WO2025102491A1 - 一种基于光声显微成像的图像增强方法 - Google Patents

一种基于光声显微成像的图像增强方法 Download PDF

Info

Publication number
WO2025102491A1
WO2025102491A1 PCT/CN2023/141312 CN2023141312W WO2025102491A1 WO 2025102491 A1 WO2025102491 A1 WO 2025102491A1 CN 2023141312 W CN2023141312 W CN 2023141312W WO 2025102491 A1 WO2025102491 A1 WO 2025102491A1
Authority
WO
WIPO (PCT)
Prior art keywords
signals
pixel
similarity
range
target
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
PCT/CN2023/141312
Other languages
English (en)
French (fr)
Inventor
刘成波
雷天成
徐智强
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Institute of Advanced Technology of CAS
Original Assignee
Shenzhen Institute of Advanced Technology of CAS
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shenzhen Institute of Advanced Technology of CAS filed Critical Shenzhen Institute of Advanced Technology of CAS
Publication of WO2025102491A1 publication Critical patent/WO2025102491A1/zh
Anticipated expiration legal-status Critical
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10056—Microscopic image
    • G06T2207/10061—Microscopic image from scanning electron microscope
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10132—Ultrasound image
    • G06T2207/10136—3D ultrasound image

Definitions

  • the present invention relates to the technical field of photoacoustic imaging, and more specifically, to an image enhancement method based on photoacoustic microscopy imaging.
  • photoacoustic imaging technology as a non-invasive imaging method, stands out among many medical imaging technologies because it combines the high contrast of optical imaging with the advantages of high resolution and penetration depth of ultrasound imaging.
  • Photoacoustic imaging technology breaks through the limit of traditional optical imaging ( ⁇ 1mm) and extends the penetration depth to the centimeter level, becoming one of the fastest-growing technologies in medical imaging.
  • Current imaging systems usually perform maximum projection on the three-dimensional images obtained by scanning to obtain projection images, and perform algorithm processing based on the projection images. The signal-to-noise ratio of the images obtained in this way is often limited.
  • Deep learning requires a large amount of image data for preparing training sets, validation sets, and test sets, which results in a lot of time spent on data collection.
  • the training process is computationally intensive and has high computational time costs.
  • PA photoacoustic
  • PAM photoacoustic microscopy
  • the purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide an image enhancement method based on photoacoustic microscopy.
  • the method comprises the following steps:
  • the advantages of the present invention are that the provided image enhancement method based on photoacoustic microscopy imaging fully utilizes the characteristics of photoacoustic microscopy imaging three-dimensional data, utilizes the high correlation of each voxel in time and space, and uses an improved non-local means (NLM) algorithm to process images, which significantly improves the image quality and the signal-to-noise ratio of the imaging system.
  • NLM non-local means
  • the present invention does not require a large amount of data training, and can fully utilize the spatial and temporal dimension characteristics of the collected data information.
  • FIG1 is a flow chart of an image enhancement method based on photoacoustic microscopy imaging according to an embodiment of the present invention
  • FIG. 2 is a diagram of an image enhancement method based on photoacoustic microscopy according to an embodiment of the present invention. Process diagram;
  • FIG3 is a flow chart of processing original three-dimensional data in the depth dimension according to one embodiment of the present invention.
  • FIG4 is a flow chart of processing a maximum projection graph in a time axis dimension according to an embodiment of the present invention
  • FIG. 5 is a schematic diagram of experimental results of an image enhancement method based on photoacoustic microscopy according to an embodiment of the present invention.
  • the present invention provides a continuous frame four-dimensional image processing algorithm based on NLM technology (hereinafter referred to as seq-NLM), which is used for image denoising and enhancement to improve the image signal-to-noise ratio.
  • NLM non-local mean
  • the present invention makes full use of the characteristics of photoacoustic microscopy data as a three-dimensional structure, and combines the continuity of the data on the time axis to improve the performance of the traditional NLM algorithm.
  • the image enhancement method comprises the following steps:
  • Step S110 using photoacoustic microscopy to obtain three-dimensional data information of the target, and obtaining the scanning result of the target on the time axis through continuous scanning imaging.
  • photoacoustic microscopy can obtain three-dimensional information of the observed object or target, and through continuous scanning imaging, it can obtain scanning results related to the time axis. Since the distribution of blood vessels in the spatial structure must be continuous, the blood vessels obtained based on this scanning system have a high degree of spatiotemporal continuity.
  • the three-dimensional information of the spatial dimension combined with the continuity of the time dimension can be considered as four-dimensional data processing for continuous frames.
  • Step S120 for the three-dimensional data information, merge the signals of adjacent areas in the spatial dimension.
