WO2020224508A1 - 一种对实时影像中图像的物体轮廓进行加强的方法 - Google Patents

一种对实时影像中图像的物体轮廓进行加强的方法 Download PDF

Info

Publication number
WO2020224508A1
WO2020224508A1 PCT/CN2020/087871 CN2020087871W WO2020224508A1 WO 2020224508 A1 WO2020224508 A1 WO 2020224508A1 CN 2020087871 W CN2020087871 W CN 2020087871W WO 2020224508 A1 WO2020224508 A1 WO 2020224508A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
brightness
gain value
value
real
Prior art date
Application number
PCT/CN2020/087871
Other languages
English (en)
French (fr)
Inventor
莫文军
苟春梅
Original Assignee
湖南省华芯医疗器械有限公司
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 湖南省华芯医疗器械有限公司 filed Critical 湖南省华芯医疗器械有限公司
Priority to EP20802689.8A priority Critical patent/EP3965053A4/en
Priority to US17/608,991 priority patent/US20230401675A1/en
Priority to JP2021565940A priority patent/JP7076168B1/ja
Publication of WO2020224508A1 publication Critical patent/WO2020224508A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/70—Denoising; Smoothing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/73—Deblurring; Sharpening
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/90—Dynamic range modification of images or parts thereof
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10016—Video; Image sequence
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images
    • G06T2207/10081—Computed x-ray tomography [CT]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images
    • G06T2207/10088—Magnetic resonance imaging [MRI]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images
    • G06T2207/10104—Positron emission tomography [PET]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10116—X-ray image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20021—Dividing image into blocks, subimages or windows
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20024—Filtering details
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20172—Image enhancement details
    • G06T2207/20192—Edge enhancement; Edge preservation

