[go: up one dir, main page]

WO2018157868A1 - Système d'imagerie ultrasonore et son procédé d'imagerie - Google Patents

Système d'imagerie ultrasonore et son procédé d'imagerie Download PDF

Info

Publication number
WO2018157868A1
WO2018157868A1 PCT/CN2018/077941 CN2018077941W WO2018157868A1 WO 2018157868 A1 WO2018157868 A1 WO 2018157868A1 CN 2018077941 W CN2018077941 W CN 2018077941W WO 2018157868 A1 WO2018157868 A1 WO 2018157868A1
Authority
WO
WIPO (PCT)
Prior art keywords
ultrasonic
image
nerve
imaging system
region
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.)
Ceased
Application number
PCT/CN2018/077941
Other languages
English (en)
Chinese (zh)
Inventor
江清泉
张瑞峰
王曼玲
黄耀贤
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.)
Individual
Original Assignee
Individual
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 Individual filed Critical Individual
Publication of WO2018157868A1 publication Critical patent/WO2018157868A1/fr
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
    • A61B8/08—Clinical applications

Definitions

  • This invention relates to an ultrasound imaging system, and more particularly to an ultrasound imaging system for automatically distinguishing peripheral nerve images and/or anesthetic drugs in real time and a method of imaging the same.
  • Peripheral nerve block is a local anesthetic injection around the target nerve to temporarily block the sensory and motor nerve conduction in a specific area, reduce pain during surgery, and strengthen postoperative recovery.
  • the Haar feature based cascade classifiers are an effective target detection method proposed by Paul Viola and Michael Jones in 2001 (P. Viola and M. Jones, "Rapid object detection using a boosted cascade of simple features," in Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001, 2001, pp. I-511-I-518 vol. 1.).
  • the machine learning method proposed by Paul and Michael for target detection can quickly process images and achieve high detection rates.
  • Target detection has three key contributions: integral imagery, AdaBoost-based learning calculations, and the ability to combine complex classifiers into a "cascading" approach.
  • AdaBoost learning calculation method requires a large number of positive images (only images of target and surrounding tissues) and negative images (with no target images) to train the classifier.
  • Fig. 1(a)-Fig.1(c) A schematic diagram of the Hal character shown.
  • Each feature is a single value that is calculated by subtracting the pixels under the black rectangle from the sum of the pixels under the white rectangle. Most of the calculated features are irrelevant. Therefore, the variant of AdaBoost is to perform the selection of a small feature set and training classifier (Y.Freund and RESchapire, "A Decision-Theoretic Generalization of On-Line Learning and an Application to Boosting," Journal of Computer and System Sciences , vol. 55, pp. 119-139, 1997/08/01 1997.), in its original form, used to improve the classification performance of weak learning calculations.
  • a weak learning calculation method was designed to select a single rectangular feature that can best separate the positive and negative images. For each feature, the weak learner determines the optimal threshold classification function.
  • a "cascade" classifier structure is used. The structure of the cascaded classifier not only achieves detection performance but also reduces computation time. The key reason is that it is smaller and more efficient, so the upgraded classifier can reject many negative subwindows (cascades) when positive instances are detected. 2 is a schematic structural view of a cascade classifier.
  • the present invention utilizes the Hal feature of level-based learner learning to detect nerves, and can help an anesthesiologist to automatically address a neural region immediately prior to surgery.
  • the present invention provides an ultrasonic imaging system, comprising: an ultrasonic probe for acquiring an ultrasonic image including a nerve and surrounding tissue; and an ultrasonic device coupled to the ultrasonic probe for collecting the ultrasonic image; An image capturing device is connected to the ultrasonic device for converting the ultrasonic image; and an arithmetic device is connected to the image capturing device for scanning the ultrasonic image to obtain an ultrasonic scanned image, and the ultrasonic wave A specific feature in the scanned image is indicated by a color; and a developing device is coupled to the computing device for displaying the ultrasonic scanned image in real time and the color for indicating the particular feature.
  • the computing device may further mark a plurality of the specific features in the ultrasonic scanned image in a plurality of the colors.
  • the developing device can display a plurality of the ultrasonic scanned images in real time, and mark a plurality of the colors of the plurality of the specific features.
  • a plurality of said specific features may be nerve regions or local anesthetic regions.
  • the ultrasonic source device and the image capturing device further have an image source line.
  • the image taking device and the calculating device further have a connecting line.
  • the present invention also provides a method for developing an ultrasonic imaging system, which can be used for detecting a nerve region, and is characterized by the following steps: Step 1, based on the characteristics of Haar, analyzing the ultrasonic image, detecting Detect possible neural candidate regions to achieve real-time imaging effects; Step 2, depth-based neural network identification method, complete neural region localization in real time; and step 3, according to image pixel intensity distribution in the neural region, by The image with high pixel value is discolored, the nerve tissue is separated from the background, and the nerve tissue is presented in real time with a color.
