Method for ultrasonic determination of a corrected image of a medium and device for carrying out this method - Patents.com

The method corrects ultrasound images of viscoelastic media by compensating for probe-induced pressure variations through geometric transformations and regression analysis, resulting in accurate measurements of media properties.

JP7579271B2Active Publication Date: 2024-11-07SUPERSONIC IMAGINE SA
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
JP2021563097
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Priority Date
2019-05-17
Filing Date
2020-05-14
Publication Date
2024-11-07
Estimated Expiration
2040-05-14

AI Technical Summary

Technical Problem

Existing ultrasound-based imaging methods for viscoelastic media are sensitive to variations in pressure exerted by the ultrasound imaging probe, leading to image distortions and errors in representation and determination.

Method used

A method for ultrasound determination of corrected images that involves determining geometric changes between successive images using cross-correlation and geometric transformations to compensate for probe-induced pressure variations, followed by calculations of corrected strain and elasticity images using linear and nonlinear regression.

Benefits of technology

The method produces images that are less affected by probe pressure variations, providing accurate and precise measurements of viscoelastic media properties.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007579271000023
    Figure 0007579271000023
  • Figure 0007579271000024
    Figure 0007579271000024
  • Figure 0007579271000025
    Figure 0007579271000025
Patent Text Reader

Abstract

A method for ultrasonic determination of corrected images of a medium, comprising the steps of: obtaining time-sequential images of the medium (I k ), determining the motion for each image of the medium, and calculating the time-sequential partial distortion images (IΔε k ) and the partial distortion image (IΔε k ) to obtain the corrected partial distortion image (IΔε k * ) and the preceding corrected partial distortion image (IΔε k * ) to obtain the corrected distortion image (Iε k * ) determining the
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present disclosure relates to a method and device for imaging with ultrasound to provide an image of a viscoelastic medium. In particular, the object of the method is to correct the image produced by taking into account variations in the pressure exerted by an ultrasound imaging probe on the outer surface of the medium. [Background technology]

[0002] More specifically, the present disclosure relates to imaging methods that use ultrasound to view diffuse viscoelastic media, including particles that reflect said ultrasonic compressional waves.

[0003] The images corrected by the method may be of different types: for example, the images may be of the distortion type or of the elastic type.

[0004] US 2005 / 252295 proposes an elastic imaging technique in which elastic shear waves are generated in a medium and the propagation of these is observed by recording multiple images of the medium. Images of at least one motion parameter of the medium (displacement, partial strain) are then determined and shear wave propagation parameters such as shear wave velocity, shear modulus, Young's modulus, shear elasticity, etc. are estimated from the images.

[0005] This last technique is widely used. However, it has been found that the images of the motion parameters or the shear wave propagation parameters are sensitive to the pressure applied by the ultrasound imaging probe to the surface of the medium. However, this probe pressure is applied by the practitioner and can vary during the examination or from examination to examination. In effect, the practitioner varies this external pressure in order to make the desired observations in the image of the object.

[0006] In addition, these images contain distortions related to variations in the applied external pressure, which give rise to errors in image representation, image repeatability, and image determination. [Prior art documents]

Patent Documents

[0007]

Patent Document 1

Patent Document 2

Non-Patent Documents

[0008]

Non-Patent Document 1

Non-Patent Document 2

Non-Patent Document 3

Non-Patent Document 4

[0009] The aim of the invention is to improve this type of imaging method, in particular to improve the image itself. [Means for solving the problem]

[0010] To that end, the present disclosure provides a method for ultrasound determination of a corrected image of a medium, the method comprising: -Images of time-sequential media (I k ), where k is an image index between 0 and N, and the first image of the sequence is assumed to be undistorted; Determination of the motion for each image of the medium with index k between 1 and N, where the motion is determined for the image (I k ) and the image of the preceding medium with index k-1 (I k-1 ) a decision that is between Image of the medium with index k (I k ) and the image of the preceding medium with index k-1 (I k-1 ) based on the time-sequential partial strain image (IΔε k ) decision, said partial distortion image (IΔε) based on a preceding motion, meaning a motion relative to the image of the medium with indices between 1 and k, inclusive; k ) to obtain the corrected partial distortion image (IΔεk * ) decision, · Corrected partial distortion image (IΔε k * ) to obtain the corrected distortion image (Iε k * ) Determination The present invention proposes a method characterized by the above.

[0011] With these configurations it is possible to ensure that the images produced are correct.

[0012] In various embodiments of methods according to the present disclosure, recourse may optionally be provided to one or more of the following configurations:

[0013] According to one embodiment, the determination of the movement at index k is Determination of the displacement area between the image of the medium and the image of the preceding medium; Determining a geometric change of an image based on the displacement region, the geometric change representing the displacement region with less than 10 parameters. It is executed by.

[0014] According to one embodiment, the displacement area is calculated by cross-correlation between the media image and the preceding media image or by an algorithm for tracking sub-images between the media image and the preceding media image.

[0015] According to one embodiment, the tracking algorithm is the Lucas-Kanade algorithm.

[0016] According to one embodiment, the geometric transformation includes at least one translation, or a translation and a similarity transformation, or a translation, a similarity transformation, and a rotation.

[0017] According to one embodiment, the geometric change T k includes translation, similarity transformation, and rotation, and the geometric transformations are put into matrix form such that:

number

[0018] According to one embodiment, the parameter of the geometric change is obtained by the median value of a population of parameters, calculated based on a group of points in the displacement region.

[0019] According to one aspect, A group contains between 3 and 10 points in the displacement domain, · A population is larger than a group of one hundred individuals.