Definitions

  • the invention belongs to the technical field of real-time video image processing, and specifically relates to a method for enhancing the contour of an object in an image, which is used for enhancing the contour in the image.
  • Medical Imaging is the study of interacting with the human body by means of a certain medium (such as X-rays, electromagnetic fields, ultrasonic waves, etc.), to display the structure and density of the internal tissues and organs of the human body in an image format for the physician to observe, and to view the human body.
  • the visualization information for research includes the relatively independent research directions of medical imaging systems and medical image processing.
  • CT ordinary CT, spiral CT
  • PET positron scan
  • ultrasound B-ultrasound, color Doppler ultrasound, cardiac color ultrasound, three-dimensional color ultrasound
  • MRI magnetic resonance imaging
  • electrocardiogram equipment EEG equipment, etc.
  • the images (images) displayed by X-ray imagers, CT, ultrasound, and MRI are all black and white images, and due to optical reasons, blood vessels cannot be distinguished in the images, and only some special places can be seen. From a certain perspective, the prior art only performs pure contour extraction on the image, ignoring the details of the image.
  • the images displayed by the equipment are all original images without contour enhancement and extraction, which increases the difficulty of distinguishing and identifying objects with relatively low recognition.
  • the purpose of the present invention is to provide a method for enhancing the contours of objects in real-time images, so as to solve the problem that the images output by the equipment in the prior art are original images, which are relatively low-recognition objects. , Improve the difficulty of distinguishing and identifying.
  • a method for enhancing the contours of objects in real-time images includes the following steps:
  • the multiple image blocks with enhanced contours are filtered as a whole to obtain a noise-removed image
  • the next frame of image in the image is divided into multiple image blocks, and the next frame of image is divided into blocks based on any of 8 ⁇ 8, 16 ⁇ 16, or 32 ⁇ 32 to obtain the corresponding Of multiple image blocks.
  • step S2 are:
  • the improved gradient calculation formula is:
  • Y1, Y2 are the brightness of two adjacent pixels, i, j are the abscissa and ordinate of the pixel, and I is the calculated gradient value.
  • step S3 are:
  • the multiple image blocks after the enhanced contour are classified as a whole;
  • the multiple image blocks after the enhanced contour are weighted and averaged, the brightness value of the entire image is calculated, and the entire image after the enhanced contour is graded according to the brightness of the brightness value; that is, according to the gain value
  • the classification includes four levels: one level for gain value less than ISO100, one level for gain value ISO200, one level for gain value ISO400, and one level for gain value ISO800.
  • step 3.2 when the gain value is less than ISO100, no filtering is performed; when the gain value is ISO200, the standard deviation is set to 16, and filtering is performed based on the Gaussian filter formula; when the gain value is ISO400, the standard deviation is set to 20 , Filter processing based on Gaussian filter formula; when gain value is ISO800, set standard deviation to 25, filter processing based on Gaussian filter formula;
  • the Gaussian filtering formula is:
  • ⁇ represents the standard deviation
  • Y represents the brightness of the current pixel
  • i and j are the abscissa and ordinate of the pixel, respectively.
  • Y represents the brightness of the current pixel
  • ⁇ represents the ⁇ coefficient
  • i and j are the abscissa and ordinate of the pixel, respectively;
  • the present invention combines the gradient-based algorithm, the Gaussian filtering method and the gamma transform to strengthen the contour of the object in the image, which can maximize the recognizability of the object in the image and facilitate the easy identification of micro-objects;
  • Each pixel in the image in the present invention is a discrete function, and the gradient of the existing gradient algorithm is the partial derivative of a continuous function on the surface. Therefore, the modification of the existing gradient algorithm in this case is convenient for better Contour enhancement of multiple image blocks;
  • hierarchical filtering is performed according to the brightness and darkness of the entire image, which solves the problem of poor image clarity and smoothness when filtering processing directly based on the same brightness and darkness through Gaussian filtering in the prior art;
  • an enhanced image of the overall observation image and an enhanced image of the details of the observation image can be obtained, which is suitable for viewing in different situations.
  • Figure 1 is a schematic flow diagram of the present invention.
  • first and second are only used for description purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Thus, the features defined with “first” and “second” may explicitly or implicitly include one or more of these features.
  • “plurality” means two or more than two, unless specifically defined otherwise.
  • the terms “installed”, “connected”, “connected”, “fixed” and other terms should be understood in a broad sense, for example, it can be a fixed connection or a detachable connection. , Or integrally connected; it can be a mechanical connection or an electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, and it can be the internal communication between two components.
  • installed can be a fixed connection or a detachable connection.
  • it can be a mechanical connection or an electrical connection
  • it can be directly connected, or indirectly connected through an intermediate medium, and it can be the internal communication between two components.
  • the specific meaning of the above-mentioned terms in the present invention can be understood according to specific circumstances.
  • a method for enhancing the contours of objects in an image includes the following steps:
