction de coût incluant au moins deux termes, un premier terme de variation totale et un second terme général de moindres carrés pondéré au plan statistique qui donne une préférence à la projection originale lorsque le bruit est faible et une préférence aux projections de faible variation totale lorsque le bruit est important.


 
2. Système selon la revendication 1, le débruiteur de données de projection comprenant en outre : un estimateur de photons (202) configuré pour estimer un nombre de photons détectés pour chaque projection, dans lequel le débruiteur de données de projection est configuré pour débruiter les données de projection sur la base du nombre estimé de photons détectés.
 
3. Système selon la revendication 2, dans lequel le nombre estimé de photons détectés représente une moyenne mobile lissée sur le nombre estimé de photons détectés.
 
4. Système selon la revendication 1, dans lequel le débruiteur (204) est configuré pour employer un algorithme itératif pour minimiser la fonction de coût.
 
5. Système selon la revendication 1, comprenant en outre :
un reconstructeur (116) configuré pour reconstruire les données de projection débruitées afin de générer des données d'image volumétriques.
 
6. Système selon la revendication 1, dans lequel le débruiteur de données de projection (114) est configuré pour débruiter des projections ayant un nombre inférieur de photons à un degré supérieur par rapport à des projections ayant un nombre plus élevé de photons.
 
7. Système selon la revendication 1, dans lequel au moins une projection inclut un nombre de photons détectés qui satisfait à un seuil de nombre prédéterminé de photons et n'est pas débruité et au moins une projection inclut un nombre de photons détecté qui ne satisfait pas au seuil de nombre déterminé de photons et est débruité.
 
8. Système selon la revendication 1, dans lequel le débruiteur de données de projection (114) est configuré pour réduire le bruit tout en maintenant une résolution d'image donnée.
 
9. Procédé comprenant :

l'obtention de données de projection générées par un système d'imagerie ;

la détermination du fait qu'une projection est ou non débruitée sur la base d'un nombre de photons détectés pour la projection ; et

le débruitage des données de projection sur la base du nombre de photons détectés pour la projection,

dans lequel les projections sont débruitées sur la base de la minimisation d'une fonction de coût, incluant au moins deux termes, un premier terme de variation totale et un second terme général de moindres carrés statistiquement pondéré qui donne une préférence à la projection originale lorsque le bruit est faible et une préférence aux projections de faible variation totale lorsque le bruit est important.


 
10. Procédé selon la revendication 9, dans lequel un degré du débruitage est proportionnel au nombre estimé de photons détectés pour la projection et au moins deux projections sont débruitées à deux différents degrés.
 
11. Procédé selon la revendication 9, comprenant en outre :
le débruitage uniquement des projections ayant un nombre estimé de photons qui ne satisfait pas à un seuil prédéterminé et le débruitage de projections restantes à un degré basé sur un nombre estimé respectif de photons.
 
12. Procédé selon la revendication 9, comprenant en outre :
le lissage du nombre estimé de photons détectés en utilisant une moyenne mobile sur le nombre estimé de photons détectés, dans lequel la projection est débruitée sur la base du nombre estimé lissé de photons détectés.
 




Drawing











Cited references

REFERENCES CITED IN THE DESCRIPTION



This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.

Patent documents cited in the description