[0020] According to one aspect, the method comprises: (e1) determining an elasticity image of the medium, and (e2) Determination of a corrected elasticity image by compensation of said elasticity image based on a previous motion, meaning a motion relative to the image of the medium having an index between 0 and k (including 0 and k itself). Further includes:

[0021] According to one embodiment, the elasticity image produced in step (e1) is obtained by the following sub-steps: (e1.1) an excitation step in which shear waves are generated in the medium by emitting at least one focused ultrasonic wave; (e1.2) an observation step in which the shear wave propagation is observed by determining intermediate images of the medium successive in time, j being an intermediate image index between 0 and M; (e1.3) a processing step in which an elasticity image is determined based on said intermediate image of the medium and based on a shear wave propagation model. Generated by:

[0022] According to one aspect, -Determination of multiple corrected distortion images, determining a plurality of compensated elasticity images, the compensated strain images and the elasticity images being interlaced in time; and Determination of an image of nonlinear parameters based on a number of corrected distortion images and a number of corrected elasticity images. is further executed.

[0023] According to one aspect, the value of each pixel of the nonlinear parameter image is determined by strain linear regression of pairs of values, a first value of each pair corresponds to a value of the same pixel in the compensated strain image among the plurality and a second value of the pair corresponds to a value of the same pixel in the elasticity image among the plurality, and the compensated strain images among the plurality and the compensated elasticity images among the plurality are consecutive in time or close in time; A strain linear regression is established based on the following relationship between strain and elasticity:

number

[0024] According to one aspect, Determination of time-successive partial stress images, which are corrected by multiplication of the corrected partial strain image of index k with the corrected elasticity image of index k, Determination of the corrected stress image by summing the corrected partial strain images with indices between 1 and k, including 1 and k itself is further executed.

[0025] According to one aspect, Determination of multiple corrected stress images, determining a plurality of compensated elasticity images, the compensated strain images and the elasticity images being interlaced in time; and Determination of nonlinear parameter images based on multiple corrected stress images and multiple corrected elasticity images is further executed.

[0026] According to one embodiment, the value of each pixel of the nonlinear parameter image is determined by stress linear regression of pairs of values, a first value of each pair corresponds to a value of the same pixel of the corrected stress image in the plurality and a second value of the pair corresponds to a value of the same pixel of the elasticity image in the plurality, and the corrected stress images in the plurality and the corrected elasticity images in the plurality are consecutive in time or close in time; A stress linear regression is established based on the following relationship between stress and elasticity:

number

[0027] The present disclosure also relates to an imaging device comprising an ultrasound probe and a microcomputer suitable for implementing the method for ultrasound determination of a corrected image of a medium as described above.

[0028] Other characteristics and advantages of the present disclosure will become apparent from the following description of various embodiments, given as non-limiting examples, with reference to the attached drawings, in which: [Brief description of the drawings]

[0029] [Figure 1] 1 is a schematic diagram of an ultrasound imaging device according to an embodiment of the present disclosure. [Figure 2A]1A-1C are schematic diagrams illustrating the movement of an image of a medium between zero external pressure and an external pressure deforming the outer surface of the medium. [Figure 2B] 1A-1C are schematic diagrams illustrating the movement of an image of a medium between zero external pressure and an external pressure deforming the outer surface of the medium. [Figure 2C] FIG. 2C is a schematic diagram of a more complex motion than that of FIG. 2B, which motion includes rotation. [Diagram 3] FIG. 2 illustrates an embodiment of a method for calculating a corrected distortion image. [Figure 4] FIG. 1 illustrates a method for determining a corrected elasticity image. [Diagram 5] FIG. 2 illustrates a first embodiment of a method for determining an image of a corrective nonlinear parameter. [Figure 6] 6 is a graph illustrating an example implementation of the method of FIG. 5. [Figure 7A] FIG. 1 illustrates a method for determining a corrected stress image. [Figure 7B] FIG. 2 illustrates a second embodiment of a method for determining an image of corrective nonlinear parameters. [Figure 8] 7C is a graph illustrating an example implementation of the method of FIG. 7B. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0030] In the various figures provided for illustration, the same reference numbers represent identical or similar elements.

[0031] Imaging device 1 The imaging device 1 shown in Fig. 1 is intended to scatter ultrasonic compressional waves and provide an image of a viscoelastic medium 2, which may be a living body, for example a part of a patient's body (chest, liver, abdomen, etc.) in the case of medical applications. This imaging device 1 is also suitable for investigating the propagation of elastic shear waves in order to provide an elastic image of the medium 2.

[0032] The image of the medium is created, for example, by a microcomputer 4 or any other electronic processing unit (comprising at least an input interface 4b, such as a keyboard, and an output interface 4a, such as a screen). The microcomputer 4 or any other electronic processing unit sends ultrasonic compressional waves into the medium 2 from its outer surface 3, which interact with scattering particles 5 contained in the medium 2, said particles being reflective to the ultrasonic compressional waves. The particles 5 may be constituted by any inhomogeneity in the medium 2, and in particular by collagen particles present in the tissues of the human body, in the case of medical applications of interest (these particles form dots on the ultrasound image known as "speckle").

[0033] To observe the medium 2 and generate an image of the medium, an ultrasonic probe 6 is used which is positioned relative to the outer surface 3 of the medium 2 to be observed. The probe transmits pulses of ultrasonic compressional waves, of the type commonly used in ultrasound, along the Z axis, at a frequency, for example, between 0.5 and 100 MHz, preferably between 0.5 and 15 MHz, for example around 4 MHz.

[0034] The ultrasonic probe 6 consists of an array of n ultrasonic transducers T1, T2, . . . , Ti, . . . , Tn, where n is an integer equal to or greater than 1.

[0035] This probe 6 may for example be in the form of a linear array, which may for example comprise n=128 transducers aligned along an X-axis perpendicular to the Z-axis. The subject probe may also be a two-dimensional array (planar or not) of transducers.

