Method for batch desensitization and format conversion of digital imaging and communications in medicine (DICOM) files of echocardiogram

Abstract

A method for batch desensitization and format conversion of digital imaging and communications in medicine (DICOM) files of an echocardiogram includes acquiring 100 a DICOM file and removing 200 personal data from it to obtain a desensitized file. A region of interest (ROI) is extracted 300 from the desensitized file using an adaptive mask image and a prior mask image to obtain a file of interest. Coding and image pixel information processing is performed 400 on the file of interest based on an image output format to obtain a DICOM file of interest in a specified storage format. The removing of sensitive information may include using machine vision to identify and remove a sensitive region in an image. The prior mask image may be used for a single frame file whereas for a multi frame file the adaptive mask may be used. The adaptive may be obtained from motion information. The methods can be used in pre-processing cardiac ultrasound image data to be stored in an image archive suitable for training artificial intelligence algorithms.

Description

[0001] The present disclosure relates to the technical field of digital imaging and communications in medicine (DICOM) file processing, and in particular, to a method for batch desensitization and format conversion of DICOM files of an echocardiogram. BACKGROUND

[0002] Echocardiograms are mainly stored in a picture archiving and communication system (PACS) in DICOM format, and include single-frame images and multi-frame videos. Echocardiography, with rich time domain information, no radiation, high cost performance, and other advantages, leads to an increasing clinical demand for ultrasound examination. In recent years, artificial intelligence has rapidly developed in the field of intelligent analysis of echocardiography. By analysis of echocardiography through artificial intelligence, multiple cardiac structural and functional parameters can be automatically obtained, which has an important clinical value for diagnosis and treatment of cardiovascular diseases. Large samples and standardized ultrasound data are cornerstones for developing artificial intelligence algorithms. Therefore, DICOM image desensitization and format conversion become key steps of intelligent analysis using large sample ultrasound data. However, DICOM images contain a lot of pixel information and metadata related to patient information, so that manual processing and management are tedious and time-consuming, with a high error rate and low accuracy and efficiency of data analysis. In addition, there are differences in image parameters (such as a dimension and a fan angle), and different medical devices and manufacturers may produce DICOM files with different standards and settings, which increases heterogeneity of data and leads to low ease of use of data. SUMMARY

[0003] In order to overcome the shortcomings of the prior art, an objective of the present disclosure is to provide a method for batch desensitization and format conversion of DICOM files of an echocardiogram, to solve the problems of low efficiency of DICOM data analysis and low ease of use of data thereof in the prior art.

[0004] To achieve the above objective, the present disclosure provides the following solutions.

[0005] A method for batch desensitization and format conversion of DICOM files of an echocardiogram includes:

[0006] acquiring a to-be-processed DICOM file;

[0007] desensitizing the to-be-processed DICOM file to obtain a desensitized DICOM file;

[0008] extracting a region of interest (ROI) from the desensitized DICOM file based on an adaptive mask image and a prior mask image to obtain a DICOM file of interest; and

[0009] performing coding and image pixel information processing on the DICOM file of interest based on a DICOM image output format to obtain a DICOM file of interest in a specified storage format.

[0010] Preferably, the desensitizing the to-be-processed DICOM file to obtain a desensitized DICOM file includes:

[0011] parsing and identifying the to-be-processed DICOM file by using a pydicom library of Python, and replacing sensitive personal information of a patient with a preset character string to obtain an initial desensitized DICOM file;

[0012] reading the initial desensitized DICOM file by using an open source computer vision (OpenCV) library of Python, and identifying sensitive information to obtain a sensitive region; and

[0013] setting a pixel value of the sensitive region to 0 to obtain the desensitized DICOM file.