  • the output result is optimized by combining the tomographic results of adjacent depths.
  • an algorithm for pixel-by-pixel processing is used. For example, in order to restore the imaging quality, before performing the maximum projection operation on the result of a single imaging, the NLM algorithm is first used on the original three-dimensional data in the spatial dimension to synthesize the relevant signals of adjacent areas within a certain search range.
  • v(j,h) represents the original pixel with noise
  • ⁇ (i,j) represents the weight relationship between two pixels
  • I represents the plane range that the algorithm can search in the original data
  • H represents the depth range that the algorithm can search in the original data
  • i represents the pixel index in the target depth plane
  • j represents the pixel index in a certain depth plane
  • h represents the depth index.
  • the weight ⁇ is introduced before signal synthesis to suppress the interference of irrelevant items.
  • the weight ⁇ (i,j) is set as:
  • Z(i) represents the normalization coefficient
  • g represents the smoothing coefficient
  • V(i) represents the neighborhood similarity block of pixel i. Only when the neighborhood similarity is high can it be said that the similarity between the two pixels is high.
  • the similarity between signals can be measured by ⁇ V(i)-V(j,h) ⁇ 2 , thereby effectively evaluating the correlation degree of adjacent region signals, which is expressed as:
  • d represents the radius of the similarity block.
  • the present invention first uses the continuity of vascular data to search and weighted average the data in a three-dimensional range, making full use of the three-dimensional data obtained by scanning, effectively expanding the selection range of similar blocks in the algorithm, and effectively reducing the noise intensity in the original data.
  • Step S130 with respect to the scanning result on the time axis, signals of adjacent regions are merged on the time dimension.
  • a non-local mean algorithm can be used to merge signals in the time dimension.
  • a single frame image can be obtained, and a continuous dynamic video can be obtained by scanning. Due to the high-speed characteristics of the imaging system, it can be considered that there is still continuity between different frames.
  • the maximum projection map that is continuously scanned and processed in step S120 is stacked in time order, and the NLM algorithm is used in the time dimension to synthesize the relevant signals of adjacent areas within a certain range, which is expressed as:
  • v(j,t) represents the original pixel with noise
  • ⁇ (i,j) represents the weight relationship between two pixels
  • I represents the plane range that the algorithm can search in the original data
  • T represents the time range that the algorithm can search in the original data
  • i represents the pixel index at the target time
  • j represents the pixel index at a certain time
  • t represents the time index.
  • the weight ⁇ is introduced before signal synthesis to suppress the interference of irrelevant items.
  • Z(i) represents the normalization coefficient
  • g represents the smoothing coefficient
  • V(i) represents the similarity block range.
  • d represents the radius of the similarity block.
  • weight ⁇ involved in steps S120 and S130 may also be in other forms.
  • it can be set as a fixed value based on experience or experiments, and the similarity between signals can also be measured by other criteria, such as structural similarity.
  • the existing NLM algorithm is usually used for denoising and enhancement of a single image.
  • extending the range of similar blocks to the time axis range can effectively expand the selection range of similar blocks in the algorithm and make it easier to find similar pixels for weighted synthesis.
  • FIG3 is a flowchart for processing raw three-dimensional data in the depth dimension, where I(m,n,h) is the raw three-dimensional data obtained by photoacoustic microscopy scanning, m and n are the image sizes, and the detection depth is h. Ds is the search block range, ds is the similarity block size, and Hs is the depth range.
  • the three-dimensional data I[i-Ds:i+Ds,j-Ds:j+Ds,h-Hs:h+Hs] is the search range Sz of the NLM algorithm
  • the two-dimensional data I[i-ds:i+ds,j-ds:j+ds,h] is the size of the reference block R
  • the two-dimensional data I[p-ds:p+ds,q-ds:q+ds,z] is the size of the similarity block S.
  • Similar blocks are searched in sequence within the search range, and the weight A is obtained according to the similarity between the similar blocks and the reference blocks, and then weighted summation is performed to achieve the purpose of updating the target pixel point I(i,j).
  • FIG4 is a flowchart for processing the maximum projection image in the time axis dimension, where I(m,n,t) is the maximum projection image to be processed at time t, and m and n are the image sizes.
  • Ds is the search block range
  • ds is the similarity block size
  • Ts is the time axis range.
  • the three-dimensional data I[i-Ds:i+Ds,j-Ds:j+Ds,t-Ts:t+Ts] is the search range Sz of the NLM algorithm
  • the two-dimensional data I[i-ds:i+ds,j-ds:j+ds,t] is the size of the reference block R
  • the two-dimensional data I[p-ds:p+ds,q-ds:q+ds, ⁇ ] is the size of the similarity block S.
  • Similar blocks are searched in sequence within the search range, and the weight A is obtained according to the similarity between the similar blocks and the reference blocks, and then weighted summation is performed to achieve the purpose of updating the target pixel point I(i,j).