  • the imaging method can be used for detecting an anesthetic drug region, and the method comprises the following steps: Step 1, based on the Haar feature, analyzing the ultrasound image to detect possible nerve candidate regions, The effect of real-time imaging is achieved; in step 2, the depth-based neural network identification method is used to complete the localization of the neural region in real time; and in step 3, according to the image pixel intensity distribution in the neural region, by changing the image with high pixel value, The nerve tissue is separated from the background, and the nerve tissue is presented in real time in one color; and in step 4, the nerve tissue is mainly used, the detection range is increased, the pixel intensity around the nerve tissue is analyzed, and the range of the anesthetic drug is confirmed, and The range is presented in real time in another color.
  • Figure 1 (a) is a schematic diagram of the Haar feature of a rectangular feature (edge feature).
  • Fig. 1(b) is a schematic diagram of Haar characteristics of a 3 rectangular feature (line feature).
  • Fig. 1(c) is a schematic diagram of the Haar feature of a 4-rectangular feature.
  • FIG. 2 is a schematic structural view of a cascade classifier.
  • Figure 3 is a schematic diagram of a common ultrasound guided peripheral nerve block system.
  • FIG. 4 is a schematic view of an embodiment of an ultrasonic imaging system of the present invention.
  • Fig. 5 is a flow chart showing a method of developing a nerve region of the ultrasonic imaging system of the present invention.
  • Figure 6 is a schematic view of a detection area in a neuroimaging image.
  • Figure 7 is a schematic diagram of the coloration of the nerve region.
  • Figure 8 is a schematic diagram of a traceable development mode.
  • Fig. 9(a) is a schematic view of an ultrasound image including nerves and surrounding tissues.
  • Fig. 9(b) is a schematic view of an ultrasonic image excluding nerves.
  • Fig. 10 is a flow chart showing an anesthetic drug region developing method of the ultrasonic imaging system of the present invention.
  • Figure 11 is a schematic view showing the coloration of the nerve region and the anesthetic drug region.
  • FIG. 3 is a schematic diagram of a conventional ultrasonic guided peripheral nerve blocking system 1 including an ultrasonic probe 11, an ultrasonic device 12, and an ultrasonic image playback device 13.
  • the ultrasonic guided peripheral nerve blocking system 1 of FIG. 3 obtains an ultrasonic image including the nerve and the surrounding tissue using the ultrasonic probe 11, and then transmits the ultrasonic image to the ultrasonic image playing device 13 via the ultrasonic device 12, and then the operator
  • the ultrasonic image on the ultrasonic image playback device 13 is visually observed, the nerve position in the ultrasonic image is judged, or the nerve stimulator is used to strike the human body, and the nerve is moved to identify the true nerve position to inject the anesthetic.
  • the shortcoming of the ultrasound-guided peripheral nerve block system 1 is that it depends on the experience of the operating physician to correctly determine the peripheral nerve position, and cannot perform real-time analysis to assist diagnosis and treatment; in addition, knocking on the human body may also cause other human body Part of the damage.
  • the ultrasonic imaging system 2 includes an ultrasonic probe 21, an ultrasonic device 22, an image capturing device 23, a calculating device 24, and a developing device 25.
  • the image source line can be DVI, HDMI, or DSUB.
  • the image capture device 23 includes, but is not limited to, a capture cartridge.
  • computing device 24 includes, but is not limited to, a notebook computer.
  • imaging device 25 includes, but is not limited to, a computer screen.
  • the ultrasonic imaging system 2 acquires an ultrasonic image including a nerve and a surrounding tissue using the ultrasonic probe 21, and transmits the ultrasonic image to the image capturing device 23 via the ultrasonic device 22 and the image source line 221, and then transmits the image capturing device. 23.
  • the ultrasonic image is converted to the calculation device 24 via the connection line 231.
  • the calculating device 24 scans the ultrasonic image by using a neural detection algorithm, and then displays the nerve tissue in the ultrasonic scanned image in real time with a significant color, and displays the result on the developing device 25.
  • the ultrasonic imaging system 2 provided by the present invention can correctly mark the nerve position in real time, the ultrasonic imaging system 2 provided by the present invention can be used without the nerve stimulation as compared with the ultrasonic guided peripheral nerve blocking system 1 of FIG.
  • an auxiliary tool such as a device, it is possible to identify a true nerve position from a soft tissue including a nerve, to put the drug into the correct area, to help an operator with and/or experience, and to more easily identify and evaluate the local anesthetic.
  • the injection position avoids the misuse of drugs and has an adverse effect on the human body.
  • FIG. 5 is a flow chart of a developing method of the ultrasonic imaging system 2 of the present invention
  • FIG. 6 is a schematic view of a detecting area in a neuroimaging image
  • FIG. 7 is a schematic view showing a coloring of a nerve region.
  • the imaging method of the ultrasonic imaging system 2 provided by the present invention has the following steps:
  • Step S241 the first stage of the neural candidate region detecting step, scanning and analyzing the ultrasonic image based on the Haar feature, and detecting a possible neural candidate region (as shown in FIG. 6), in order to achieve real-time imaging effect.