  • S1 acquires the captured image, performs block processing on the next frame of the image to obtain multiple image blocks; that is, divides the next frame of image into multiple image blocks, and the next frame of image is based on 8 ⁇ 8, 16 ⁇ Blocking is performed in either of 16 or 32 ⁇ 32 methods to obtain multiple corresponding image blocks.
  • 16 ⁇ Blocking is performed in either of 16 or 32 ⁇ 32 methods to obtain multiple corresponding image blocks.
  • other blocking methods can also be used, but these methods are selected for blocking, which makes the processing speed faster;
  • S2 is based on the gradient algorithm to perform contour enhancement on each pixel in each image block; the specific steps are:
  • the gradient in the gradient algorithm in the prior art is a partial derivative of a continuous function on a curved surface, and each pixel of the image in the present invention is a discrete function, it is necessary to use specific pixels to calculate the gradient.
  • the gradient algorithm in the prior art needs to be improved, and the improved gradient calculation formula is:
  • Y1, Y2 are the brightness of two adjacent pixels, i, j are the abscissa and ordinate of the pixel, and I is the calculated gradient value.
  • the multiple image blocks after the enhanced contour are filtered as a whole to obtain an image after the noise is filtered, the purpose is to increase the smoothness of the image;
  • Denoising is due to sensor sampling, gain and other factors will inevitably introduce noise, or own electronic noise. According to the manifestation of noise, there are salt and pepper noise, white noise, Gaussian noise, etc., which will reduce the clarity and smoothness of the image. Therefore, it is necessary to filter the image, but if the filter is directly based on the same brightness and darkness, the clarity and smoothness of the image will also be reduced.
  • the multiple image blocks with enhanced contours are filtered as a whole to obtain a noise-filtered image.
  • the specific steps are as follows:
  • the gain value (brightness value) is graded, including four levels: a gain value of less than ISO100 is a level, a gain value of ISO200 is a level, a gain value of ISO400 is a level, and a gain value is ISO800 is a level.
  • the gain values of the multiple image blocks (that is, the entire image) after the grading meet the bright and dark requirements, and the entire image that meets the requirements is filtered based on the Gaussian filtering formula. That is, when the gain value is less than ISO100, no filtering is performed; when the gain value is ISO200, the standard deviation is set to 16, and the filtering is performed based on the Gaussian filter formula; when the gain value is ISO400, the standard deviation is set to 20, and the filtering is performed based on the Gaussian filter formula Processing; when the gain value is ISO800, set the standard deviation to 25, and perform filtering based on the Gaussian filter formula;
  • the Gaussian filtering formula is:
  • ⁇ represents the standard deviation
  • Y represents the brightness of the current pixel
  • i and j are the abscissa and ordinate of the pixel, respectively.
  • Y represents the brightness of the current pixel
  • ⁇ represents the ⁇ coefficient
  • i and j are the abscissa and ordinate of the pixel, respectively;
  • the enhanced image effect will be different. It is suitable for different environmental applications.
  • a larger gamma parameter a larger brightness range can be seen, which is suitable for observing the overall image; use a smaller
  • the gamma parameter focuses on the details in the limited brightness range and is suitable for observing image details. Therefore, based on the gamma transformation formula, set a global ⁇ parameter value for the observation image and a ⁇ parameter value for the observation image details, using two ⁇ respectively
  • the parameter value performs light-dark contrast processing on the entire image after filtering the noise, and obtains the final enhanced image T(Y).
  • next frame of image that is, based on 32 ⁇ 32, divide the next frame of image (1920 ⁇ 1080 pixels) into If there are multiple 60 ⁇ 33.75 pixel image blocks, if there is a decimal point, the image blocks are divided by rounding, that is, 60 ⁇ 34 pixels are used as an image block, and the next frame of image is divided, specifically: Based on 32 ⁇ 31 and 32 ⁇ 26, the next frame of image 1920 ⁇ 1080 pixels are divided into 1920 ⁇ 1054 and 1920 ⁇ 26 for block processing, and the final image block is obtained after processing.
  • the multiple image blocks after the enhanced outline are classified as a whole; that is, the entire image after the enhanced outline is classified according to the gain value, including four levels: the gain value is less than ISO100 is one level, gain value is ISO200 for one level, gain value is ISO400 for one level, and gain value is ISO800 for one level.
  • the gain value of the entire image after grading meets the requirements of brightness and darkness, based on Gaussian filtering formula Perform filtering processing on the entire image that meets the requirements, otherwise, no filtering processing is performed. That is, when the gain value is less than ISO100, no filtering is performed; when the gain value is ISO200, the standard deviation is set to 16, and the filtering is performed based on the Gaussian filter formula; when the gain value is ISO400, the standard deviation is set to 20, and the filtering is performed based on the Gaussian filter formula Processing; when the gain value is ISO800, set the standard deviation to 25, and perform filtering based on the Gaussian filter formula;
  • the present invention is not limited to the above-mentioned alternative embodiments.
  • anyone can derive other products in various forms under the enlightenment of the present invention, but regardless of any changes in its shape or structure, all that fall into the scope of the claims of the present invention The technical solutions within the scope fall within the protection scope of the present invention.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Image Processing (AREA)
  • Apparatus For Radiation Diagnosis (AREA)
  • Picture Signal Circuits (AREA)