[0014] Preferably, the extracting an ROI from the desensitized DICOM file based on an adaptive mask image and a prior mask image to obtain a DICOM file of interest includes:

[0015] acquiring a single-frame DICOM file and a multi-frame DICOM file of the desensitized DICOM file;

[0016] extracting an ROI from the single-frame DICOM file based on the prior mask image to obtain a first DICOM file;

[0017] extracting an ROI from the multi-frame DICOM file based on the adaptive mask image to obtain a second DICOM file; and

[0018] obtaining the DICOM file of interest based on the first DICOM file and the second DICOM file.

[0019] Preferably, the extracting the ROI from the multi-frame DICOM file based on the adaptive mask image to obtain the second DICOM file includes:

[0020] acquiring motion information of the multi-frame DICOM file, and obtaining an initial mask image based on the motion information;

[0021] extracting a bounding rectangle of the mask image based on an OpenCV library of Python and constructing the adaptive mask image; and

[0022] extracting the ROI from the multi-frame DICOM file based on the adaptive mask image to obtain the second DICOM file.

[0023] Preferably, the performing coding and image pixel information processing on the DICOM file of interest based on a DICOM image output format to obtain a DICOM file of interest in a specified format includes:

[0024] parsing the DICOM file of interest by using a pydicom library to obtain color space information of the DICOM file of interest;

[0025] determining whether the color space information is a red green blue (RGB) color space, and if so, acquiring pixel information and bit depth information of the DICOM file of interest; or if not, converting the color space information into a specified color space by using a conversion function and acquiring the pixel information and the bit depth information;

[0026] acquiring a number of frames and a dynamic range of a pixel value of the DICOM file of interest based on the pixel information and the bit depth information;

[0027] determining whether the number of frames of the DICOM file of interest is 1, and if so, obtaining an image storage format of the first DICOM file; or if not, acquiring a frame rate, an image length and an image width of the second DICOM file, and acquiring an image storage format of the second DICOM file based on the bit depth information, and the frame rate, the image length and the image width of the second DICOM file; and

[0028] obtaining the DICOM file of interest with a specified image storage format based on the image storage format of the second DICOM file and the image storage format of the first DICOM file.

[0029] According to specific embodiments of the present disclosure, the present disclosure has the following technical effects.

[0030] The present disclosure provides a method for batch desensitization and format conversion of DICOM files of an echocardiogram. The method includes: acquiring a to-be-processed DICOM file; desensitizing the to-be-processed DICOM file to obtain a desensitized DICOM file; extracting an ROI from the desensitized DICOM file based on an adaptive mask image and a prior mask image to obtain a DICOM file of interest; and performing coding and image pixel information processing on the DICOM file of interest based on a DICOM image output format to obtain a DICOM file of interest in a specified storage format. According to the present disclosure, the to-be-processed DICOM file is desensitized, so that data security is improved. An ROI of data is extracted, so that efficiency and accuracy of data analysis are improved. Batch format conversion is performed on the data, thereby improving ease of use of the data. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] To describe the technical solutions in embodiments of the present disclosure or in the prior art more clearly, the accompanying drawings required in the embodiments are briefly described below. Apparently, the accompanying drawings in the following description show merely some embodiments of the present disclosure, and other drawings can be derived from these accompanying drawings by those of ordinary skill in the art without creative efforts.

[0032] FIG. 1 is a flowchart of a method for batch desensitization and format conversion of DICOM files of an echocardiogram according to an embodiment of the present disclosure;

[0033] FIG. 2 is a flowchart showing batch desensitization processing of DICOM data according to an embodiment of the present disclosure;

[0034] FIG. 3 is a flowchart showing batch extraction ofROIs from an echocardiogram according to an embodiment of the present disclosure; and

[0035] FIG. 4 is a flowchart showing batch format conversion of DICOM data of an echocardiogram according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0036] The technical solutions of the embodiments of the present disclosure are clearly and completely described below with reference to the accompanying drawings in the embodiments of the present disclosure. Apparently, the described embodiments are merely some rather than all of the embodiments of the present disclosure. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present disclosure without creative efforts shall fall within the protection scope of the present disclosure.