  • the present invention has the following advantages:
  • the present invention proposes a Seq-NLM algorithm, which makes full use of the data information of three-dimensional voxels for image processing of photoacoustic microscopy, and effectively improves the image signal-to-noise ratio of the photoacoustic microscopy system.
  • the present invention utilizes this data feature and uses the seq-NLM algorithm to obtain images with excellent signal-to-noise ratio.
  • the algorithm of the present invention does not use complex filtering algorithms and can still obtain excellent results. Less time cost.
  • the algorithm in the present invention does not rely on a large amount of data sets, but only requires processing of the original photoacoustic data, which reduces the difficulty of data acquisition and the computational burden.
  • the seq-NLM algorithm proposed in the present invention is easy to implement and has high computational efficiency, and can be easily extended to other PAM systems based on devices such as scanning galvanometers.
  • the present invention may be a system, a method and/or a computer program product.
  • the computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present invention.
  • Computer readable storage medium can be a tangible device that can hold and store instructions used by an instruction execution device.
  • Computer readable storage medium can be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination thereof.
  • Non-exhaustive list of computer readable storage medium include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disk read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, for example, a punch card or a convex structure in a groove on which instructions are stored, and any suitable combination thereof.
  • RAM random access memory
  • ROM read-only memory
  • EPROM or flash memory erasable programmable read-only memory
  • SRAM static random access memory
  • CD-ROM compact disk read-only memory
  • DVD digital versatile disk
  • memory stick a floppy disk
  • mechanical encoding device for example, a punch card or a convex structure in a groove on which instructions are stored, and any suitable combination thereof.
  • the computer readable storage medium used here is not interpreted as a transient signal itself, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagated by a waveguide or other transmission medium (for example, a light pulse by an optical fiber cable), or an electrical signal transmitted by a wire.
  • the computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing/processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and/or a wireless network.
  • the network can include copper transmission cables, optical fiber transmissions, wireless transmissions, routers, firewalls, switches, gateway computers, and/or edge servers.
  • the network adapter card or network interface in each computing/processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing/processing device.
  • the computer program instructions for performing the operations of the present invention may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state setting instructions, or other instructions.
  • the computer readable program instructions may be executed completely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer, partially on a remote computer, or completely on a remote computer or server.
  • the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect through the Internet).
  • LAN local area network
  • WAN wide area network
  • an electronic circuit such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), may be personalized by utilizing the state information of the computer readable program instructions.
  • the electronic circuit may execute the computer readable program instructions, thereby realizing various aspects of the present invention.
  • These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processor of the computer or other programmable data processing device, a device that implements the functions/actions specified in one or more boxes in the flowchart and/or block diagram is generated.
  • These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause the computer, programmable data processing device, and/or other equipment to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured product, which includes instructions for implementing various aspects of the functions/actions specified in one or more boxes in the flowchart and/or block diagram.
  • Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operating steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions/actions specified in one or more boxes in the flowchart and/or block diagram.
  • each box in the flow chart or block diagram can represent a part of a module, a program segment or an instruction, and a part of the module, a program segment or an instruction contains one or more executable instructions for realizing the specified logical function.
  • the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved.
  • each box in the block diagram and/or flow chart, and the combination of the boxes in the block diagram and/or flow chart can be implemented by a dedicated hardware-based system that performs the specified function or action, or can be implemented by a combination of dedicated hardware and computer instructions. It is well known to those skilled in the art that it is equivalent to implement it by hardware, implement it by software, and implement it by combining software and hardware.