  • the area selected in step S241 is not completely correct or unique;
  • Step S242 the second stage of the neural region identification step, using the depth-like neural network to complete the positioning of the neural region in real time for the ultrasound scan image;
  • Step S243 a step of identifying the nerve tissue in the nerve region, and identifying the nerve tissue according to the image pixel intensity distribution in the neural region, and dissociating the nerve tissue from the background by discoloring the image with high pixel value, and highlighting The color presents the neural tissue in real time.
  • step S241 to step S243 is performed by the calculating device 24 of the ultrasonic developing system 2, and the developing result is displayed on the developing device 25 in real time by the developing device 25 (as shown in FIG. 7). Show) to assist the operator in real-time analysis of the correct nerve location.
  • FIG. 8 is a schematic diagram of a traceable development mode.
  • the imaging method of the ultrasonic imaging system 2 provided by the present invention can continuously mark the nerve tissue scanned in the ultrasonic image on the developing device 25 in real time with different significant colors.
  • the neural tissue of the first ultrasonic scanned image is presented in the first color on the developing device 25, and the second ultrasonic scanned image is obtained from the plurality of ultrasonic scanned images confirmed in steps S241 to S243.
  • the neural tissue will be presented in a second color on the imaging device 25... and so on, providing the operator with a traceable imaging mode.
  • Fig. 9(a) is a schematic view of an ultrasonic image including nerves and surrounding tissues
  • Fig. 9(b) is a schematic view of an ultrasonic image not including nerves.
  • the present invention utilizes a Hal feature and a cascading learning calculation extracted from training images (including but not limited to the ultrasound images of Figures 9(a) and 9(b) to find regions of the nerve to The development method of steps S241 to S243 is implemented.
  • Fig. 10 is a flow chart showing another development method of the ultrasonic imaging system 2 of the present invention
  • Fig. 11 is a schematic view showing the coloration of the nerve region and the anesthetic drug region.
  • the method of the present invention for finding a nerve region can be used to find an area of anesthetic drug that is applied around a nerve region.
  • the imaging method of the ultrasonic imaging system 2 provided by the present invention in some embodiments, has the following steps:
  • Step S241 the first stage of the neural candidate region detecting step, scanning and analyzing the ultrasonic image based on the Haar feature, and detecting a possible neural candidate region (as shown in FIG. 6), in order to achieve real-time imaging effect.
  • the area selected in step S241 is not completely correct or unique;
  • Step S242 the second stage of the neural region identification step, using the depth-like neural network to complete the positioning of the neural region in real time for the ultrasound scan image;
  • Step S2431 a step of identifying the nerve tissue in the nerve region, identifying the nerve tissue according to the image pixel intensity distribution in the neural region, and dissociating the nerve tissue from the background by discoloring the image with high pixel value, and highlighting The color presents the neural tissue in real time;
  • Step S2432 a step of identifying the anesthetic drug around the nerve tissue, focusing on the nerve tissue, increasing the detection range, analyzing the pixel intensity around the nerve tissue, confirming the range of the anesthetic drug, and presenting the effect in real time in another significant color. range.
  • step S241 to step S2432 is performed by the calculating device 24 of the ultrasonic developing system 2, and the developing result is displayed on the developing device 25 in real time by the developing device 25 (as shown in FIG. 11). To assist the operator in confirming the scope of the anesthetic that has been applied to determine whether the anesthetic is completely covered with nerve tissue.
  • the identified pixel value of the anesthetic is differentiated by 60 pixels, and if the pixel value is below 60 pixels, it is not rendered in a significant color.
  • the imaging device 25 can display not only the original ultrasound image, but also an ultrasound scan image that colors the nerve region and/or the anesthetic drug region for the operator to confirm the original ultrasound image and the colored ultrasound scan. Whether the image positions match each other.
  • the present invention assists an unexperienced peripheral nerve block operation by automatically addressing the neural region immediately and marking the position of the nerve region and/or the location of the anesthetic drug region in a significant color by using the Hal feature of the cascade learning calculation method.
  • the physician identifies the true nerve position in the soft tissue including the nerve; the invention can determine the correct nerve position without using the nerve stimulator to strike the human body, and avoids the need for the traditional peripheral nerve block surgery to hit the human body with the auxiliary tool The problem of damage to the ligament or soft tissue, etc.; the invention also avoids the adverse effects of the anesthetic drug on the human body, and confirms the scope of the anesthetic drug, thereby determining that the drug can completely cover the nerve tissue, further Improve the success rate and effectiveness of nerve blockade. Therefore, the present invention not only contributes to the technical education of peripheral nerve blockade, but also is beneficial to clinical diagnosis and treatment, solves the existing technical problems, and achieves better effects.