[0036] The transducers T1, T2, ... Tn can be controlled independently of one another by a microcomputer 4, possibly a central processing unit CPU, for example contained in an electronic cabinet 7, which is connected to the probe 6 by a flexible cable. The transducers T1 to Tn can therefore be selectively discharged as follows: "Planar" ultrasonic compressional waves (whose wavefronts are linear in the XZ plane in this case) illuminating the entire field of view in the medium 2, or any other type of unfocused wave, e.g. waves generated by causing random acoustic signals to be emitted by the various transducers T1 to Tn. Or ultrasonic compression waves focused at one or more points in the medium 2

[0037] US Patent Application Publication No. 2009 / 234230 proposes a composite imaging technique that uses several unfocused compressional waves, e.g. plane wave type waves at different angles, which combine the return waves of these plane waves to very quickly obtain an image of the medium with improved quality.

[0038] Media Image I k Decision To generate an image (I) of the medium, the imaging device performs, for example, the following steps: (i1) the microcomputer 4 causes the probe 6 to emit at least one unfocused ultrasonic compressional wave into the viscoelastic medium; (i2) the microcomputer 4 causes the probe 6 to detect and record in real time acoustic signals received from the viscoelastic medium 2, including reverberations generated by the unfocused ultrasonic compressional waves interacting with the reflective particles 5 of the viscoelastic medium; (i3) a processing step in which the microcomputer 4 processes the acoustic signals received from the viscoelastic medium 2 during sub-step (i2) in order to determine one or more images (I) of the medium.

[0039] Unfocused ultrasound compressional waves propagate through a medium 2 with a very high propagation speed in the human body, for example about 1500 m / s, and interact with reflective particles 5. This produces echoes or other similar interferences in the signal, which are known per se in the ultrasound field by the name "speckle noise".

[0040] Such "speckle noise" is captured by the transducers T1, ..., Tn during sub-step (i2) after emitting unfocused ultrasonic compressional waves. The signal si(t) thus captured by each transducer Ti is first sampled at a high frequency (e.g., 30-100 MHz) and digitized in real time by samplers E1, E2, ... En, respectively, which are part of the cabinet 7 and are connected to this transducer.

[0041] The signal si(t) thus sampled and digitized is stored, also in real time, in a memory Mi which is part of the cabinet 7 and which is dedicated to the transducer Ti.

[0042] For example, each memory Mi has a capacity of about 128 MB and contains all the received signals si(t).

[0043] After storing all the signals si(t), the central processing unit CPU causes these signals to be reprocessed by a summing circuit S which is part of the cabinet 7 (or which performs this processing itself, or else said processing can be performed in the microcomputer 4), by a conventional beamforming process corresponding to substep (i3).

[0044] Thus, signals S(x,z) are generated, each corresponding to an image of a field of view of the medium after emission of unfocused ultrasound waves.

[0045] For example, the signal S(t) may be determined by the following equation:

number

[0046] The above formulas apply, mutatis mutandis, when the field of view is three-dimensional (a two-dimensional array of transducers), by replacing the spatial coordinates in the plane (x,z) with the spatial coordinates (x,y,z).

[0047] After an optional beam shaping step, the central processing unit CPU stores the image signals S(x,z) corresponding to the last firing in a central memory M that is part of the cabinet 7. These signals can also be stored in the microcomputer 4 so that the microcomputer 4 itself performs the calculation of the image of the medium (I).

[0048] There are other techniques for generating an image (I) of the medium, such as synthetic imaging techniques. Any imaging technique that makes it possible to obtain an image of the medium can be used. Preferably, a technique is used that makes it possible to obtain an image at high speed.

[0049] The imaging device 1 and method according to the present disclosure obtains time-sequential images of the medium to track variations in an external pressure P applied to the outer surface 3. Thus, N+1 images of the medium are captured at successive times, which are not necessarily separated by a constant period of time, and the images of the medium can thus be identified by an image index k between 0 and N.

[0050] Therefore, the image of the medium is I k It will be identified by the notation.

[0051] This index k is used for any other images or quantities determined based on the image of the media at index k, and possibly for previous images of the media, and thus this index can be used to look up the time associated with the image of this media at index k.

[0052] In addition, by convention, the first image of this sequence is denoted as the image of the medium with index 0. In this first image, it is presumed that the external pressure P is low or zero and that the strain induced on the outer surface 3 is low or zero. This first image is represented, for example, in FIG. 2A.

[0053] 2A-2C illustrate the problem solved by the present disclosure.

[0054] In Fig. 2A, use of the imaging device 1 is depicted, where a probe 6 applies a low or zero external pressure P on the outer surface 3. The outer surface 3 remains approximately horizontal (in the X direction). An image I0 of the medium includes, for example, an inclusion 2i at a depth Z1 relative to the outer surface 3.

[0055] In Fig. 2B, at a later time than in Fig. 2A, use of the imaging device 1 with an external pressure P' greater than P is represented, which deforms the outer surface 3 inwardly in a direction Z into the medium 2. The imaging device 1 creates an image I1 of the medium of the same size as the image I0 created in Fig. 2A, but the upper layer of the medium above the inclusion 2i is naturally more compressed than the lower layer of the medium 2 below the inclusion 2i, so that the inclusion 2i in the image I1 of this medium in Fig. 2B is displaced towards the probe 6 compared to the first image I0 in Fig. 2A. In other words, the inclusion 2i is now at a depth Z2 relative to the outer surface 3, which is shallower than the depth Z1 in the first image I0.

[0056] Thus, during the variation of the external pressure, the internal elements of the medium are displaced in the image of the medium due to the elasticity of said medium 2. Thus, in order to perform an accurate tracking and analysis of the properties of these internal elements, it is necessary to know their general motion.

[0057] The examples of Figures 2A and 2B show that Z translation can provide suitable correction.