[0037] An objective of the present disclosure is to provide a method for batch desensitization and format conversion of DICOM files of an echocardiogram, to solve the problems of low efficiency of DICOM data analysis and low ease of use of data thereof in the prior art.

[0038] In order to make the above objective, features and advantages of the present disclosure clearer and more comprehensible, the present disclosure is further described in detail below with reference to the accompanying drawings and specific implementations.

[0039] As shown in FIG. 1, the present disclosure provides a method for batch desensitization and format conversion of DICOM files of an echocardiogram, including the following steps.

[0040] Step 100: Acquire a to-be-processed DICOM file.

[0041] Step 200: Desensitize the to-be-processed DICOM file to obtain a desensitized DICOM file.

[0042] Step 300: Extract an ROI from the desensitized DICOM file based on an adaptive mask image and a prior mask image to obtain a DICOM file of interest.

[0043] Step 400: Perform coding and image pixel information processing on the DICOM file of interest based on a DICOM image output format to obtain a DICOM file of interest in a specified storage format.

[0044] Further, the desensitizing the to-be-processed DICOM file to obtain a desensitized DICOM file includes:

[0045] parsing and identifying the to-be-processed DICOM file by using a pydicom library of Python, and replacing sensitive personal information of a patient with a preset character string to obtain an initial desensitized DICOM file;

[0046] reading the initial desensitized DICOM file by using an OpenCV library of Python, and identifying sensitive information to obtain a sensitive region; and

[0047] setting a pixel value of the sensitive region to 0 to obtain the desensitized DICOM file.

[0048] Specifically, batch desensitization processing of DICOM data: in research of intelligent ultrasound images, data desensitization is one of key steps to protect patients' personal privacy. The specific steps are as follows. 1) Use the pydicom library of Python to parse DICOM files collected by different ultrasonic devices, identify personal information (such as name, age, and date of birth) of patients in DICOM Tag files, and replace sensitive information with "XXXX". 2) Use the OpenCV library of Python to read the DICOM files, identify patient-related sensitive information above an image, and set a pixel value in this region to 0 (that is, all black), so as to achieve the purpose of data desensitization and ensure safe sharing of medical data, as shown in FIG. 2.

[0049] Further, the extracting an ROI from the desensitized DICOM file based on an adaptive mask image and a prior mask image to obtain a DICOM file of interest includes:

[0050] acquiring a single-frame DICOM file and a multi-frame DICOM file of the desensitized DICOM file;

[0051] extracting an ROI from the single-frame DICOM file based on the prior mask image to obtain a first DICOM file;

[0052] extracting an ROI from the multi-frame DICOM file based on the adaptive mask image to obtain a second DICOM file; and

[0053] obtaining the DICOM file of interest based on the first DICOM file and the second DICOM file.

[0054] Specifically, with respect to batch extraction of ROIs from an echocardiogram, except for fan-shaped ROIs, a clinically collected echocardiogram contains a lot of useless pixel information therearound. To improve performance of an intelligent analysis model for ultrasound images, in this patent, the pydicom library of Python is used to parse large sample DICOM ultrasound data.

[0055] For single-frame images, ROIs are extracted from single-sector, double-sector or triple-sector DICOM files collected by various ultrasonic devices (with device models including EPIQ 5, EPIQ 7, EPIQ 7C, IE Elite, and iE33) respectively based on prior mask images.

[0056] Further, the extracting the ROI from the multi-frame DICOM file based on the adaptive mask image to obtain the second DICOM file includes:

[0057] acquiring motion information of the multi-frame DICOM file, and obtaining an initial mask image based on the motion information;

[0058] extracting a bounding rectangle of the mask image based on an OpenCV library of Python and constructing the adaptive mask image; and

[0059] extracting the ROI from the multi-frame DICOM file based on the adaptive mask image to obtain the second DICOM file.

[0060] Specifically, for multi-frame images, an initial mask image is obtained based on motion information of two consecutive image frames, then a bounding rectangle of the mask image is extracted by using the OpenCV library of Python, to construct an adaptive mask image (a pixel value in the rectangle region is 255, and the other pixel values are 0), and finally the ROI is extracted from the multi-frame DICOM file based on the adaptive mask image.

