Sound spectrum image processing method and system for ultrasonic diagnosis

Abstract

The invention discloses a sound spectrum image processing method for ultrasonic diagnosis. The method comprises five steps of sound spectrum image acquisition, image preprocessing, feature extraction, intelligent analysis and result output. The method comprises the following steps: firstly, acquiring sound spectrum image data through a diasonograph, denoising through median filtering or wavelet transform, and enhancing image quality by using histogram equalization or gamma correction; then, sound spectrum features such as frequency, amplitude and phase are extracted through a signal processing algorithm, and shape features of the lesion area are extracted through morphological operation; thirdly, analyzing the feature data by adopting a machine learning algorithm, automatically identifying and classifying lesion areas, and comparing a result with standard data to verify accuracy; and finally, displaying an analysis result in a graph and text form for reference of a doctor. According to the method, the image definition, the processing efficiency and the diagnosis accuracy are remarkably improved through efficient automatic processing and an optimization algorithm, and a user-friendly interface is configured, so that a doctor can quickly interpret and operate conveniently.

Description

Technical Field

[0001] The present application relates to the technical field of medical image processing, and in particular to a method and system for processing sonographic images for ultrasonic diagnosis. Background Art

[0002] Ultrasonic diagnostic equipment is a widely used device for medical examinations. It forms images through the propagation and reflection of sound waves in the body to help doctors make diagnoses. Spectral images are important information carriers in ultrasonic diagnosis, which can reflect the acoustic characteristics of tissues and pathological conditions. However, traditional spectral image processing methods have the following problems:

[0003] 1. Poor image quality: Due to noise and interference, the clarity and readability of the spectrum image are poor.

[0004] 2. Low processing efficiency: Existing image processing methods are mostly manual, time-consuming and error-prone.

[0005] 3. Lack of intelligent analysis: Traditional methods lack the ability to intelligently analyze spectral images and are unable to automatically identify and classify lesions.

[0006] Therefore, a new type of acoustic spectrum image processing method is needed to improve the accuracy and efficiency of ultrasonic diagnosis. Summary of the invention

[0007] In view of the shortcomings of the prior art, the present invention aims to provide a method and system for processing acoustic spectrum images for ultrasonic diagnosis, so as to solve the problems raised in the above-mentioned background technology.

[0008] According to one aspect of the present application, a method for processing an ultrasonic spectrum image for ultrasonic diagnosis comprises the following steps:

[0009] Step S1, acoustic spectrum image acquisition: acquiring acoustic spectrum image data through an ultrasonic diagnostic instrument;

[0010] Step S2: Image preprocessing:

[0011] Step S21, denoising: using median filtering or wavelet transform algorithm to remove noise in the spectrum image;

[0012] Step S22, image enhancement: using histogram equalization or gamma correction method to enhance the contrast and brightness of the spectrum image;

[0013] Step S3: feature extraction:

[0014] Step S31, extracting sound spectrum features: extracting the frequency, amplitude and phase features of the sound spectrum image through a signal processing algorithm;

[0015] Step S32, morphological processing: using morphological operations to extract shape features of the lesion area in the spectrum image;

[0016] Step S4: Intelligent analysis:

[0017] Step S41: Analyze the extracted features using a machine learning algorithm to automatically identify and classify the lesion area in the acoustic spectrum image;

[0018] Step S42, result evaluation: comparing the analysis results with standard data to verify accuracy;

[0019] Step S5, result output: the processing results are displayed in the form of graphics and text for the doctor's reference.

[0020] Preferably, the wavelet transform algorithm in step S21 adopts adaptive threshold denoising to retain key features of the spectrum image.

[0021] Preferably, the gamma correction method in step S22 dynamically adjusts the gamma value according to the brightness distribution of the image to optimize the visual effect of the image.

[0022] Preferably, the morphological processing comprises the following steps:

[0023] Dilation operation: expanding the edge of the lesion area to fill possible gaps;

[0024] Erosion operation: Remove noise points to improve the clarity of the lesion area outline.

[0025] Preferably, the machine learning algorithm in step S4 uses a convolutional neural network (CNN) or a support vector machine (SVM), and improves the model's ability to recognize specific lesions through a specially designed training data set.

[0026] Preferably, step S5 supports the following functions:

[0027] Multiple visualization forms, including heat maps, curve graphs, and 3D models;

[0028] User interaction function allows doctors to annotate and correct recognition results.

[0029] A sonogram image processing system for ultrasonic diagnosis, comprising:

[0030] An image acquisition module is used to acquire acoustic spectrum image data in real time through an ultrasonic diagnostic instrument;

[0031] An image processing module, including a denoising unit, an image enhancement unit and a feature extraction unit;

[0032] Intelligent analysis module, used for machine learning analysis and result evaluation of spectrum images;

[0033] The user interface module is used to display the processing results of the spectrum image and support user interaction.

[0034] Preferably, the image enhancement unit in the image processing module further comprises:

[0035] The local histogram equalization submodule is used to optimize the contrast of the local area of ​​the spectrum image;

[0036] The adaptive gamma correction submodule is used to adjust the gamma value according to the dynamic range of the spectrum image.

