A photoacoustic non-line-of-sight imaging method

Abstract

The application discloses a photoacoustic non-line-of-sight imaging method, which utilizes the strong reflection of the inner wall of the skull to ultrasonic waves to receive the photoacoustic signals generated by the tissues (such as brain blood vessels) in the non-line-of-sight range. After detecting the photoacoustic signals in the non-line-of-sight range reflected by the skull, the photoacoustic image in the non-line-of-sight range can be recovered through corresponding signal and image processing algorithms. Compared with the prior art, the application has the following beneficial effects: the application utilizes the strong reflection effect of the inner wall of the skull to the sound wave signals to analyze the multiple reflection propagation processes of the sound waves in the skull from the limited imaging angle of the temporal bone region, thereby reconstructing the intracranial image outside the straight line angle and realizing the whole brain photoacoustic imaging.

Description

Technical Field

[0001] This invention relates to a photoacoustic imaging method, and more particularly to a transcranial photoacoustic imaging method for the human brain. Background Technology

[0002] Photoacoustic imaging is a rapidly developing hybrid imaging modality that combines the high specificity of multi-wavelength optical absorption with the deep penetration of ultrasound. Currently, photoacoustic imaging plays an important role in many biomedical applications, including whole-body imaging in mice, breast imaging, and skin imaging. In the past two years, transcranial photoacoustic imaging of the human brain has become a research hotspot. Compared to traditional techniques such as CTA and MRA, transcranial photoacoustic imaging can achieve high-resolution reconstruction of vascular distribution without contrast agents, and the overall system cost is lower.

[0003] However, the biggest challenge facing transcranial photoacoustic imaging is the severe attenuation and distortion of sound signals by the skull, posing a significant challenge to effective reception and image reconstruction. Therefore, selecting the thinnest region of the human skull, the temporal bone (only 0.1 cm at its thinnest point), for photoacoustic signal detection is one way to address this problem, minimizing the attenuation and distortion of sound waves by the skull. Nevertheless, because the temporal bone is located in a small area on the side of the brain, only localized photoacoustic signals can be detected from here, greatly limiting the application scenarios of photoacoustic brain imaging.

[0004] Figure 1 This is a schematic diagram illustrating photoacoustic signal detection from the temporal bone region. Figure 1 As can be seen, due to the limited area of ​​the temporal bone, the field of view of the ultrasound probe can only cover a part of the human brain (within the area of ​​the two dotted lines), and the photoacoustic signals generated by the brain tissue outside the dotted line area cannot be directly detected. Summary of the Invention

[0005] The technical problem that this invention aims to solve is that, since the temporal bone is located in a small area on the side of the human brain, only local photoacoustic signals of the human brain can be detected from here, which greatly limits the application scenarios of photoacoustic brain imaging.

[0006] To solve the above-mentioned technical problems, the technical solution of the present invention is to provide a photoacoustic non-line-of-sight imaging method, characterized by comprising the following steps:

[0007] Step 1: Perform ultrasound imaging of the human brain skull and model the inner surface of the skull based on the obtained images of the human brain skull.

[0008] Step 2: Based on the geometric features of the inner surface of the skull reflected by the model, combined with the arrangement of ultrasound probes at the temporal bone and the directionality of the ultrasound probes during photoacoustic imaging, select the corresponding inner surface of the skull that the ultrasound probes can receive as the reflection interface, and regard the points on the interface as virtual probes.

[0009] Step 3: Receive photoacoustic signals generated by tissues outside the visual range reflected from the skull;

[0010] Step 4: Perform time-domain truncation on the obtained photoacoustic signal to remove signal interference from multiple reflections;

[0011] Step 5: Combining the established virtual ultrasound probe, the photoacoustic signal received by the actual probe is projected onto the virtual probe according to the positional relationship between the virtual probe and each actual ultrasound probe. This obtains the weighted signal received by the virtual probe that is related to the spatial position, so that the reconstruction process changes from a process that includes reflection to the direct reception of the signal at the virtual probe position. The weighted signal is then reconstructed using a photoacoustic reconstruction algorithm.

[0012] Step 6: Post-process the reconstructed image.

[0013] Preferably, step 4 includes the following steps:

[0014] Remove the coupling signal portion from the obtained photoacoustic signal and truncate the photoacoustic signal to a fixed length in the time domain.

[0015] Preferably, when truncating a photoacoustic signal to a fixed length in the time domain, for signals with multiple reflections, the maximum path that the reflected signal may travel is calculated based on the known geometric structure of the inner surface of the skull, and the maximum length of the signal is truncated in the time domain based on this maximum path.