Abstract

本发明属于图像处理技术领域，本发明公开了一种对实时影像中图像的物体轮廓进行加强的方法，解决现有技术中设备输出的图像是原始图像，对于辨识度比较低的物体，提高了分辨、识别的难度的问题。本发明获取拍摄的影像，对影像中的下一帧图像进行分块处理，得到多个图像块；基于梯度算法，对每个图像块中的每个像素进行轮廓增强；基于高斯滤波方法，对增强轮廓后的多个图像块作为一个整体进行噪声的滤除，得到滤除噪声后的图像；基于gamma变换，对滤除噪声后的整个图像进行亮暗对比度处理，得到最终加强后的图像。本发明用于对图像中的轮廓进行加强。

Description

一种对实时影像中图像的物体轮廓进行加强的方法 技术领域
本发明属于实时视频图像处理技术领域，具体涉及一种对图像中的物体轮廓进行加强的方法，用于对图像中的轮廓进行加强。
背景技术
医学图像学Medical Imaging，是研究借助于某种介质(如X射线、电磁场、超声波等)与人体相互作用，把人体内部组织器官结构、密度以图像方式表现出来，供医师进行观看，从而对人体进行研究的可视化信息，包括医学成像系统和医学图像处理两方面相对独立的研究方向。
主要包括X光成像仪器、CT(普通CT、螺旋CT)、正子扫描(PET)、超声(分B超、彩色多普勒超声、心脏彩超、三维彩超)、核磁共振成像(MRI)、心电图仪器、脑电图仪器等。
X光成像仪、CT、超声、核磁共振成像等显示的影像(图像)均为黑白影像，且由于光学原因，图像中不能分辨血管组织，只能看到一些特殊的地方。从某个角度讲，现有技术对图像只是进行纯粹的轮廓提取，忽略了影像细节。
综上所述，现有技术中的医疗器械产品中，设备所显示出来的影像都是原始图像，没有经过轮廓的加强和提取，对于辨识度比较低的物体，提高了分辨、识别的难度。
发明内容
为了解决现有技术存在的上述问题，本发明目的在于提供一种对实时影像中图像的物体轮廓进行加强的方法，解决现有技术中设备输出的影像是原始图像，对于辨识度比较低的物体，提高了分辨、识别的难度的问题。
本发明所采用的技术方案为：
一种对实时影像中图像的物体轮廓进行加强的方法，包括如下步骤：
S1、获取拍摄的影像，对影像中的下一帧图像进行分块处理，得到多个图像块；
S2、基于梯度算法，对每个图像块中的每个像素进行轮廓增强；
S3、基于高斯滤波方法，对增强轮廓后的多个图像块作为一个整体进行噪声的滤除，得到滤除噪声后的图像；
S4、基于gamma变换，对滤除噪声后的整个图像进行亮暗对比度处理，得到最终加强后的图像。
进一步，所述步骤S1中，将影像中的下一帧图像分成多个图像块，下一帧图像基于8×8、16×16或32×32中的任一种方式进行分块，得到对应的多个图像块。
进一步，所述步骤S2的具体步骤为：
S2.1、基于改进后的梯度计算公式对每个图像块的每一个像素的亮度分量进行梯度计算，得到每个像素的梯度值；
S2.2、判断每个梯度值是否大于设置的阈值，若大于，则判定为边缘，并根据梯度值的下负判定是设置在白边还是黑边；若小于，则不做任何处理。
进一步，所述步骤S2.1中，改进后的梯度计算公式为：
I＝∑ i∑ j((Y1-Y2)/(Y1+Y2))，
其中，Y1，Y2为相邻两个像素的亮度，i、j分别为像素的横坐标和纵坐标，I为计算的梯度值。
进一步，所步骤S3的具体步骤为：
S3.1、根据增强轮廓后的多个图像块的亮暗程度，将增强轮廓后的多个图像块作为一个整体进行分级；
S3.2、分级后的整个图像的gain值达到亮暗要求，基于高斯滤波公式对达到要求的整个图像进行滤波处理，否则，不作滤波处理。
进一步，所述步骤3.1中，将增强轮廓后的多个图像块进行加权平均，计算整个图像的亮度值，根据其亮度值的亮暗程度将增强轮廓后的整个图像进行分级；即根据gain值进行分级，包括四个级别：gain值小于ISO100为一个级别、gain值为ISO200为一个级别、gain值为ISO400为一个级别、gain值为ISO800为一个级别。
进一步，所述步骤3.2中，gain值小于ISO100时，不做滤波处理；gain值为ISO200时，设置标准差为16，基于高斯滤波公式进行滤波处理；gain值为ISO400时，设置标准差为20，基于高斯滤波公式进行滤波处理；gain值为ISO800时，设置标准差为25，基于高斯滤波公式进行滤波处理；
其中，高斯滤波公式为：
Figure PCTCN2020087871-appb-000001
其中，σ表示标准差，Y表示当前像素的亮度，i、j分别为像素的横坐标和纵坐标。
进一步，所述步骤4中，gamma变换的公式为：
T(Y)＝Σ iΣ j(1+Y) γ，
其中，Y表示当前像素的亮度，γ表示γ系数，i、j分别为像素的横坐标和纵坐标；
基于gamma变换的公式，设定一个观察图像全局的γ参数值和一个观察图像细节的γ参数值，分别利用两个γ参数值对滤除噪声后的整个图像进行亮暗对比度处理，得到最终的加强后图像T(Y)。
本发明的有益效果为：
一、本发明结合基于梯度算法、高斯滤波方法和基于gamma变换，对图像中的物体轮廓进行加强，能最大限度的提高图像中物体的可辨识度，便于容易辨识微物体；
二、本发明中的图像中的每个像素是离散函数，而现有的梯度算法的梯度在曲面上是连续函数的偏导数，所以本案对现有的梯度算法进行的修改，便于更好的对多个图像块进行轮廓增强；
三、本发明中，根据整个图像的亮暗程度进行分级滤波，解决了现有技术中直接基于同一亮暗程度通过高斯滤波进行滤波处理，会造成图像清晰度和平滑度差的问题；
四、本发明中，通过设置两个γ参数值，可得到观察图像全局的加强图像和观察图像细节的加强图像，适用于不同情况下的观看。
附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1是本发明流程示意图。
具体实施方式
下面详细描述本发明的实施例，所述实施例的示例在附图中示出，其中自始至终相同或类似的标号表示相同或类似的元件或具有相同或类似功能的元件。下面通过参考附图描述的实施例是示例性的，旨在用于解释本发明，而不能理解为对本发明的限制。
在本发明的描述中，需要理解的是，术语“第一”、“第二”仅用于描述目的，而不能理解为指示或暗示相对重要性或者隐含指明所指示的技术特征的数量。由此，限定有“第一”、“第二”的特征可以明示或者隐含地包括一个或者更多个该特征。在本发明的描述中，“多个”的含义是两个或两个以上，除非另有明确具体的限定。