[0058] In Fig. 2C, the use of the imaging device 1 is shown at another time, e.g. subsequent to the previous time shown in Figs. 2A and 2B, with a different external pressure P and above all with a tilt of the probe 6 with respect to the vertical direction Z (the tilt is exaggerated in this Fig. 2C). The outer surface 3 is deformed in the Z direction and tilted at an angle θ. The imaging device 1 creates an image I2 of the medium of the same size, but the upper layer of the medium is still compressed and the inclusions 2i are laterally displaced in the image, i.e. in the X direction due to the tilt of the probe 6.

[0059] Thus, pressure variations and the position of the probe 6 on the outer surface 3 will displace internal elements of the medium in the image of the medium and it appears necessary to know these general motions (which are very slow compared to ultrasound) in order to track and / or compensate for the internal elements in the captured image of the medium so as to perform an accurate analysis of the properties of the internal elements.

[0060] The probe 6 may in some cases have more complex motions than a simple translation in the Z direction or a combination of translation and rotation, where the required corrective motions become more complex, and the details of determining these motions will be explained further below in this disclosure.

[0061] Displacement area u k Decision Media image (I k ) are two of the images of the medium with index k (I k ) and the image of the medium with index k-1 (I k-1 ) can be processed by correlation, advantageously by cross-correlation.

[0062] The cross-correlation can be carried out in a dedicated electronic circuit DSP, for example which is part of the cabinet 7, or can be programmed in the central processing unit CPU or microcomputer 4.

[0063] During this cross-correlation process, the cross-correlation function is calculated to determine the displacement caused by each particle 5 that causes the echo of the ultrasound. k-1 (x,z),S k (x,z)> is maximized.

[0064] Examples of such cross-correlation calculations are given in the state of the art, in particular in the following documents: · "Internal displacement and strain imaging using speckle tracking", O'Donnell et al., IEEE transactions on ultrasonics, ferroelectrics, and frequency control, vol. 41, no. 3, May 1994, pp. 314-325 ·"Elastography: a quantitative method for imaging the elasticity of biological tissues", Ophir et al., Ultrasound imaging, Vol. 13, pp. 111-134, 1991 ·"Pyramidal Implementation of the Lucas Kanade Feature Tracker. Description of the algorithm.", JY Bouguet, Intel Corp.

[0065] Thus, a set of displacement vectors, i.e., u(x,z,t) at each position (x,z) of the medium 2 is obtained, and by replacing the variable at time t with the image index k, u k (x,z). These displacement vectors can be optionally simplified to a single component, or to two or three components. In the example under consideration, the displacement region u at image index k can be expressed as k is the following equation.

number

[0066] This displacement field (set of displacement vectors) is stored in the memory M or in the microcomputer 4.

[0067] Image of the displacement region Iu k and an image of the displacement region Iu corresponding to the time-sequential images of the medium. k can in particular be viewed by means of a screen 4a of a microcomputer, for example in the form of a slow-motion film, in which the displacement values ​​are illustrated by optical parameters such as a grey scale or a colour scale.

[0068] Partial distortion image IΔε k Decision It is now possible to calculate the partial strain Δε at each point of the medium 2, i.e. a vector whose components are the derivatives of the components of the displacement vector with respect to the spatial variables (X,Z coordinates in the example under consideration), i.e.

number

[0069] Optionally, according to one exemplary implementation, the strain / partial strain is calculated only in the Z direction. In fact, the compression related to the external pressure of the probe is mainly in this direction. In the same manner, in the subsequent stress calculation, the stress / partial stress is optionally calculated only in the Z direction. The calculation is thus simplified.

[0070] As applied to the displacement vectors, the partial strain images (IΔε k ) can be constructed from a set of partial distortions computed at any point of coordinates (x,z) in the XZ plane (image plane).

[0071] These partial distortion vectors or partial distortion images (IΔε k) can in particular be viewed in the form of a slow-motion film on the screen 4a of the microcomputer.

[0072] Determination of the motion of a medium due to an external pressure The strain of the medium 2 linked to variations in the external pressure P applied to the outer surface 3 is a slow, almost uniform elastic strain. This strain propagates inside the viscoelastic medium 2 and causes displacements of the particles 5 and elements contained within the medium 2.

[0073] This displacement is detrimental to the analysis of the properties of various points of the image of the medium, in particular to the evaluation of the distortion at these points and / or the elasticity of these points.

[0074] The present disclosure therefore contemplates correcting the image to compensate for these movements, thereby allowing the particles and / or internal elements of medium 2 to be maintained in an almost constant position in the image, thus allowing their characteristics to be examined more accurately.

[0075] The external distortion caused by an external pressure P is expressed as the image of the medium with index k (I k ) and the image of the preceding medium with index k-1 (I k-1 ) can be thought of as causing motion between the

[0076] Here, the motion is the displacement area (u k ) can be determined by

[0077] In particular, this motion corresponds to a geometric change T k This can be modeled by, in a general and simple way, at the step of image index k, the total displacement of the displacement domain, u k (x,z), i.e., a very small number of parameters, e.g., less than 10.

[0078] Now, compensation for any type of image based on prior motion can be done by applying the geometric transformation above, i.e., the geometric transformation T i , where index i runs from 1 to k.

[0079] According to a first variant, the geometric change T k includes translation as seen between FIG. 2A and FIG. 2B.

[0080] According to a second variant, the geometric change T k includes translation and similarity transformations.

[0081] According to a second variant, the geometric change T k includes translations, similarity transformations, and rotations.