[0016] Preferably, after step 4 and before step 5, the method further includes the following step: filtering the photoacoustic signal obtained in the previous step to reduce interference.

[0017] Preferably, in step 6, for the multiple reflection signals that still exist, after the image reconstruction is completed, the portion of the reconstruction result located outside the skull is cut off according to the known skull geometry, so as to further remove image artifacts.

[0018] Preferably, in step 5, the weighted signal received by the virtual probe is represented as:

[0019]

[0020] In the formula: D n R represents the nth actual ultrasound probe. m Let m be the m-th reflection point on the reflecting interface. This represents the weighted signal received by the virtual probe corresponding to the m-th reflection point; F R () represents the reflection function of the interface reflecting the image; v RS v DR These represent the direction vector from the reflecting surface to a point in space and the actual direction vector from the ultrasonic probe to the reflecting surface, respectively. R represents the m-th reflection point. m The corresponding normal vector direction; P0() represents the spatial pressure distribution; (x, y) are the position coordinates of any point in space; t represents the current time, c DR c represents the actual distance from the ultrasonic probe to a point on the reflecting interface, and c0 represents the uniform sound velocity; c RS This represents the distance from the reflecting interface to any point in space.

[0021] This invention utilizes the strong reflection effect of the inner wall of the skull on sound wave signals from the limited imaging perspective of the temporal bone region to analyze the multiple reflection and propagation process of sound waves in the skull, thereby reconstructing intracranial images outside the straight-line viewpoint and realizing whole-brain photoacoustic imaging. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of photoacoustic imaging via the temporal bone.

[0023] Figure 2 This is a schematic diagram illustrating the principle of transtemporal photoacoustic non-visual distance imaging.

[0024] Figure 3 The framework of the intracranial photoacoustic non-line-of-sight imaging algorithm is illustrated.

[0025] Figure 4a A numerical simulation model of a cross-section of a blood vessel as the target being measured;

[0026] Figure 4b A numerical simulation model for a short blood vessel as the target being tested;

[0027] Figure 5a and Figure 5b The diagram illustrates the detected photoacoustic signals within the non-line-of-sight region, where, Figure 5a The signal received by the first sensor unit. Figure 5b This is the signal received by the 28th sensor;

[0028] Figure 6a and Figure 6b This is a diagram showing the results of non-line-of-sight photoacoustic imaging, in which... Figure 6a This is a reconstructed image of a cross-section of a blood vessel. Figure 6b This is a reconstruction diagram of a short blood vessel. Detailed Implementation

[0029] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims.

[0030] In performing photoacoustic transcranial imaging, this invention does not consider the attenuation and distortion of photoacoustic signals by the skull as interfering factors. Instead, it uses a reverse approach to fully utilize the strong reflection of photoacoustic signals by the skull to enhance the ability of the ultrasound probe at the temporal bone to detect photoacoustic signals in non-visual regions, thereby better completing the reconstruction of whole-brain photoacoustic images. Figure 2 As shown, a portion of the photoacoustic signal generated by the blood vessel being tested propagates downwards to the brain tissue. When it encounters the inner wall of the skull, it undergoes a strong reflection and is then detected by the ultrasound probe placed on the temporal bone.

[0031] This invention discloses a photoacoustic non-line-of-sight imaging method that utilizes the strong reflection of ultrasound waves by the inner wall of the skull to receive photoacoustic signals generated by tissues (such as cerebral blood vessels) outside the line of sight. Once the photoacoustic signal reflected from the skull outside the line of sight is detected, the photoacoustic image within the non-line-of-sight range can be recovered through appropriate signal and image processing algorithms.

[0032] Because the inner wall of the skull is not an ideal plane, and the propagation direction of photoacoustic signals is also anisotropic and uncertain, the photoacoustic signals will be reflected multiple times on different parts of the skull at different times. Therefore, traditional image reconstruction based directly on the received reflected signals often fails to achieve good imaging results.

[0033] Furthermore, the present invention proposes as follows: Figure 3 The algorithm framework shown is used to achieve high-quality intracranial photoacoustic non-line-of-sight imaging. Unlike traditional in-line-of-sight imaging, non-line-of-sight imaging cannot directly reconstruct the target image using traditional back projection algorithms. Here, this invention proposes an algorithm that uses ultrasound skull structure imaging as prior information, combined with photoacoustic signal preprocessing and image postprocessing, to achieve high-quality photoacoustic non-line-of-sight imaging.

[0034] Step 1: First, perform ultrasound imaging of the human skull, segment the obtained skull image, and then model the inner surface of the skull.