在本发明中，除非另有明确的规定和限定，术语“安装”、“相连”、“连接”、“固定”等术语应做广义理解，例如，可以是固定连接，也可以是可拆卸连接，或一体地连接；可以是机械连接，也可以是电连接；可以是直接相连，也可以通过中间媒介间接相连，可以是两个元件内部的连通。对于本领域的普通技术人员而言，可以根据具体情况理解上述术语在本发明中的具体含义。
本发明的描述中，有的结构或器件或步骤未作具体描述的，理解为现有技术中有能实现的结构或器件或步骤。
一种对图像中的物体轮廓进行加强的方法，包括如下步骤：
S1获取拍摄的影像，对影像中的下一帧图像进行分块处理，得到多个图像块；即将影像中的下一帧图像分成多个图像块，下一帧图像基于8×8、16×16或32×32中的任一种方式进行分块，得到对应的多个图像块，当然，还可为其它分块方式，只是选择这几种方式进行分块，使得处理速度更快；
S2基于梯度算法，对每个图像块中的每个像素进行轮廓增强；具体步骤为：
S2.1、基于改进后的梯度计算公式对每个图像块的每一个像素的亮度分量进行梯度计算，得到每个像素的梯度值；
因现有技术中的梯度算法中的梯度在曲面上是连续函数的偏导数，而本发明中的图像的每个像素是离散函数，在梯度计算时，需要采用具体的像素点来计算，所以需要对现有技术中的梯度算法进行改进，改进后的梯度计算公式为：
I＝Σ iΣ j((Y1-Y2)/(Y1+Y2))，
其中，Y1，Y2为相邻两个像素的亮度，i、j分别为像素的横坐标和纵坐标，I为计算的梯度值。
S2.2、判断每个梯度值是否大于设置的阈值，若大于，则判定为边缘，并根据梯度值的下负判定是设置在白边还是黑边；若小于，则不做任何处理。
S3、基于高斯滤波方法，对增强轮廓后的多个图像块作为一个整体进行噪声的滤除，得到滤除噪声后的图像，目的是增加图像的平滑度；
去噪是由于传感器采样，gain等因素必然会引入噪点，或者自己的电子噪声，根据噪点的表现形式，有椒盐噪声、白噪声、高斯噪声等，这样就会降低图像的清晰度和平滑度，因此必须要对图像进行滤波处理，但如果直接基于同一亮暗程度进行滤波处理，同样会降低图像的清晰度和平滑度。
为了解决上述问题，基于高斯滤波方法，对增强轮廓后的多个图像块作为一个整体进行噪声的滤除，得到滤除噪声后的图像，采用的具体步骤为：
S3.1、将增强轮廓后的多个图像块进行加权平均，计算整个图像的亮度值，根据其亮度值的亮暗程度将增强轮廓后的整个图像进行分级；即将增强轮廓后的整个图像根据gain值(亮度值)进行分级，包括四个级别：gain值小于ISO100为一个级别、gain值为ISO200为一个级别、gain值为ISO400为一个级别、gain值为ISO800为一个级别。
S3.2、分级后的多个图像块(即整个图像)的gain值达到亮暗要求，基于高斯滤波公式对达到要求的整个图像进行滤波处理。即gain值小于ISO100时，不做滤波处理；gain值为ISO200时，设置标准差为16，基于高斯滤波公式进行滤波处理；gain值为ISO400时，设置标准差为20，基于高斯滤波公式进行滤波处理；gain值为ISO800时，设置标准差为25，基于高斯滤波公式进行滤波处理；
其中，高斯滤波公式为：
Figure PCTCN2020087871-appb-000002
其中，σ表示标准差，Y表示当前像素的亮度，i、j分别为像素的横坐标和纵坐标。
S4、基于gamma变换，对滤除噪声后的整个图像进行亮暗对比度处理，得 到最终加强后的图像。gamma变换的公式为：
T(Y)＝∑ i∑ j(1+Y) γ，
其中，Y表示当前像素的亮度，γ表示γ系数，i、j分别为像素的横坐标和纵坐标；
当选用不同的γ参数时，加强后的图像效果表现不同，适用于不同的环境应用中，采用较大的gamma参数，可以看到更大的亮度范围，适用于观察图像全局；用较小的gamma参数，重点突出在有限亮度范围的细节，适用于观察图像细节，因此基于gamma变换的公式，设定一个观察图像全局的γ参数值和一个观察图像细节的γ参数值，分别利用两个γ参数值对滤除噪声后的整个图像进行亮暗对比度处理，得到最终的加强后图像T(Y)。
实施例
因我们要实时处理影像中的图像，并进行实时显示，所以在显示当前图像时，就需要对下一帧图像进行处理，即基于32×32，将下一帧图像(1920×1080像素)分成多个60×33.75像素的图像块，若有小数点的情况，通过四舍五入的方式进行图像块的化分，即以60×34像素作为一个图像块，对下一帧图像进行化分，具体为：分别基于32×31和32×26，将下一帧图像1920×1080像素分成的1920×1054和1920×26进行分块处理，处理后得到最终的图像块。
基于改进后的梯度计算公式I＝Σ iΣ j((Y1-Y2)/(Y1+Y2))对每个图像块中的每个像素的亮度分量进行梯度计算，得到每个像素的梯度值；
判断每个梯度值是否大于设置的阈值，若大于，则判定为边缘，并根据梯度值的下负判定是设置在白边还是黑边；若小于，则不做任何处理。
根据增强轮廓后的多个图像块的亮暗程度，将增强轮廓后的多个图像块作为一个整体进行分级；即将增强轮廓后的整个图像根据gain值进行分级，包括四个级别：gain值小于ISO100为一个级别、gain值为ISO200为一个级别、gain值为ISO400为一个级别、gain值为ISO800为一个级别。
分级后的整个图像的gain值达到亮暗要求，基于高斯滤波公式
Figure PCTCN2020087871-appb-000003
对达到要求的整个图像进行滤波处理，否则，不作滤波处理。即gain值小于ISO100时，不做滤波处理；gain值为ISO200时，设置 标准差为16，基于高斯滤波公式进行滤波处理；gain值为ISO400时，设置标准差为20，基于高斯滤波公式进行滤波处理；gain值为ISO800时，设置标准差为25，基于高斯滤波公式进行滤波处理；
基于gamma变换公式T(Y)＝∑ i∑ j(1+Y) γ，对滤除噪声后的整个图像进行亮暗对比度处理，得到最终加强后的图像。
本发明不局限于上述可选实施方式，任何人在本发明的启示下都可得出其他各种形式的产品，但不论在其形状或结构上作任何变化，凡是落入本发明权利要求界定范围内的技术方案，均落在本发明的保护范围之内。