[0082] Thus, in the XZ plane (image plane), the geometric transformation can be put into a matrix form of the following type:

number

[0083] Thus, considering three points P1, P2, and P3 in the image with coordinates (x1,z1), (x2,z2), and (x3,z3) respectively, we obtain the following relationship:

number

[0084] In addition, based on the above matrices A and B, there is a relation that makes it possible to calculate the geometric transformation Tk, where matrix A corresponds to the coordinates of three points in the image, and matrix B corresponds to the coordinates of these same three points with their displacement (movement). Thus, the matrix of geometric transformation is obtained by the matrix product of matrix B and the inverse matrix of matrix A. T k =BA -1

[0085] By applying this formula, the geometric change T k It becomes possible to calculate

[0086] Conversely, once the geometric change T k Once established, this relation makes it possible, by inverse correlation, to know the coordinates (x,z) of any point P. B=T k .A

[0087] Here, by the matrix product of the preceding geometric transformations, i.e., the geometric transformation T i where index i runs from 1 to k, compensation can be performed for any type of image based on prior motion.

[0088] The above relationships established for three points can be generalized to groups of points in an image, where the groups of points contain 3, 4, 5, or 6 points in an image. The groups of points consist of between 3 and 10 points in an image.

[0089] In addition, in a variant, a cluster of three points (groups of points) of the image can be advantageously chosen, said cluster comprising a number Ng of high-value groups of points, for example the size of this cluster, the number Ng of groups of points, being greater than one hundred.

[0090] Here, the geometric change T k The parameters of are obtained by the median of the parameters calculated from the population of a group of image points.

[0091] In particular, the technique estimates the motion between an image of a medium with index k and a preceding image of a medium with index k−1 by using a displacement region u k can be applied to images.

[0092] By using a collection of groups of points, it is possible to determine a geometric change that more generally represents the motion between an image of a medium at index k and an image of a preceding medium at index k-1.

[0093] Advantageously, the selection of points for each group is performed randomly in the images. Due to this arrangement, it is possible to determine the geometric changes which represent the motion between the images in a more general and reliable manner.

[0094] Corrected partial distortion image IΔε k * Decision Partial distortion image IΔε k are now corrected to return these images to the situation according to the image I0 of the first medium, thus obtaining these partial distortion images IΔε k Remove the motion caused by the strain from the external pressure P in the

[0095] This correction is performed for each partial distortion image IΔε k This is done by using a set of general movements determined between each image of the medium for I0. More precisely, the inverse movements must be applied to return to the situation in the first image I0.

[0096] In other words, the corrected partial distortion image IΔε k * is the corrected partial distortion Δε k *and based on the motion associated with the image of the medium having indices between 1 and k, inclusive, said partial distortion image IΔε k The compensation is calculated based on the following:

[0097] Corrected distortion image (Iε k * ) Determination Strain ε k (x,z) is obtained by summing the successive partial distortions, i.e.

number

[0098] In this case, the corrected continuous partial distortion Δε k * By summing up the strain ε k (x,z) (corrected distortion) is obtained.

number

[0099] Corrective distortion ε calculated at any point with coordinates (x,z) in the XZ plane (image plane) k From the (x,z) pair, the corrected distortion image (IΔε k * ) can be constructed.

[0100] These corrected distortion images Iε k * can be seen as a slow-motion video.

[0101] Method for determining a corrected image - Patent application In summary, according to a first embodiment of the method according to the present disclosure and shown in FIG. 3, the method performed includes the following steps: (d1) Images of time-sequential media (I k ), where k is an image index between 0 and N, and the first image of the sequence is assumed to be undistorted; (d2) Determination of a motion for each image of the medium with index k between 1 and N, the motion being determined for the image (I k ) and the image of the preceding medium with index k-1 (I k-1 ) a decision that is between (d3) Image of the medium with index k (I k ) and the image of the preceding medium with index k-1 (I k-1 ) based on the time-sequential partial strain image (IΔε k ) decision, (d4) determining the partial distortion image (IΔε) based on the preceding motion, i.e., the motion relative to the image of the medium having an index between 1 and k, inclusive; k ) to obtain the corrected partial distortion image (IΔε k * ) decision, (d5) Corrected partial distortion image (IΔε k * ) to obtain the corrected distortion image (Iε k * ) Determination

[0102] Image summation or summation is understood to mean that the values ​​of pixels having the same spatial coordinates in said image are added (i.e. added together) to form the value of the corresponding pixel in the resulting image.

[0103] With these arrangements, the corrected distortion image Iε of the medium 2 that is not affected or only slightly affected by the deformations and movements induced by the external pressure P k * which makes it possible to obtain more accurate and more precise distortion values ​​than those of the prior art.

[0104] Elastic image of the medium IE k Decision US 2005 / 252295 proposes a technique for imaging the elasticity of a viscoelastic medium 2. An imaging device 1 investigates the propagation of elastic shear waves in this medium. The motion of the elastic shear waves is tracked by the means described above, in particular by a microcomputer 4.

[0105] This is done in several steps. (e1.1) an excitation step in which the microcomputer 4 generates elastic shear waves in the viscoelastic medium 2 by emitting, by the probe 6, at least one ultrasonic wave focused in the viscoelastic medium; (e1.2) an observation step in which the propagation of shear waves is observed simultaneously at multiple points in the field of view in the viscoelastic medium 2, said step comprising the step of obtaining a plurality of time-successive intermediate images II j,k The step includes the sub-steps of generating,j,where j is an intermediate image index between 0 and M (inclusive), and M+1 is the number of intermediate images to be created.

[0106] Each intermediate image of the medium may be, for example, an image I of the medium as described above. k The method for determining the intermediate image of the medium (II), in which at least one shot of unfocused ultrasonic compressional waves is emitted by the probe 6, received acoustic signals are detected and recorded by the probe 6, and these acoustic signals are used to generate an intermediate image of the medium (II j,k ) are processed to construct

[0107] The focus and timing of the focused ultrasound emitted in step (e1) and the timing of the unfocused ultrasound emitted in step (e2) are adapted so that the unfocused ultrasound reaches the field of view during the propagation of the shear wave in this field of view. Thus, the shear wave is reflected by the intermediate image II of the generated medium. j,k You can see it inside.