[0061] Based on position information of a minimum bounding rectangle in a sector region of the mask image, an original image is cropped to increase a ratio of the ROI to the background, which is helpful to quantitatively analyze the echocardiogram by means of an artificial intelligence algorithm.

[0062] Further, the performing coding and image pixel information processing on the DICOM file of interest based on a DICOM image output format to obtain a DICOM file of interest in a specified format includes:

[0063] parsing the DICOM file of interest by using a pydicom library to obtain color space information of the DICOM file of interest;

[0064] determining whether the color space information is an RGB color space, and if so, acquiring pixel information and bit depth information of the DICOM file of interest; or if not, converting the color space information into a specified color space by using a conversion function and acquiring the pixel information and the bit depth information;

[0065] acquiring a number of frames and a dynamic range of a pixel value of the DICOM file of interest based on the pixel information and the bit depth information;

[0066] determining whether the number of frames of the DICOM file of interest is 1, and if so, obtaining an image storage format of the first DICOM file; or if not, acquiring a frame rate, an image length and an image width of the second DICOM file, and acquiring an image storage format of the second DICOM file based on the bit depth information, and the frame rate, the image length and the image width of the second DICOM file; and

[0067] obtaining the DICOM file of interest with a specified image storage format based on the image storage format of the second DICOM file and the image storage format of the first DICOM file, as shown in FIG. 3.

[0068] Specifically, batch format conversion of DICOM data of an echocardiogram: a DICOM file is parsed by using the pydicom library of Python, and color space information of each piece of data is extracted from (0028,0004) Photometricinterpretation, where (0028,0004) is a Tag of DICOM data, and denotes an address where corresponding data is stored, and Photometricinterpretation is data description, denoting that the position (0028,0004) stores the color space information of the data, which is mainly divided into "MONOCHROME", "YBR_FULL" or "YBRFULL 422", and "RGB", where "MONOCHROME" denotes a single-channel grayscale image; and "YBR FULL" or "YBRFULL 422" denotes a YCbCr (luminance, chrominance-blue, and chrominance-red) color space image. An opencv function, a cv2.cvtColor function, is used to convert the image color space into "RGB". Then pixel information is extracted from (7FE0,0010)PixelData. Adynamic range of a pixel value is extracted from (0028,0100)Bits Allocated, and is generally divided into 8 bits (each pixel is denoted by 8 bits) or 16 bits (each pixel is denoted by 16 bits). Due to the ability of bit depth to affect a subtle change in image representation intensity. Based on bit depth information of an original DICOM file, this patent can save a single-frame image as an image in a specified format (JPG, PNG, BMP or TIFF) having a corresponding bit depth. For multi-frame images, pixel information of each image frame is extracted and stored in an array. In addition, a frame rate is obtained from (0008,2144) Recommended DisplayFrame Rate, and a width and a length of each image are obtained from (0028,0010)Rows and (0028,001 l)Columns respectively. Based on data such as the pixel information, the frame rate, the length, and the width extracted from DICOM files, multi-frame images are encoded into a video with a specified format (AVI, MP4, MOV, or WMV) by using the OpenCV library of Python. (7FE0,0010), (0028,0100), (0008,2144), (0028,0010), and (0028,0011) are all addresses for storing pixel data in DICOM files. PixelData, BitsAllocated, Recommended DisplayFrame Rate, Rows, and Columns are Tag descriptions of the pixel data stored in the DICOM files. In addition, the present disclosure supports resaving of desensitized data as a DICOM format to increase the universality and ease of use of data, as shown in FIG. 4.