Landscapes

  • Image Processing (AREA)

Abstract

本发明公开了一种基于光声显微成像的图像增强方法。该方法包括：针对目标物体，执行多次光声显微扫描，获得目标的三维数据，并通过连续扫描成像获得目标在时间轴上的扫描结果；在单次成像结果进行投影操作之前，在空间维度上对所述三维数据在设定的搜索范围内合成相邻区域的信号；针对所述时间轴上的扫描结果，在设定的搜索范围内，在时间维度上合并相邻区域的信号，获得增强图像。本发明有效提高了光声显微系统的图像信噪比，并且提高了计算效率。

Description

一种基于光声显微成像的图像增强方法 技术领域
本发明涉及光声成像技术领域，更具体地，涉及一种基于光声显微成像的图像增强方法。
背景技术
随着医学影像技术的飞速发展，光声成像技术作为一项非侵入式成像方法因其兼具光学成像高对比度，以及超声成像高分辨率和穿透深度的优势，在众多医疗成像技术中脱颖而出。光声成像技术突破了传统光学成像极限(～1mm)，将穿透深度扩展到厘米级，成为目前医学影像技术中发展最为迅速的技术之一。目前的成像系统通常对扫描获得的三维图像进行最大值投影获得投影图，在投影图的基础上进行算法处理，这种方式获得的图像的信噪比往往有限。
在现有技术中，常用的图像增强方法有两类：一类是基于三维块匹配算法(Block Matching 3D，BM3D)对原始数据进行去噪，利用图像的自相似性进行频域变换去噪，提高图像的信噪比。另一类方法是深度学习，将视野内的低噪声图像作为真实数据集，利用卷积神经网络来增强图像的信噪比。
经分析，现有图像增强方案主要存在以下缺陷：
(1)构成算法的协同滤波与维纳滤波运算耗时长。当噪声强度较大时，降噪图像往往会丢失边缘、纹理等细节信息，造成图像模糊等问题
(2)深度学习需要大量的图像数据用于制备训练集、验证集和测试集，导致消耗大量的时间进行数据采集，而且训练过程计算量大，计算时间成本高。
(3)已有方案都是对扫描后获得的单帧最大值投影图进行处理，损失了空间信息。
(4)对于光声(Photoacoustic，PA)成像技术，特别是光声显微成像技术(Photoacoustic Microscopy，PAM)，现有的图像增强方法仍面临挑战。例如，PAM可以进行亚微米级别的成像以进行血管层面的分析，但由于背景噪声的影响，该技术观察毛细血管等微小血管网络仍存在困难。
发明内容
本发明的目的是克服上述现有技术的缺陷，提供一种基于光声显微成像的图像增强方法。该方法包括以下步骤：
针对目标物体，执行多次光声显微扫描，获得目标的三维数据，并通过连续扫描成像获得目标在时间轴上的扫描结果；
在单次成像结果进行投影操作之前，在空间维度上对所述三维数据在设定的搜索范围内合成相邻区域的信号；
针对所述时间轴上的扫描结果，在设定的搜索范围内，在时间维度上合并相邻区域的信号，获得增强图像。
与现有技术相比，本发明的优点在于，所提供的基于光声显微成像的图像增强方法，充分利用光声显微成像三维数据的特征，利用各个体素在时空间上的高度相关性，使用改进的非局部均值(Non-Local Means，NLM)算法处理图像，显著提高了图像质量和成像系统的信噪比。此外，与深度学习方法相比，本发明无需使用大量的数据训练，并且能充分利用采集到的数据信息的空间维度和时间维度特征。
通过以下参照附图对本发明的示例性实施例的详细描述，本发明的其它特征及其优点将会变得清楚。
附图说明
被结合在说明书中并构成说明书的一部分的附图示出了本发明的实施例，并且连同其说明一起用于解释本发明的原理。
图1是根据本发明一个实施例的基于光声显微成像的图像增强方法的流程图；
图2是根据本发明一个实施例的基于光声显微成像的图像增强方法的 过程示意图；
图3是根据本发明一个实施例的在深度维度上处理原始三维数据的流程图；
图4是根据本发明一个实施例的在时间轴维度上处理最大值投影图的流程图；
图5是根据本发明一个实施例的基于光声显微成像的图像增强方法的实验结果示意图。
具体实施方式
现在将参照附图来详细描述本发明的各种示例性实施例。应注意到：除非另外具体说明，否则在这些实施例中阐述的部件和步骤的相对布置、数字表达式和数值不限制本发明的范围。
以下对至少一个示例性实施例的描述实际上仅仅是说明性的，决不作为对本发明及其应用或使用的任何限制。
对于相关领域普通技术人员已知的技术、方法和设备可能不作详细讨论，但在适当情况下，所述技术、方法和设备应当被视为说明书的一部分。
在这里示出和讨论的所有例子中，任何具体值应被解释为仅仅是示例性的，而不是作为限制。因此，示例性实施例的其它例子可以具有不同的值。
应注意到：相似的标号和字母在下面的附图中表示类似项，因此，一旦某一项在一个附图中被定义，则在随后的附图中不需要对其进行进一步讨论。
本发明提供了一种基于NLM技术的连续帧四维图像处理算法(以下简称seq-NLM)，用于图像去噪增强，以提高图像信噪比。在本发明中，结合PAM获得的数据特点，分别在空间上的深度方向与时间轴方向使用NLM(非局部均值)算法，扩大邻域选择的范围，获得更好的相似块选取结果。本发明充分利用了光声显微数据是三维结构的特点，并且结合数据在时间轴上的连续性，提升了传统NLM算法的性能。
具体地，参见结合图1和图2所示，所提供的基于光声显微成像的图 像增强方法包括以下步骤：
步骤S110，利用光声显微扫描，获得目标的三维数据信息，并通过连续扫描成像获得目标在时间轴上的扫描结果。
例如，光声显微扫描能够获得观察对象或目标的三维信息，并通过连续扫描成像可以获得时间轴上相关联的扫描结果。由于血管在空间结构上的分布一定是连续的，基于这种扫描系统获得的血管具有高度的时空连续性。空间维度的三维信息结合时间维度的连续性，可以认为是针对连续帧的四维数据处理。
步骤S120，针对三维数据信息，在空间维度上合并相邻区域的信号。
利用光声显微成像三维数据具有连续性的特征，逐层对三维数据进行去噪处理时，进而结合相邻深度的层析结果优化输出结果。在一个实施例中，采用逐像素点进行处理的算法。例如，为了恢复成像质量，在单次成像的结果进行最大值投影操作之前，首先在空间维度上对原始三维数据使用NLM算法，在一定的搜索范围内将相邻区域的有关信号进行合成，
其中，表示去噪后的像素点结果，v(j,h)表示含噪声的原始像素点，ω(i,j)表示两个像素点间的权重关系，I表示原始数据中算法可搜索的平面范围，H表示原始数据中算法可搜索的深度范围，i表示目标深度平面内的像素点索引，j表示某一深度平面内的像素点索引，h表示深度索引。
为了合理地利用相邻信号，在信号合成前引入权重ω来抑制无关项的干扰。例如，权重ω(i,j)设置为：