[0035] Step 2: Based on the geometric features of the inner surface of the skull, combined with the arrangement of ultrasound probes at the temporal bone during photoacoustic imaging, and the directionality and size of the probes, the corresponding inner surface of the skull that the probes can receive is selected as the reflection interface. The points on the interface are regarded as virtual probes. According to the relationship between the virtual probes and each actual probe, the signals received by the actual probes are projected onto the virtual probes. This changes the reconstruction process from a process involving reflection to a direct reception of signals at the virtual probe positions, preparing for the next step of photoacoustic non-line-of-sight back projection imaging.

[0036] Step 3: Photoacoustic signal acquisition: Receive photoacoustic signals generated by tissues outside the line of sight and reflected from the skull.

[0037] Step 4: First, remove the coupled signal portion from the signal and truncate it to a fixed length in the time domain. For signals with multiple reflections, calculate the maximum possible path traveled by the reflected signal based on the known geometry of the inner surface of the skull, and then truncate the signal to its maximum length in the time domain based on this path.

[0038] Step 5: Filter the photoacoustic signal obtained in the previous step to reduce interference such as white noise.

[0039] Step 6: Based on the established ultrasonic virtual probe geometric model, the classic back projection imaging algorithm is improved. After projecting the photoacoustic signal of the actual probe onto the virtual probe, the weighted signal related to the spatial position received by the virtual probe is obtained. The reconstruction process of this signal is consistent with the traditional photoacoustic reconstruction process. Thus, the classic algorithm can be used to reconstruct the weighted signal, making the classic algorithm applicable to the reconstruction process of photoacoustic non-line-of-sight imaging.

[0040] In a discrete setup, define N probes D on a plane. n and its location With normal vector direction M reflection points R on the reflection boundary m and its location and the corresponding normal vector direction Then probe D n Received signal for

[0041]

[0042] Where: c DR c RS These represent the distance from the probe to a point on the reflecting interface and the distance from a point on the reflecting interface to any point (x, y) in space, respectively.

[0043] c0 represents the uniform speed of sound;

[0044] v RS v DRThese represent the direction vector from the reflecting surface to a point in space and the direction vector from the probe to the reflecting surface, respectively.

[0045] F R F represents the reflection function of the reflective interface. D Let represent the probe's directionality function, and we have:

[0046]

[0047]

[0048] Among them, v in v out Both N and k are unit vectors. R k D These are the surface unfolding index of the reflective interface and the probe directivity coefficient, respectively.

[0049] P0 represents the spatial pressure distribution.

[0050] For each pair of probes and reflection points, establish a virtual probe. The received signal can be calculated using the following formula:

[0051]

[0052] Since the signal from the virtual probe can be represented as

[0053]

[0054] in This problem can be viewed as a weighted photoacoustic reconstruction problem, thus enabling classical photoacoustic reconstruction algorithms to be used for photoacoustic non-line-of-sight imaging.

[0055] Step 7: Post-process the reconstructed image, including but not limited to image noise reduction, smoothing filtering, and artifact removal. Additionally, for any remaining multiple reflection signals, after image reconstruction, we truncate the portion of the reconstructed image located outside the skull based on the known skull geometry, further removing image artifacts.

[0056] It is worth noting that the imaging method for modeling the geometric structure of the skull in this invention is not limited to ultrasound imaging. Other medical imaging modalities such as CT and MRI can also be used for skull modeling. Different skull modeling methods can be selected according to different scenarios.

[0057] Furthermore, the method described above for establishing a geometric model of an intracranial ultrasound virtual probe (step two) can optimize and improve photoacoustic signal preprocessing, image reconstruction, and image postprocessing methods for non-line-of-sight imaging. Additionally, other photoacoustic image reconstruction methods, such as convex optimization methods and deep learning methods, can also be used in photoacoustic non-line-of-sight imaging based on the improved and optimized ultrasound virtual probe geometric model (step two). All signal / image processing methods used in intracranial photoacoustic non-line-of-sight imaging are within the scope of this invention.

[0058] The pseudocode for the above-mentioned intracranial photoacoustic non-visual imaging method is shown in Table 1 below.

[0059] Table 1. Pseudocode for Non-Line-of-Sight Imaging Reconstruction

[0060]

[0061] Numerical simulation experiments were conducted on the above-mentioned intracranial photoacoustic non-line-of-sight imaging method in a computer, including the following steps.

[0062] 1) Establish a photoacoustic numerical model of the human brain, including brain structure segmentation, assignment of optical and acoustic parameters, etc.