Claims (8)

  1. 一种对实时影像中图像的物体轮廓进行加强的方法，其特征在于，包括如下步骤：
    S1、获取拍摄的影像，对影像中的下一帧图像进行分块处理，得到多个图像块；
    S2、基于梯度算法，对每个图像块中的每个像素进行轮廓增强；
    S3、基于高斯滤波方法，对增强轮廓后的多个图像块作为一个整体进行噪声的滤除，得到滤除噪声后的图像；
    S4、基于gamma变换，对滤除噪声后的整个图像进行亮暗对比度处理，得到最终加强后的图像。
  2. 根据权利要求1所述的一种对实时影像中图像的物体轮廓进行加强的方法，其特征在于，所述步骤S1中，将影像中的下一帧图像分成多个图像块，下一帧图像基于8×8、16×16或32×32中的任一种方式进行分块，得到对应的多个图像块。
  3. 根据权利要求2所述的一种对实时影像中图像的物体轮廓进行加强的方法，其特征在于，所述步骤S2的具体步骤为：
    S2.1、基于改进后的梯度计算公式对每个图像块的每一个像素的亮度分量进行梯度计算，得到每个像素的梯度值；
    S2.2、判断每个梯度值是否大于设置的阈值，若大于，则判定为边缘，并根据梯度值的下负判定是设置在白边还是黑边；若小于，则不做任何处理。
  4. 根据权利要求3所述的一种对实时影像中图像的物体轮廓进行加强的方法，其特征在于，所述步骤S2.1中，改进后的梯度计算公式为：
    I＝∑ i∑ j((Y1-Y2)/(Y1+Y2))，
    其中，Y1，Y2为相邻两个像素的亮度，i、j分别为像素的横坐标和纵坐标，I为计算的梯度值。
  5. 根据权利要求1或4所述的一种对实时影像中图像的物体轮廓进行加强的方法，其特征在于，所步骤S3的具体步骤为：
    S3.1、根据增强轮廓后的多个图像块的亮暗程度，将增强轮廓后的多个图像块作为一个整体进行分级；
    S3.2、分级后的整个图像的gain值达到亮暗要求，基于高斯滤波公式对达 到要求的整个图像进行滤波处理，否则，不作滤波处理。
  6. 根据权利要求5所述的一种对实时影像中图像的物体轮廓进行加强的方法，其特征在于，所述步骤3.1中，将增强轮廓后的多个图像块进行加权平均，计算整个图像的亮度值，根据其亮度值的亮暗程度将增强轮廓后的整个图像进行分级；即根据gain值进行分级，包括四个级别：gain值小于ISO100为一个级别、gain值为ISO200为一个级别、gain值为ISO400为一个级别、gain值为ISO800为一个级别。
  7. 根据权利要求5所述的一种对实时影像中图像的物体轮廓进行加强的方法，其特征在于，所述步骤3.2中，gain值小于ISO100时，不做滤波处理；gain值为ISO200时，设置标准差为16，基于高斯滤波公式进行滤波处理；gain值为ISO400时，设置标准差为20，基于高斯滤波公式进行滤波处理；gain值为ISO800时，设置标准差为25，基于高斯滤波公式进行滤波处理；
    其中，高斯滤波公式为：
    Figure PCTCN2020087871-appb-100001
    其中，σ表示标准差，Y表示当前像素的亮度，i、j分别为像素的横坐标和纵坐标。
  8. 根据权利要求1或7所述的一种对实时影像中图像的物体轮廓进行加强的方法，其特征在于，所述步骤4中，gamma变换的公式为：
    T(Y)＝∑ i∑ j(1+Y) γ,
    其中，Y表示当前像素的亮度，γ表示γ系数，i、j分别为像素的横坐标和纵坐标；
    基于gamma变换的公式，设定一个观察图像全局的γ参数值和一个观察图像细节的γ参数值，分别利用两个γ参数值对滤除噪声后的整个图像进行亮暗对比度处理，得到最终的加强后图像T(Y)。
PCT/CN2020/087871 2019-05-05 2020-04-29 一种对实时影像中图像的物体轮廓进行加强的方法 WO2020224508A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
EP20802689.8A EP3965053A4 (en) 2019-05-05 2020-04-29 METHOD OF ENHANCING THE OBJECT CONTOUR OF AN IMAGE IN REAL-TIME VIDEO
US17/608,991 US20230401675A1 (en) 2019-05-05 2020-04-29 Method for enhancing object contour of image in real-time video
JP2021565940A JP7076168B1 (ja) 2019-05-05 2020-04-29 リアルタイム映像における画像の物体輪郭を強調する方法