Landscapes

  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Pathology (AREA)
  • Radiology & Medical Imaging (AREA)
  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Ultra Sonic Daignosis Equipment (AREA)

Abstract

La présente invention concerne un système d'imagerie ultrasonore (2), comprenant : une sonde ultrasonore (21) servant à acquérir une image ultrasonore comprenant des nerfs et des tissus environnants; un dispositif ultrasonore (22) relié à la sonde ultrasonore (21) et servant à recueillir l'image ultrasonore; un dispositif de capture d'image (23) relié au dispositif ultrasonore (22) et servant à convertir l'image ultrasonore; un dispositif de calcul (24) relié au dispositif de capture d'image (23), et servant à balayer l'image ultrasonore et à marquer une caractéristique particulière de l'image ultrasonore avec une couleur; et un dispositif d'imagerie (25) relié au dispositif de calcul (24) et servant à afficher en temps réel la couleur de la caractéristique particulière et de l'image balayée de l'image ultrasonore. La présente invention concerne en outre un procédé d'imagerie du système d'imagerie ultrasonore.
PCT/CN2018/077941 2017-03-02 2018-03-02 Système d'imagerie ultrasonore et son procédé d'imagerie Ceased WO2018157868A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US201762465845P 2017-03-02 2017-03-02
US62/465,845 2017-03-02

Publications (1)

Publication Number Publication Date
WO2018157868A1 true WO2018157868A1 (fr) 2018-09-07

Family

ID=63369860

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/077941 Ceased WO2018157868A1 (fr) 2017-03-02 2018-03-02 Système d'imagerie ultrasonore et son procédé d'imagerie

Country Status (1)

Country Link
WO (1) WO2018157868A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20210077061A1 (en) * 2019-09-18 2021-03-18 GE Precision Healthcare LLC Method and system for analyzing ultrasound scenes to provide needle guidance and warnings