[0108] Then proceed to the next. (e1.3) Elastic image of the medium IE k Processing the intermediate image of the medium to determine

[0109] Intermediate Image II j,k is calculated by, for example, correlation or cross-correlation as above, in the displacement domain u j,k For example, the intermediate image (II j,k ) and the intermediate image of the medium with index j-1 (I j-1 , k) by cross-correlation between the displacement vectors. The determined displacement vectors can be used to visualize the propagation of the shear waves in the form of a video. Optionally, the partial strains Δε at each point of the medium 2 can be further calculated based on the displacement fields.

[0110] Based on the displacement area or the strain, the microcomputer 4 can then advantageously proceed to a mapping step during which at least one propagation parameter of the shear wave is calculated, either at a point in a selected field of view or over the entire field of view, based on the evolution of the motion parameter (displacement or partial strain) over time in the field of view.

[0111] Then, elasticity images I E corresponding to sets of shear wave propagation parameters at various points in the field of view are k can be constructed.

[0112] Elastic image of this medium IE k can be viewed in particular by means of a screen 4a of a microcomputer, where the values ​​of the propagation parameters are illustrated by optical parameters, such as a grey scale or a colour scale.

[0113] The shear wave propagation parameters calculated during the mapping step are, for example, selected from: · Shear wave speed Cs, or Shear modulus μ, or Young's modulus E = 3μ, or ·shear wave damping α, or Shear modulus μ1, or Shear viscosity μ2, or · Mechanical relaxation time τs of the tissue in the medium

[0114] For example, the following can be calculated at different points in the field of view: The shear wave velocity Cs that approximates the tissue stiffness Value of the mechanical relaxation time τs of the tissue, characteristic of the local viscosity of the medium

[0115] For this purpose, we use a shear wave propagation model, represented for example by the following propagation equation according to the displacement u generated by the shear wave at each position r in the medium:

number

number

[0116] By solving this propagation equation for the set of displacements u, it is possible to obtain the propagation parameters (cS, τS) mentioned above.

[0117] In calculating the propagation parameters, variants are possible: in particular, wave equations can be used in the Fourier domain, for example by averaging the values ​​over a frequency band. Partial strains can also be used instead of displacements.

[0118] Different shear waves can also be used to establish a map of the propagation parameters, i.e., elasticity images, which can then be combined to obtain a more accurate mapping, for example by averaging the elasticity images.

[0119] The propagation speed of the shear waves in the medium 2 is sufficiently fast (shear wave propagation parameter) taking into account that there are no variations in the external pressure P applied on the outer surface 3 during this process during which several images of the medium (intermediate images) are captured with the intention of determining an elasticity image IE. These images of the medium are thus not corrected for the motion of the medium.

[0120] On the other hand, the elastic image of the medium IE k is the comparison with the initial medium image I0 or the correction distortion image Iε k * To be able to compare with, we must correct for the motion (represented by the index k) at or near the time of interest.

[0121] Corrected Elasticity Image IE k * Decision Here, the elastic image I.E. k is corrected in the manner represented in FIG. 4, where the goal is to convert this elastic image I.E. into a situation according to the image I0 of the first medium in order to remove the motion caused by the distortion from the external pressure P. k The aim is to bring back.

[0122] This correction is performed for each elasticity image I.E. k This is done by using a set of general motions determined between each image of the medium for I0. More specifically, to return to the situation in the first image I0, the inverse motions must be applied.

[0123] In other words, the method includes the following steps: (e1) Elasticity image (IE k ) decision, (e2) Elastic images (IE k ) to obtain the corrected elasticity image (IE k * ) Determination

[0124] Image of nonlinear parameters INL kDecision In the linear elastic region, the medium 2 deforms proportionally to the stress σ and follows Hooke's law. σ=E0.ε Where: E0 is the Young's modulus in the linear region, ε is the strain.

[0125] In the non-linear elastic region, this proportionality is no longer valid. In many cases, the Young's modulus E of the material of medium 2 increases when compressed.

[0126] Here, the coefficient of elasticity, or Young's modulus E, is defined as the slope of the stress-strain curve, i.e.,

number

[0127] The nonlinear parameter of elasticity may for example be the Landau coefficient, called the third order elastic coefficient A, mentioned in the following document: · “Acoustoelasticity in soft solids: Assessment of the non-linear shear modulus with the acoustic radiation force”, Gennisson et al., J. Acoust. Soc. Am (122), December 2007, pp. 3211-3219 · “Quantitative Imaging of Nonlinear Shear Modulus by Combining Static Elastography and Shear Wave Elastography”, H. Latorre-Ossa et al., IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, Vol. 51, no. 4, pp. 833-839

[0128] In particular, equation (1) in the latter document can be rewritten in the form of a relationship (R1) between the elastic modulus (Young's modulus) E and the stress σ.

number

[0129] By differentiating the previous relation R1, we obtain:

number

[0130] Division and integration in the definition of Young's modulus from equation (Eq. 1) gives the relationship R2 between the elastic modulus (Young's modulus) E and strain ε.

number

[0131] According to a first variant NLa of the method represented in FIG. 5, the method executed by the microcomputer 4 on the basis of the strain image and the elasticity image further comprises the following steps: (NL1a) As shown above, multiple corrected distortion images (Iε k * ) decision, As shown above (NL2a), multiple corrected elasticity images (IE k * ), wherein the compensated strain image and the elasticity image are interlaced in time; (NL3a) Multiple corrected distortion images (Iε k * ) and multiple corrected elasticity images (IE k * ) based on the image of nonlinear parameters (INL k ) Determination

[0132] In effect, the nonlinear parameters of a pixel in an image are determined by the set of compensated distortion and elasticity values ​​for that pixel.

[0133] For example, the nonlinear parameters may be determined by a linear regression of the plurality of value pairs, which may be referred to as a skewed linear regression.