其中，Z(i)表示归一化系数，g表示平滑系数，V(i)表示像素i的邻域相似块。只有邻域相似度高才能说明这两个像素的相似度高，可使用‖V(i)-V(j,h)‖2度量信号间的相似度，从而有效地评估相邻区域信号的相关程度，表示为：
其中，d表示相似块半径。
区别于传统最大值投影图的获取，本发明首先利用血管数据的连续性，在三维范围内对数据进行搜索与加权平均，充分利用了扫描得到的三维数据，有效扩充了算法中相似块的选取范围，可以有效减小原始数据中的噪声强度。
步骤S130，针对时间轴上的扫描结果，在时间维度上合并相邻区域的信号。
例如，可采用非局部均值算法在时间维度上合并信号。具体地，每一轮扫描获得的三维数据进行最大值投影操作后，可以获得单帧图像，通过扫描可以获得连续的动态视频。由于成像系统的高速特性，可以认为不同帧之间仍然存有连续性，为了提高最大值投影图的成像质量，将连续扫描且经过步骤S120处理的最大值投影图按照时间顺序堆叠，在时间维度上使用NLM算法，在一定范围内将相邻区域的有关信号进行合成，表示为：
其中，表示去噪后的像素点结果，v(j,t)表示含噪声的原始像素点，ω(i,j)表示两个像素点间的权重关系，I表示原始数据中算法可搜索的平面范围，T表示原始数据中算法可搜索的时间范围，i表示目标时刻像素点索引，j表示某一时刻像素点索引，t表示时间索引。
为了合理地利用相邻信号，在信号合成前引入权重ω来抑制无关项的干扰，