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910368593.X 2019-05-05
CN201910368593.XA CN110335203B (zh) 2019-05-05 2019-05-05 一种对实时影像中图像的物体轮廓进行加强的方法

Publications (1)

Publication Number Publication Date
WO2020224508A1 true WO2020224508A1 (zh) 2020-11-12

Family

ID=68139330

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/087871 WO2020224508A1 (zh) 2019-05-05 2020-04-29 一种对实时影像中图像的物体轮廓进行加强的方法

Country Status (5)

Country Link
US (1) US20230401675A1 (zh)
EP (1) EP3965053A4 (zh)
JP (1) JP7076168B1 (zh)
CN (1) CN110335203B (zh)
WO (1) WO2020224508A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115409745A (zh) * 2022-10-31 2022-11-29 深圳市亿康医疗技术有限公司 一种应用于放疗准备的ct图像的增强方法
CN116433537A (zh) * 2023-06-13 2023-07-14 济南科汛智能科技有限公司 基于物联网和云计算的智慧病房监控系统

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110335203B (zh) * 2019-05-05 2022-02-01 湖南省华芯医疗器械有限公司 一种对实时影像中图像的物体轮廓进行加强的方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5557635B2 (ja) * 2010-07-21 2014-07-23 Ｈｏｙａ株式会社 輪郭強調装置
CN104574284A (zh) * 2013-10-24 2015-04-29 南京普爱射线影像设备有限公司 一种数字x射线图像对比度增强处理方法
CN106920218A (zh) * 2015-12-25 2017-07-04 展讯通信（上海）有限公司 一种图像处理的方法及装置
CN108898152A (zh) * 2018-05-14 2018-11-27 浙江工业大学 一种基于多通道多分类器的胰腺囊性肿瘤ct图像分类方法
CN110335203A (zh) * 2019-05-05 2019-10-15 湖南省华芯医疗器械有限公司 一种对实时影像中图像的物体轮廓进行加强的方法

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6059729A (en) * 1998-10-19 2000-05-09 Stonger; Kelly A. Method and apparatus for edge enhancement in ultrasound imaging
US7088474B2 (en) * 2001-09-13 2006-08-08 Hewlett-Packard Development Company, Lp. Method and system for enhancing images using edge orientation
US6891549B2 (en) * 2002-01-11 2005-05-10 Applied Materials, Inc. System and method for edge enhancement of images
US7529422B2 (en) * 2004-09-22 2009-05-05 Siemens Medical Solutions Usa, Inc. Gradient-based image restoration and enhancement
US7587099B2 (en) * 2006-01-27 2009-09-08 Microsoft Corporation Region-based image denoising
JP2012515952A (ja) * 2009-01-20 2012-07-12 コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ 強化画像を生成するための方法及び装置
JP5828649B2 (ja) 2011-03-09 2015-12-09 キヤノン株式会社 画像処理装置、画像処理方法、及びコンピュータプログラム
JP6552325B2 (ja) * 2015-08-07 2019-07-31 キヤノン株式会社 撮像装置、撮像装置の制御方法、及びプログラム