[0069] The present disclosure has the following beneficial effects:

[0070] (1) With respect to batch desensitization processing of DICOM data: At present, single-frame or multi-frame DICOM images collected by many ultrasonic devices are faced with certain challenges in desensitization. A conventional technology cannot easily meet requirement for batch desensitization of DICOM files collected by different devices, especially in deleting patient-related sensitive information in DICOM Tag files or images. In order to solve this problem, this patent proposes a batch desensitization processing method, which can efficiently and reliably remove sensitive information to ensure safe sharing of medical data.

[0071] (2) With respect to batch extraction of ROIs from an echocardiogram, except for fan-shaped ROIs, the echocardiogram usually contains a lot of useless pixel information therearound. The information not only occupies a storage space, but also may interfere with performance of an intelligent analysis model. However, conventional software and technologies usually cannot design adaptive mask images for different ultrasonic devices or extract ROIs effectively, which affects accuracy and efficiency of intelligent analysis to some extent. According to this patent, adaptive mask images are innovatively constructed for single-frame and multi-frame DICOM images, and pixel information of RIOs can be adaptively extracted from DICOM images collected by different ultrasonic devices, which is helpful to improve accuracy and efficiency of intelligent analysis.

[0072] (3) With respect to batch format conversion of DICOM data of an echocardiogram, generally, reading DICOM files usually requires use of a special parser. This process is not only complicated, but also reduces ease of use of the DICOM data. In the prior art, DICOM files are mainly converted into single-frame images in JPG format. This patent proposes a method for batch format conversion of DICOM data. By parsing of DICOM files, multi-frame ultrasound images are extracted and encoded into a universal playable file format. Single-frame images can support JPG, PNG, BMP, and TIFF, while multi-frame videos can support AVI, MP4, MOV, and WMV. In addition, this patent supports resaving of desensitized data as a DICOM format. This technology not only greatly simplifies a processing flow of the DICOM data, but also provides more flexible data format selection for wide application of medical images.

[0073] Embodiments of this description are described in a progressive manner, and each embodiment focuses on the difference from other embodiments. For the same and similar parts between the embodiments, reference may be made to each other.

[0074] Specific examples are used herein for illustration of principles and implementations of the present disclosure. The descriptions of the above embodiments are merely used for assisting in understanding the method of the present disclosure and its core ideas. In addition, those of ordinary skill in the art can make changes in terms of specific implementations and the scope of application in accordance with the ideas of the present disclosure. In conclusion, the content of the description shall not be construed as limitations to the present disclosure.

Claims

1. A method for batch desensitization and format conversion of digital imaging and communications in medicine (DICOM) files of an echocardiogram, comprising:acquiring a to-be-processed DICOM file;desensitizing the to-be-processed DICOM file to obtain a desensitized DICOM file;extracting a region of interest (ROI) from the desensitized DICOM file based on an adaptive mask image and a prior mask image to obtain a DICOM file of interest; andperforming coding and image pixel information processing on the DICOM file of interest based on a DICOM image output format to obtain a DICOM file of interest in a specified storage format.

2. The method for batch desensitization and format conversion of DICOM files of an echocardiogram according to claim 1, wherein the desensitizing the to-be-processed DICOM file to obtain a desensitized DICOM file comprises:parsing and identifying the to-be-processed DICOM file by using a pydicom library of Python, and replacing sensitive personal information of a patient with a preset character string to obtain an initial desensitized DICOM file;reading the initial desensitized DICOM file by using an open source computer vision (OpenCV) library of Python, and identifying sensitive information to obtain a sensitive region; andsetting a pixel value of the sensitive region to 0 to obtain the desensitized DICOM file.

3. The method for batch desensitization and format conversion of DICOM files of an echocardiogram according to claim 1, wherein the extracting an ROI from the desensitized DICOM file based on an adaptive mask image and a prior mask image to obtain a DICOM file of interest comprises:acquiring a single-frame DICOM file and a multi-frame DICOM file of the desensitized DICOM file;extracting an ROI from the single-frame DICOM file based on the prior mask image to obtain a first DICOM file;extracting an ROI from the multi-frame DICOM file based on the adaptive mask image to obtain a second DICOM file; andobtaining the DICOM file of interest based on the first DICOM file and the second DICOM file.