[0134] Each value pair is formed for a pixel of the image and includes: the first value of this same pixel from the rectified distortion image, and a second value of this same pixel from the corrected elasticity image, The corrected strain image and the corrected elasticity image of the object are temporally consecutive or nearby images, and may have the same or similar image indices.

[0135] Optionally, the strain and elasticity images can have different image indices, and as explained above, images should be taken with indices that correspond to successive or close in time instants, i.e., a correspondence is made between the time and index of each image of interest.

[0136] FIG. 6 shows a number of points Pa for the pixels of an image, where each point Pa is a pair of such values ​​(E k ,ε k ) All of these points Pa are approximately aligned along a straight line Da in a graph having on the abscissa the corrective distortion value ε (for this pixel) and on the ordinate the natural logarithm of the value of Young's modulus E (for this same pixel of the image).

[0137] The strain linear regression then uses the relationship R2 between the elastic modulus (Young's modulus) E and the strain ε (Equation 3).

number

[0138] Multiple corrected distortion images (Iε k * ) and corrected elasticity images (IE k *The solution of this skewed linear regression on the set of pairs of values ​​extracted from θ, θ, θ gives the value of the nonlinear parameter A for each pixel in the image.

[0139] These calculations allow us to calculate the nonlinearity parameter INL for medium 2 using the corrected image. k An image of the nonlinear parameter A is obtained, which does not have the artifacts from the uncorrected image of the prior art. The image is not affected, or only slightly affected, by the deformations and movements induced by the external pressure P. A more accurate and more correct nonlinear parameter value A is obtained than that of the prior art method.

[0140] According to a second variant NLb of the method represented in FIGS. 7A and 7B, the method executed by the microcomputer 4, based on the strain and elasticity images, comprises: A first phase in which the partial strain images are converted into a corrected stress image, and In the second phase, the image of the nonlinear parameters is determined.

[0141] During the first phase depicted in FIG. 7A, the following is performed: (c1) Corrected partial distortion image of index k (IΔε k * ) and the corrected elasticity image of index k (IE k * ) to obtain the time-sequential corrected partial stress image (IΔσ k * ) decision, (c2) Corrected partial stress image (IΔσ k * ) to obtain the corrected stress image (Iσ k * ) Determination

[0142] Multiplication of images is understood here to mean that the values ​​of pixels having the same spatial coordinates in said images are multiplied (i.e. multiplied together) to form the value of the corresponding pixel of the resulting image, whereby the corrected partial distortion value Δε calculated by the steps of the method described above and illustrated in FIG. * , the corrected elasticity value E calculated by the method steps described above and illustrated in FIG. * (corrected Young's modulus) and the corrected stress value Δσ for the pixel * get.

[0143] Thus, the partial stress is calculated at each pixel of the image by the following formula: Δσ * =E * .Δε *

[0144] Image summation or summation is understood to mean that the values ​​of pixels having the same spatial coordinates in said image are added (i.e. added together) to form the value of the corresponding pixel in the resulting image.

[0145] This means that

number

[0146] This formula is the corrected stress image (Iσk * ) at any pixel in the image.

[0147] During the second phase depicted in FIG. 7B, the following is performed: (NL1b) Multiple corrected stress images (Iσ k * ) decision, (NL2b) Multiple corrected elasticity images (IE k * ), wherein the compensated strain image and the elasticity image are interlaced in time; (NL3b) Multiple corrected stress images (Iε k* ) and multiple corrected elasticity images (IE k * ) based on the image of nonlinear parameters (INL k ) Determination

[0148] In effect, the nonlinear parameters of a pixel in an image are determined by the set of compensated stress and elasticity values ​​for that pixel.

[0149] For example, the nonlinear parameters are determined by a linear regression of the plurality of value pairs, which is referred to as a stress linear regression.

[0150] Each value pair is formed for a pixel of the image and includes: the first value of this same pixel of the corrected stress image, and a second value of this same pixel in the corrected elasticity image, The corrected stress image and the corrected elasticity image of the object are images that are successive in time or close in time, and may have the same or similar image indices.

[0151] Optionally, the stress and elasticity images can have different image indices, and as explained above, images should be taken with indices that correspond to successive or close instants in time, i.e., there is a correspondence between the time and index of each image of interest.

[0152] FIG. 8 shows a number of points Pb for pixels of an image, where each point Pb represents a pair of values ​​(E k ,σ k ) All of these points Pb are approximately aligned along a straight line Db in the graph having the corrected stress value σ (for this pixel) on the abscissa and the value of Young's modulus E (for this same pixel of the image) on the ordinate.

[0153] The strain linear regression then uses the relationship R1 between elasticity (Young's modulus) E and stress σ (Equation 2).

number

[0154] Multiple corrected stress images (Iσ k * ) and corrected elasticity images (IE k * The solution of this stress linear regression on the set of value pairs extracted from θ, θ, θ gives the value of the nonlinear parameter A for each pixel in the image.

[0155] These calculations allow us to calculate the nonlinearity parameter INL for medium 2 using the corrected image. k An image of the non-corrected image of the prior art is obtained. This image does not have the artifacts from the uncorrected image of the prior art. This image is not affected or only slightly affected by the deformation and movement induced by the external pressure P. A more accurate and more correct non-linear parameter value A is obtained than that of the prior art. [Explanation of symbols]

[0156] 1. Imaging device 2. Media, viscoelastic media 2i Inclusions 3 External surface 4. Microcomputer 4a Output Interface 4b Input Interface 5 Scattering particles, reflective particles 6 Ultrasound probe 7 Cabinet