式中，Z(i)表示归一化系数，g表示平滑系数，V(i)表示相似块范围。使用‖V(i)-V(j,t)‖2度量信号间的相似度，可以有效地评估相邻区域信号的相关程度，表示为：
其中，d表示相似块半径。
需要说明的是，步骤S120和S130中涉及到权重ω也可采用其他形式， 例如，根据经验或实验设置为固定值，并且信号间的相似度也可采用其他衡量标准，如结构相似性等。
现有的NLM算法通常被使用于单张图像的去噪增强，在处理光声显微图像时，将相似块的范围扩展到时间轴范围，可以有效扩充算法中相似块的选取范围，更容易找到具有相似性的像素点进行加权合成。
图3是在深度维度上处理原始三维数据的流程图，其中I(m,n,h)为光声显微扫描获得的原始三维数据，m和n为图像大小，探测深度为h。Ds为搜索块范围，ds为相似块大小，Hs为深度范围。三维数据I[i-Ds:i+Ds,j-Ds:j+Ds,h-Hs:h+Hs]是NLM算法的搜索范围Sz，二维数据I[i-ds:i+ds,j-ds:j+ds,h]是参考块R的大小，二维数据I[p-ds:p+ds,q-ds:q+ds,z]是相似块S的大小。在搜索范围内依次寻找相似块，根据相似块与参考块之间的相似程度求得权重A后进行加权求和，实现更新目标像素点I(i,j)的目的。
图4是在时间轴维度上处理最大值投影图的流程图，其中I(m,n,t)为t时刻待处理的最大值投影图，m和n为图像大小。Ds为搜索块范围，ds为相似块大小，Ts为时间轴范围。三维数据I[i-Ds:i+Ds,j-Ds:j+Ds,t-Ts:t+Ts]是NLM算法的搜索范围Sz，二维数据I[i-ds:i+ds,j-ds:j+ds,t]是参考块R的大小，二维数据I[p-ds:p+ds,q-ds:q+ds,τ]是相似块S的大小。在搜索范围内依次寻找相似块，根据相似块与参考块之间的相似程度求得权重A后进行加权求和，实现更新目标像素点I(i,j)的目的。
为进一步验证本发明的效果，进行了实验模拟。例如，对裸鼠耳朵血管、裸鼠大脑皮层血管进行光声显微成像实验。参见图5所示，实验结果表明，获得的图像有去噪增强效果，有助于提高PAM系统的信噪比。
综上所述，相对于现有技术，本发明具有以下优势：
(1)本发明提出了Seq-NLM算法，充分利用了三维体素的数据信息可以用于光声显微成像的图像处理，有效提高了光声显微系统的图像信噪比。本发明利用这种数据特点，使用seq-NLM算法，可以获得信噪比极佳的图像。
(2)本发明中算法未使用复杂的滤波算法仍能获得优秀的结果，减 少了时间成本。
(3)本发明中算法不依赖于大量的数据集，只需要原始光声数据处理，降低了数据获取难度和计算负担。
(4)本发明提出的seq-NLM算法易于实现，计算效率高，可以很容易地扩展到其他基于扫描振镜等器件的PAM系统。
本发明可以是系统、方法和/或计算机程序产品。计算机程序产品可以包括计算机可读存储介质，其上载有用于使处理器实现本发明的各个方面的计算机可读程序指令。
计算机可读存储介质可以是可以保持和存储由指令执行设备使用的指令的有形设备。计算机可读存储介质例如可以是但不限于电存储设备、磁存储设备、光存储设备、电磁存储设备、半导体存储设备或者上述的任意合适的组合。计算机可读存储介质的更具体的例子(非穷举的列表)包括：便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、静态随机存取存储器(SRAM)、便携式压缩盘只读存储器(CD-ROM)、数字多功能盘(DVD)、记忆棒、软盘、机械编码设备、例如其上存储有指令的打孔卡或凹槽内凸起结构、以及上述的任意合适的组合。这里所使用的计算机可读存储介质不被解释为瞬时信号本身，诸如无线电波或者其他自由传播的电磁波、通过波导或其他传输媒介传播的电磁波(例如，通过光纤电缆的光脉冲)、或者通过电线传输的电信号。
这里所描述的计算机可读程序指令可以从计算机可读存储介质下载到各个计算/处理设备，或者通过网络、例如因特网、局域网、广域网和/或无线网下载到外部计算机或外部存储设备。网络可以包括铜传输电缆、光纤传输、无线传输、路由器、防火墙、交换机、网关计算机和/或边缘服务器。每个计算/处理设备中的网络适配卡或者网络接口从网络接收计算机可读程序指令，并转发该计算机可读程序指令，以供存储在各个计算/处理设备中的计算机可读存储介质中。
用于执行本发明操作的计算机程序指令可以是汇编指令、指令集架构(ISA)指令、机器指令、机器相关指令、微代码、固件指令、状态设置数 据、或者以一种或多种编程语言的任意组合编写的源代码或目标代码，所述编程语言包括面向对象的编程语言—诸如Smalltalk、C++、Python等，以及常规的过程式编程语言—诸如“C”语言或类似的编程语言。计算机可读程序指令可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络—包括局域网(LAN)或广域网(WAN)—连接到用户计算机，或者，可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。在一些实施例中，通过利用计算机可读程序指令的状态信息来个性化定制电子电路，例如可编程逻辑电路、现场可编程门阵列(FPGA)或可编程逻辑阵列(PLA)，该电子电路可以执行计算机可读程序指令，从而实现本发明的各个方面。
这里参照根据本发明实施例的方法、装置(系统)和计算机程序产品的流程图和/或框图描述了本发明的各个方面。应当理解，流程图和/或框图的每个方框以及流程图和/或框图中各方框的组合，都可以由计算机可读程序指令实现。
这些计算机可读程序指令可以提供给通用计算机、专用计算机或其它可编程数据处理装置的处理器，从而生产出一种机器，使得这些指令在通过计算机或其它可编程数据处理装置的处理器执行时，产生了实现流程图和/或框图中的一个或多个方框中规定的功能/动作的装置。也可以把这些计算机可读程序指令存储在计算机可读存储介质中，这些指令使得计算机、可编程数据处理装置和/或其他设备以特定方式工作，从而，存储有指令的计算机可读介质则包括一个制造品，其包括实现流程图和/或框图中的一个或多个方框中规定的功能/动作的各个方面的指令。
也可以把计算机可读程序指令加载到计算机、其它可编程数据处理装置、或其它设备上，使得在计算机、其它可编程数据处理装置或其它设备上执行一系列操作步骤，以产生计算机实现的过程，从而使得在计算机、其它可编程数据处理装置、或其它设备上执行的指令实现流程图和/或框图中的一个或多个方框中规定的功能/动作。
附图中的流程图和框图显示了根据本发明的多个实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段或指令的一部分，所述模块、程序段或指令的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个连续的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或动作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。对于本领域技术人员来说公知的是，通过硬件方式实现、通过软件方式实现以及通过软件和硬件结合的方式实现都是等价的。
以上已经描述了本发明的各实施例，上述说明是示例性的，并非穷尽性的，并且也不限于所披露的各实施例。在不偏离所说明的各实施例的范围和精神的情况下，对于本技术领域的普通技术人员来说许多修改和变更都是显而易见的。本文中所用术语的选择，旨在最好地解释各实施例的原理、实际应用或对市场中的技术改进，或者使本技术领域的其它普通技术人员能理解本文披露的各实施例。本发明的范围由所附权利要求来限定。