4. The method for batch desensitization and format conversion of DICOM files of anechocardiogram according to claim 3, wherein the extracting an ROI from the multi-frame DICOM file based on the adaptive mask image to obtain a second DICOM file comprises:acquiring motion information of the multi-frame DICOM file, and obtaining an initial mask image based on the motion information;extracting a bounding rectangle of the mask image based on an OpenCV library of Python and constructing the adaptive mask image; andextracting the ROI from the multi-frame DICOM file based on the adaptive mask image to obtain the second DICOM file.

5. The method for batch desensitization and format conversion of DICOM files of an echocardiogram according to claim 1, wherein the performing coding and image pixel information processing on the DICOM file of interest based on a DICOM image output format to obtain a DICOM file of interest with a specified storage format comprises:parsing the DICOM file of interest by using a pydicom library to obtain color space information of the DICOM file of interest;determining whether the color space information is a red green blue (RGB) color space, and if so, acquiring pixel information and bit depth information of the DICOM file of interest; or if not, converting the color space information into a specified color space by using a conversion function and acquiring the pixel information and the bit depth information;acquiring a number of frames and a dynamic range of a pixel value of the DICOM file of interest based on the pixel information and the bit depth information;determining whether the number of frames of the DICOM file of interest is 1, and if so, obtaining an image storage format of the first DICOM file; or if not, acquiring a frame rate, an image length and an image width of the second DICOM file, and acquiring an image storage format of the second DICOM file based on the bit depth information, and the frame rate, the image length and the image width of the second DICOM file; andobtaining the DICOM file of interest with a specified image storage format based on the image storage format of the second DICOM file and the image storage format of the first DICOM file.Application No: GB2411595.8Examiner: Mr George MathewsClaims searched: 1-5Date of search: 20 December 2024Patents Act 1977: Search Report under Section 17Documents considered to be relevant:Category Relevant to claims Identity of document and passage or figure of particular relevance Y 1-5 in combinati on with any of document s 4-6 US 11270027 B2 (ROSENBERG et al.) See whole document particularly column 10 lines 4-23 and column 13 line 56 to column 14 line 10 Y 1 and 3-5 in combinati on with any of document s 4-6 CN 115359877 A (CHANGSHA YUNFUTONG TECH CO LTD) See whole document particularly paragraph [0056] Y 1 and 3-5 in combinati on with any of document s 4-6 CN 111145170 B (UNIV ELECTRONIC SCI &TECH CHINA) See whole document particularly paragraphs [0032]-[0034] Y 1 and 3-5 in combinati on with any of document s 1-3, 2 in combinati on with document 1 CN 111915626 A (DALIAN NEUEDU TECH GROUP CO LTD) See whole document particularly paragraphs [0062]-[0068] Y 1 and 3-5 in combinati on with any of CN 111583354 A (SHANGHAI UNITED IMAGING HEALTHCARE CO LTD) See whole document particularly paragraph [0080]document s 1-3,2 in combinati on with document 1 Y 1 and 3-5 CN 113112559 A in (SHENZHEN INST OF ADV TECH CAS) See whole document combinati on with any of document s 1-3, 2 in combinati on with document 1 particularly paragraphs [0133]-[0138]Categories:X Document indicating lack of novelty or inventive step A Document indicating technological background and / or state of the art. Y Document indicating lack of inventive step if P Document published on or after the declared priority date but combined with one or more other documents of same category. before the filing date of this invention. & Member of the same patent family E Patent document published on or after, but with priority date earlier than, the filing date of this application.Field of Search:International Classification:Subclass Subgroup Valid From G06T 0007 / 00 01 / 01 / 2017 G06T 0007 / 10 01 / 01 / 2017 G06T 0007 / 215 01 / 01 / 2017 G06V 0010 / 25 01 / 01 / 2022 G16H 0030 / 20 01 / 01 / 2018 G16H 0030 / 40 01 / 01 / 2018

Images