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5557635B2 (ja) * 2010-07-21 2014-07-23 Ｈｏｙａ株式会社 輪郭強調装置
CN104574284A (zh) * 2013-10-24 2015-04-29 南京普爱射线影像设备有限公司 一种数字x射线图像对比度增强处理方法
CN106920218A (zh) * 2015-12-25 2017-07-04 展讯通信（上海）有限公司 一种图像处理的方法及装置
CN108898152A (zh) * 2018-05-14 2018-11-27 浙江工业大学 一种基于多通道多分类器的胰腺囊性肿瘤ct图像分类方法
CN110335203A (zh) * 2019-05-05 2019-10-15 湖南省华芯医疗器械有限公司 一种对实时影像中图像的物体轮廓进行加强的方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3965053A4 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115409745A (zh) * 2022-10-31 2022-11-29 深圳市亿康医疗技术有限公司 一种应用于放疗准备的ct图像的增强方法
CN116433537A (zh) * 2023-06-13 2023-07-14 济南科汛智能科技有限公司 基于物联网和云计算的智慧病房监控系统
CN116433537B (zh) * 2023-06-13 2023-08-11 济南科汛智能科技有限公司 基于物联网和云计算的智慧病房监控系统

Also Published As

Publication number Publication date
EP3965053A1 (en) 2022-03-09
JP2022527868A (ja) 2022-06-06
CN110335203A (zh) 2019-10-15
JP7076168B1 (ja) 2022-05-27
US20230401675A1 (en) 2023-12-14
CN110335203B (zh) 2022-02-01
EP3965053A4 (en) 2022-07-06

Similar Documents

Publication Publication Date Title
WO2020224508A1 (zh) 一种对实时影像中图像的物体轮廓进行加强的方法
Crihalmeanu et al. Enhancement and registration schemes for matching conjunctival vasculature
CN110772286B (zh) 一种基于超声造影识别肝脏局灶性病变的系统
Li et al. Robust retinal image enhancement via dual-tree complex wavelet transform and morphology-based method
CN105741241B (zh) 基于合成增强图像的肿瘤区域图像增强方法及系统
CN110992377B (zh) 图像分割方法、装置、计算机可读存储介质和设备
US8351667B2 (en) Methods of contrast enhancement for images having blood vessel structures
CN111223110A (zh) 一种显微图像增强方法、装置及计算机设备
CN110097610B (zh) 基于超声与磁共振成像的语音合成系统和方法
Ratheesh et al. Advanced algorithm for polyp detection using depth segmentation in colon endoscopy
Furizal et al. Analysis of the Influence of Number of Segments on Similarity Level in Wound Image Segmentation Using K-Means Clustering Algorithm
Almi'ani et al. Automatic segmentation algorithm for brain MRA images
Lee et al. A graph-based segmentation method for breast tumors in ultrasound images
CN114627136B (zh) 一种基于特征金字塔网络的舌象分割与对齐方法
CN114187241A (zh) 一种基于肺部超声的胸膜线识别方法及系统
EP2313848A1 (en) Methods for enhancing vascular patterns in cervical imagery
Chang et al. A novel retinal blood vessel segmentation method based on line operator and edge detector
KR102053890B1 (ko) PCM 기반 양자화를 이용한 X-Ray 영상 기반 장폐색 검출방법
Muthu et al. Morphological operations in medical image pre-processing
Patibandla et al. CT Image Precise Denoising Model with Edge Based Segmentation with Labeled Pixel Extraction Using CNN Based Feature Extraction for Oral Cancer Detection
Wirth et al. Combination of color and focus segmentation for medical images with low depth-of-field
Goyal Gaussian filtering based image integration for improved disease diagnosis and treatment planning
Wei et al. Teniae coli extraction in human colon for computed tomographic colonography images
Shi et al. A method for enhancing lung nodules in chest radiographs by use of LoG Filter
ATHIRA et al. Liver Abnormality Detection using Segmentation based Fractal Texture Analysis

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: 20802689

Country of ref document: EP

Kind code of ref document: A1

DPE2 Request for preliminary examination filed before expiration of 19th month from priority date (pct application filed from 20040101)
ENP Entry into the national phase

Ref document number: 2021565940

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2020802689

Country of ref document: EP

Effective date: 20211130