Claims (10)

  1. 一种基于光声显微成像的图像增强方法，包括以下步骤：
    针对目标物体，执行多次光声显微扫描，获得目标的三维数据，并通过连续扫描成像获得目标在时间轴上的扫描结果；
    在单次成像结果进行投影操作之前，在空间维度上对所述三维数据在设定的搜索范围内合成相邻区域的信号；
    针对所述时间轴上的扫描结果，在设定的搜索范围内，在时间维度上合并相邻区域的信号，获得增强图像。
  2. 根据权利要求1所述的方法，其特征在于，在空间维度上对所述三维数据在设定的搜索范围内合成相邻区域的信号采用以下公式：
    其中，表示合成后的像素点结果，v(j,h)表示合成之前的原始像素点，ω(i,j)表示两个像素点间的权重，I表示设定的搜索平面范围，H表示搜索的深度范围，i表示目标深度平面内的像素点索引，j表示某一深度平面内的像素点索引，h表示深度索引。
  3. 根据权利要求2所述的方法，其特征在于，所述两个像素点间的权重间的权重ω(i,j)设置为：

    其中，Z(i)表示归一化系数，g表示平滑系数，V(i)表示像素i的邻域相似块，‖V(i)-V(j,h)‖2是信号间的相似度。
  4. 根据权利要求3所述的方法，其特征在于，所述信号间的相似度‖V(i)-V(j,h)‖2设置为：
    其中，d表示相似块半径，k是像素点索引。
  5. 根据权利要求1所述的方法，其特征在于，针对所述时间轴上的扫描结果，在设定的搜索范围内，在时间维度上合并相邻区域的信号采用以下公式：
    其中，表示合并后的像素点结果，v(j,t)表示合并之前的原始像素点，ω(i,j)表示两个像素点间的权重，I表示搜索的平面范围，T表示搜索的时间范围，i表示目标时刻像素点索引，j表示某一时刻像素点索引，t表示时间索引。
  6. 根据权利要求5所述的方法，其特征在于，所述两个像素点间的权重设置为：

    其中，Z(i)表示归一化系数，g表示平滑系数，V(i)表示相似块范围。‖V(i)-V(j,t)‖2是信号间的相似度。
  7. 根据权利要求6所述的方法，其特征在于，所述信号间的相似度‖V(i)-V(j,t)‖2设置为：
    其中，d表示相似块半径，k是像素点索引。
  8. 根据权利要求1所述的方法，其特征在于，所述目标物体是血管。
  9. 一种计算机可读存储介质，其上存储有计算机程序，其中，该计算机程序被处理器执行时实现根据权利要求1至8中任一项所述的方法的步骤。
  10. 一种计算机设备，包括存储器和处理器，在所述存储器上存储有能够在处理器上运行的计算机程序，其特征在于，所述处理器执行所述计算机程序时实现权利要求1至8中任一项所述的方法的步骤。
PCT/CN2023/141312 2023-11-16 2023-12-23 一种基于光声显微成像的图像增强方法 Pending WO2025102491A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202311533121.8A CN117541499A (zh) 2023-11-16 2023-11-16 一种基于光声显微成像的图像增强方法
CN202311533121.8 2023-11-16