[0063] 2) The process of photons passing through the skull and their distribution within the skull are predicted by Monte Carlo simulation to obtain the intracranial light intensity distribution.

[0064] 3) Based on the preset optical parameters of brain tissue (such as light absorption coefficient), obtain the distribution of the initial sound pressure in intracranial tissue.

[0065] 4) Import the initial sound pressure data into the k-wave photoacoustic simulation software, perform photoacoustic simulation, and receive the photoacoustic signal through an array ultrasound probe placed at the temporal bone.

[0066] 5) After performing the above signal preprocessing and image reconstruction on the photoacoustic signal, a photoacoustic non-line-of-sight image is obtained.

[0067] Human brain numerical simulation Figure 4a and Figure 4b As shown, it includes the skull, blood vessel targets, and linear ultrasound probes. Figure 5a and Figure 5b The image shows a photoacoustic signal obtained through simulation. As can be seen, the ultrasound probe located in the temporal bone successfully received the photoacoustic signal from the non-visual region due to the strong reflection from the inner surface of the skull. Using this non-visual photoacoustic signal, combined with the geometry of the inner surface of the skull, a photoacoustic image from the non-visual region can be reconstructed, such as... Figure 6a and Figure 6b As shown. From Figure 6a and Figure 6bAs can be seen in the images shown, non-visual photoacoustic imaging accurately locates the blood vessels. Although artifacts are present, this is sufficient to demonstrate the feasibility of intracranial non-visual photoacoustic imaging.

Claims

1. A photoacoustic non-line-of-sight imaging method, characterized in that, Includes the following steps: Step 1: Perform ultrasound imaging of the human brain skull and model the inner surface of the skull based on the obtained images of the human brain skull. Step 2: Based on the geometric features of the inner surface of the skull reflected by the model, combined with the arrangement of ultrasound probes at the temporal bone and the directionality of the ultrasound probes during photoacoustic imaging, select the corresponding inner surface of the skull that the ultrasound probes can receive as the reflection interface, and regard the points on the interface as virtual probes. Step 3: Receive photoacoustic signals generated by tissues outside the visual range reflected from the skull; Step 4: Perform time-domain truncation on the obtained photoacoustic signal to remove signal interference from multiple reflections; Step 5: Combining the established virtual ultrasound probe, and based on the positional relationship between the virtual probe and each actual ultrasound probe, the photoacoustic signals received by the actual probes are projected onto the virtual probe to obtain the spatially related weighted signal received by the virtual probe. This transforms the reconstruction process from one involving reflection to one of direct signal reception at the virtual probe's location. The weighted signal is then reconstructed using a photoacoustic reconstruction algorithm. The weighted signal received by the virtual probe is represented as follows: In the formula: D n R represents the nth actual ultrasound probe. m Let m be the m-th reflection point on the reflecting interface. This represents the weighted signal received by the virtual probe corresponding to the m-th reflection point; F R () represents the reflection function of the interface reflecting the image; v RS v DR These represent the direction vector from the reflecting surface to a point in space and the actual direction vector from the ultrasonic probe to the reflecting surface, respectively. R represents the m-th reflection point. m The corresponding normal vector direction; P0() represents the spatial pressure distribution; (x,y) is the position coordinate of any point in space; t represents the current time, c DR c represents the actual distance from the ultrasonic probe to a point on the reflecting interface, and c0 represents the uniform sound velocity; c RS This represents the distance from a point on the reflecting interface to any point (x, y) in space. Step 6: Post-process the reconstructed image.

2. The photoacoustic non-line-of-sight imaging method as described in claim 1, characterized in that, Step 4 includes the following steps: Remove the coupling signal portion from the obtained photoacoustic signal and truncate the photoacoustic signal to a fixed length in the time domain.

3. The photoacoustic non-line-of-sight imaging method as described in claim 2, characterized in that, When performing a fixed-length truncation of a photoacoustic signal in the time domain, for signals with multiple reflections, the maximum possible path traveled by the reflected signal is calculated based on the known geometric structure of the inner surface of the skull, and the maximum length of the signal is truncated in the time domain based on this maximum path.

4. The photoacoustic non-line-of-sight imaging method as described in claim 1, characterized in that, The process includes the following steps after step 4 and before step 5: filtering the photoacoustic signal obtained in the previous step to reduce interference.

5. The photoacoustic non-line-of-sight imaging method as described in claim 1, characterized in that, In step 6, for the multiple reflection signals that still exist, after the image reconstruction is completed, the part of the reconstruction result located outside the skull is cut off according to the known skull geometry, so as to further remove image artifacts.

Images