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070167760A1 (en) * 2005-12-01 2007-07-19 Medison Co., Ltd. Ultrasound imaging system and method for forming a 3d ultrasound image of a target object
US20110228997A1 (en) * 2010-03-17 2011-09-22 Microsoft Corporation Medical Image Rendering
CN103479398A (zh) * 2013-09-16 2014-01-01 华南理工大学 一种基于超声射频流分析的肝组织微结构的检测方法
CN104840209A (zh) * 2014-02-19 2015-08-19 三星电子株式会社 用于病变检测的设备和方法
CN105232081A (zh) * 2014-07-09 2016-01-13 无锡祥生医学影像有限责任公司 医学超声辅助自动诊断装置及方法
CN105574820A (zh) * 2015-12-04 2016-05-11 南京云石医疗科技有限公司 一种基于深度学习的自适应超声图像增强方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070167760A1 (en) * 2005-12-01 2007-07-19 Medison Co., Ltd. Ultrasound imaging system and method for forming a 3d ultrasound image of a target object
US20110228997A1 (en) * 2010-03-17 2011-09-22 Microsoft Corporation Medical Image Rendering
CN103479398A (zh) * 2013-09-16 2014-01-01 华南理工大学 一种基于超声射频流分析的肝组织微结构的检测方法
CN104840209A (zh) * 2014-02-19 2015-08-19 三星电子株式会社 用于病变检测的设备和方法
CN105232081A (zh) * 2014-07-09 2016-01-13 无锡祥生医学影像有限责任公司 医学超声辅助自动诊断装置及方法
CN105574820A (zh) * 2015-12-04 2016-05-11 南京云石医疗科技有限公司 一种基于深度学习的自适应超声图像增强方法

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20210077061A1 (en) * 2019-09-18 2021-03-18 GE Precision Healthcare LLC Method and system for analyzing ultrasound scenes to provide needle guidance and warnings
CN112515747A (zh) * 2019-09-18 2021-03-19 通用电气精准医疗有限责任公司 用于分析超声场景以提供针引导和警告的方法和系统

Similar Documents

Publication Publication Date Title
CN113506294B (zh) 一种医疗图像评估方法、系统、计算机设备及存储介质
CN111227864B (zh) 使用超声图像利用计算机视觉进行病灶检测的装置
CN111214255B (zh) 一种医学超声图像计算机辅助方法
Ghosh et al. An economised craniofacial identification system
Horstmann et al. More efficient rejection of happy than of angry face distractors in visual search
RU2633320C2 (ru) Отбор изображений для оптического исследования шейки матки
US12456282B2 (en) Ultrasound imaging method and system for identifying an anatomical feature of a spine
CN113707279B (zh) 医学影像图片的辅助分析方法、装置、计算机设备及介质
JP2000331167A (ja) 顔画像照合方法及び装置
WO2018157868A1 (fr) Système d'imagerie ultrasonore et son procédé d'imagerie
US20240020823A1 (en) Assistance diagnosis system for lung disease based on deep learning and assistance diagnosis method thereof
KR102095730B1 (ko) 딥러닝 기반의 대장 병변 검출 방법
TWI694812B (zh) 臨床牙科教育訓練系統及其方法
WO2021166094A1 (fr) Système de marquage pour image médicale, et dispositif d'assistance au marquage
Dittmar et al. A feedback estimation approach for therapeutic facial training
CN114242243A (zh) 用户健康评估方法、装置、设备以及存储介质
Droste et al. Discovering salient anatomical landmarks by predicting human gaze
Liu et al. Hepatitis diagnosis using facial color image
Vilarino et al. Automatic labeling of colonoscopy video for cancer detection
CN115035086B (zh) 一种基于深度学习的结核皮试智能筛查分析方法和装置
KR102501665B1 (ko) 입술 움직임 정보 분석 방법 및 장치
Mukku et al. A specular reflection removal technique in cervigrams
KR20230073499A (ko) 핵심영역 선택 기반 시각장면 분석 방법
Avent et al. Machine vision recognition of facial affect using backpropagation neural networks
Hadjerci et al. On-line learning dynamic models for nerve detection in ultrasound videos

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: 18760927

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18760927

Country of ref document: EP

Kind code of ref document: A1