Publications (1)

Publication Number Publication Date
WO2025102491A1 true WO2025102491A1 (zh) 2025-05-22

Family

ID=89795268

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/141312 Pending WO2025102491A1 (zh) 2023-11-16 2023-12-23 一种基于光声显微成像的图像增强方法

Country Status (2)

Country Link
CN (1) CN117541499A (zh)
WO (1) WO2025102491A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN120259130A (zh) * 2025-06-03 2025-07-04 陕西省人民医院(陕西省临床医学研究院) 基于医疗影像的卵巢早衰患者电针疗效检测方法及系统

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130136328A1 (en) * 2011-11-30 2013-05-30 General Electric Company Methods and systems for enhanced tomographic imaging
CN110610147A (zh) * 2019-08-30 2019-12-24 中国科学院深圳先进技术研究院 血管图像提取方法、相关装置及存储设备
CN115299967A (zh) * 2022-08-10 2022-11-08 深圳市安健科技股份有限公司 基于x光设备的三维成像方法、设备和存储介质
CN115456950A (zh) * 2022-08-15 2022-12-09 上海交通大学医学院附属第九人民医院 一种穿支血管自动识别和定位方法、设备及存储介质
CN116681589A (zh) * 2023-06-07 2023-09-01 常州博恩中鼎医疗科技有限公司 一种cbct大视野数据拼接方法和装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130136328A1 (en) * 2011-11-30 2013-05-30 General Electric Company Methods and systems for enhanced tomographic imaging
CN110610147A (zh) * 2019-08-30 2019-12-24 中国科学院深圳先进技术研究院 血管图像提取方法、相关装置及存储设备
CN115299967A (zh) * 2022-08-10 2022-11-08 深圳市安健科技股份有限公司 基于x光设备的三维成像方法、设备和存储介质
CN115456950A (zh) * 2022-08-15 2022-12-09 上海交通大学医学院附属第九人民医院 一种穿支血管自动识别和定位方法、设备及存储介质
CN116681589A (zh) * 2023-06-07 2023-09-01 常州博恩中鼎医疗科技有限公司 一种cbct大视野数据拼接方法和装置

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN120259130A (zh) * 2025-06-03 2025-07-04 陕西省人民医院(陕西省临床医学研究院) 基于医疗影像的卵巢早衰患者电针疗效检测方法及系统

Also Published As

Publication number Publication date
CN117541499A (zh) 2024-02-09

Similar Documents

Publication Publication Date Title
Varghese et al. Video denoising based on a spatiotemporal Gaussian scale mixture model
CN108921800B (zh) 基于形状自适应搜索窗口的非局部均值去噪方法
CN108198154B (zh) 图像去噪方法、装置、设备及存储介质
El-Shafai et al. Traditional and deep-learning-based denoising methods for medical images
JP2001057677A (ja) 画像処理方法および装置並びに記録媒体
CN114022444B (zh) 超声图像的处理方法及装置
CN112053394B (zh) 图像处理方法、装置、电子设备及存储介质
CN108305268B (zh) 一种图像分割方法及装置
JP2014199664A (ja) グレースケール画像の処理方法及び装置
Tofighi et al. Denoising images corrupted by impulsive noise using projections onto the epigraph set of the total variation function (PES-TV)
Chen et al. Speckle attenuation by adaptive singular value shrinking with generalized likelihood matching in optical coherence tomography
Shao et al. Optimization of ultrasound information imaging algorithm in cardiovascular disease based on image enhancement
CN111630569B (zh) 双目匹配的方法、视觉成像装置及具有存储功能的装置
Zhou et al. Synchronizing detection and removal of smoke in endoscopic images with cyclic consistency adversarial nets
KR102195047B1 (ko) 3d이미지 품질을 향상시키는 방법과 장치
WO2025102491A1 (zh) 一种基于光声显微成像的图像增强方法
CN110163812B (zh) 肌肉超声图像降噪方法、装置及存储介质
Chen et al. Pathological image super-resolution using mix-attention generative adversarial network
Alam et al. Space-variant blur kernel estimation and image deblurring through kernel clustering
Liu et al. One-dimensional image surface blur algorithm based on wavelet transform and bilateral filtering
CN120388095B (zh) 一种mri图像重建方法、装置、电子装置和存储介质
Chaudhry et al. Underwater visibility restoration using dehazing, contrast enhancement and filtering
Raj et al. SpFusionNet: deep learning-driven brain image fusion with spatial frequency analysis
Direk Image denoising by linear regression on non-local means algorithm
CN113920023B (zh) 图像处理方法及装置、计算机可读介质和电子设备

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23958759

Country of ref document: EP

Kind code of ref document: A1