Claims

1. 1. A method for ultrasonic determination of a corrected image of a medium, comprising: The images of the medium (I k ), where k is an image index between 0 and N, and the first image of the sequence is assumed to be undistorted; Determination of a motion for each image of the medium with index k between 1 and N, said motion being determined by the image of the medium with index k (I k ) and the preceding image of said medium with index k-1 (I k-1 ) between the decision and The image (I k ) and the preceding image of the medium with index k−1 (I k-1 ) based on the time-sequential partial strain image (IΔε k ) and The partial distortion image (IΔε) is based on a prior motion, which means a motion relative to the image of the medium with an index between 1 and k, inclusive. k ) to obtain the corrected partial distortion image (IΔε k * ) and The corrected partial distortion image (IΔε) of the indices between 1 and k, including 1 and k itself k * ) to obtain the corrected distortion image (Iε k * ) and The method includes:

2. said determining the motion of index k being The image of the medium (I k ) and the preceding image of the medium (I k-1 ) displacement area (u k ) and The displacement region (u k ) based on the geometric change of the image (T k ), wherein the geometric change is less than 10 parameters in the displacement region (u k ), represents the decision and The method of claim 1 , wherein the method is performed by

3. The displacement region is an image of the medium (I k ) and the preceding image of the medium (I k-1 ) or by cross-correlation between the images of said medium (I k ) and the preceding image of the medium (I k-1 3. The method of claim 2, wherein the sub-images are calculated by an algorithm for tracking the sub-images between the first and second images.

4. The method of claim 3 , wherein the tracking algorithm is the Lucas-Kanade algorithm.

5. The geometric change (T k 5. The method according to claim 2, wherein the transformation comprises at least one translation, or a translation and a similarity transformation, or a translation, a similarity transformation, and a rotation.

6. The geometric change (T k ) includes translation, similarity transformation, and rotation, and the geometric transformation is [0010] and put into matrix form such that Here, the geometric change in the image plane (T k ) the following parameters: Tx and Ty are translation coefficients, Hx and Hz are the coefficients of the similarity transformation. θ is the angle of the rotation axis perpendicular to the image plane The method according to any one of claims 2 to 4, wherein

7. The geometric change (T k ) are the parameters of the displacement domain (u k 3. The method of claim 2, wherein the parameter is obtained by a population median calculated based on a group of points in the

8. One group is the displacement region (u k ), including points between 3 and 10, The method of claim 7, wherein the population is greater than one hundred groups.

9. (e1) Elasticity image of the medium (IE k ) and (e2) the elasticity image (IE k ) to obtain the corrected elasticity image (IE k * ) and The method of claim 1, further comprising:

10. The elasticity image generated in step (e1) is (IE k ) is then processed in the following sub-steps: (e1.1) an excitation step in which shear waves are generated in the medium by emitting at least one focused ultrasonic wave; (e1.2) Intermediate images of the medium successive in time (II j,k an observation step in which the propagation of the shear wave is observed by determining (e1.3) the intermediate image of the medium (II j,k ) based on the shear wave propagation model, the elasticity image (IE k ) is determined; The method of claim 9, wherein the compound is produced by

11. Multiple corrected distortion images (Iε k * ) and Multiple corrected elasticity images (IE k * ), wherein the compensated strain image and the elasticity image are interlaced in time; The plurality of corrected distortion images (Iε k * ) and the plurality of corrected elasticity images (IE k *) based on the image of nonlinear parameters (INL k ) and The method of claim 9 further comprising:

12. The image of the nonlinear parameters (INL k ) a value for each pixel in the plurality of images is determined by strain linear regression of pairs of values, a first value of each pair corresponds to a value of the same pixel in a compensated strain image in the plurality and a second value of the pair corresponds to a value of the same pixel in an elasticity image in the plurality, the compensated strain images in the plurality and the compensated elasticity images in the plurality are consecutive in time or close in time; The strain linear regression is established based on the following relationship between strain and elasticity: [0025] Where: E is the Young's modulus of the pixel of the corrected elasticity image of the object; E 0 is the Young's modulus of the pixel of the first corrected elasticity image, ln() is the natural logarithm function, ε is the distortion of the pixel of the corrected distorted image of the object, The method of claim 11 , wherein A is the nonlinear parameter of the pixel determined by the warped linear regression.

13. The corrected partial distortion image (IΔε k * ) and the corrected elasticity image (IE k * ) to obtain the time-sequential corrected partial stress image (IΔσ k * ) and The corrected partial distortion image (IΔσ k * ) to obtain the corrected stress image (Iσ k * ) and The method of claim 9 further comprising:

14. Multiple corrected stress images (Iσ k * ) and Multiple corrected elasticity images (IE k * ), wherein the compensated strain image and the elasticity image are interlaced in time; The plurality of corrected stress images (Iε k * ) and the plurality of corrected elasticity images (IE k * ) based on the image of nonlinear parameters (INL k ) and The method of claim 13 further comprising:

15. The image of the nonlinear parameters (INL k ) a value for each pixel in the plurality of images is determined by stress linear regression of pairs of values, a first value of each pair corresponds to a value of the same pixel in a corrected stress image in the plurality and a second value of the pair corresponds to a value of the same pixel in an elasticity image in the plurality, the corrected stress images in the plurality and the corrected elasticity images in the plurality are consecutive in time or close in time; A stress linear regression is established based on the following relationship between stress and elasticity: [0030] Where: E is the Young's modulus of the pixel of the corrected elasticity image of the object; E 0 is the Young's modulus of the pixel of the first corrected elasticity image, ln() is the natural logarithm function, ε is the distortion of the pixel of the corrected distorted image of the object, The method of claim 14 , wherein A is a non-linear parameter determined for the pixel by the regression.

16. An imaging device (1) comprising an ultrasound probe (6) and a microcomputer (4) suitable for carrying out a method for ultrasound determination of a corrected image of a medium (2) according to any one of claims 1 to 15.

Citation Information

Patent Citations

  • Production of hot-extruded alloy material

    JP1983019428A

  • Imaging method and device employing sherar waves

    US20050252295A1

  • Method and Apparatus for Ultrasound Synthetic Imagining

    US20090234230A1

  • Ultrasonographic device

    WO2008075740A1