[0037] Preferably, the intelligent analysis module further comprises:

[0038] A feature matching unit, used to compare the extracted sound spectrum features with a preset database;

[0039] Model training unit, used to dynamically update machine learning models based on new data to improve analysis accuracy.

[0040] Preferably, the user interface module supports real-time voice input and control to improve the doctor's operating efficiency.

[0041] Compared with the prior art, the sound spectrum image processing method and system of the present invention have the following advantages:

[0042] 1. Image quality improvement: Through denoising and enhancement processing, the clarity and readability of the spectrum image are effectively improved.

[0043] 2. Improved processing efficiency: The automated processing process greatly reduces manual intervention and significantly improves efficiency.

[0044] 3. Enhanced diagnostic accuracy: The use of machine learning algorithms enhances the ability to identify lesion areas and reduces the misdiagnosis rate.

[0045] 4. User-friendliness: The interactive user interface enables doctors to quickly understand the processing results and perform further operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 This is a workflow diagram of a method for processing sonographic images for ultrasonic diagnosis according to an embodiment of the present application.

[0047] Figure 2 It is a schematic diagram of the composition of an acoustic spectrum image processing system for ultrasonic diagnosis according to an embodiment of the present application. DETAILED DESCRIPTION

[0048] In order to make the contents of the present application easier to understand, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0049] like Figure 1 As shown, a method for processing an ultrasonic spectrum image for ultrasonic diagnosis comprises the following steps:

[0050] Step S1, Acquisition of Spectrum Image: Acquiring spectrum image data through an ultrasonic diagnostic apparatus. The spectrum image data contains spectrum information of the lesion area and can provide basic data support for subsequent processing.

[0051] Step S2: Image preprocessing:

[0052] Step S21, denoising: using median filtering or wavelet transform algorithm to denoise the spectrum image; the wavelet transform algorithm uses adaptive threshold denoising to retain key features in the spectrum image and ensure the integrity of the lesion area.

[0053] Step S22, image enhancement: enhance the image using histogram equalization or gamma correction method; dynamically adjust gamma correction parameters, optimize visual effects according to the brightness distribution of the spectrum image, and improve the contrast and brightness of the lesion area.

[0054] Step S3: feature extraction:

[0055] Step S31, spectrum feature extraction: extract the frequency, amplitude and phase features of the spectrum image through signal processing algorithm to provide key parameters for intelligent analysis.

[0056] Step S32, morphological processing: using morphological operations to extract shape features of the lesion area, including:

[0057] Dilation operation: expand the edge of the lesion area and fill in any gaps that may exist;

[0058] Erosion operation: eliminate noise points and improve the clarity of the contour of the lesion area.

[0059] Step S4: Intelligent analysis:

[0060] Step S41, lesion area identification and classification: Use convolutional neural network (CNN) or support vector machine (SVM) to analyze the extracted features, automatically identify and classify the lesion area; through a specially designed training data set, improve the model's ability to recognize specific lesions.

[0061] Step S42, result evaluation: compare the analysis results with standard data, verify their accuracy and output a credibility score.

[0062] Step S5: Result output: Display the processing results in the form of graphics and text, including:

[0063] Multiple visualization forms: heat maps, curve graphs, and 3D models.

[0064] User interaction function: allows doctors to mark and correct the recognition results to further optimize the diagnosis effect.

[0065] Combination Figure 2 , an acoustic spectrum image processing system for ultrasonic diagnosis, including an image acquisition module, an image processing module, an intelligent analysis module and a user interface module.

[0066] The image acquisition module is used to collect acoustic spectrum image data in real time through an ultrasonic diagnostic instrument and supports multiple ultrasonic imaging modes (such as color Doppler, spectral Doppler, etc.).

[0067] The image processing module includes a denoising unit, an image enhancement unit and a feature extraction unit; specifically, the denoising unit is used to implement denoising processing on the spectrum image, and adopts a median filter or a wavelet transform algorithm to retain key features; the image enhancement unit includes a local histogram equalization submodule and an adaptive gamma correction submodule, the local histogram equalization submodule optimizes the contrast of the local area of ​​the spectrum image, and the adaptive gamma correction submodule adjusts the gamma value according to the dynamic range of the spectrum image to improve the visual effect of the image; the feature extraction unit is used to extract spectrum features such as frequency, amplitude, and phase, and obtain the shape features of the lesion area through morphological operations.

[0068] The intelligent analysis module is used to perform machine learning analysis and result evaluation on the acoustic spectrum image, including a feature matching unit and a model training unit; specifically, the feature matching unit is used to compare the extracted acoustic spectrum features with the preset database to assist in diagnosis; the model training unit is used to dynamically update the machine learning model according to the newly added data to improve the recognition accuracy of specific lesions.

[0069] The user interface module is used to display the processing results of the spectrum image and support user interaction. Specific functions include: real-time display of heat maps, curve graphs and 3D models; support doctors to complete diagnosis annotations through voice input and control to improve operation efficiency; provide an interactive interface to allow doctors to correct the analysis results and store the corrected data for subsequent learning.

[0070] The above embodiments are only used to illustrate the technical solutions of the embodiments of the present application, rather than to limit them. Although the embodiments of the present application are described in detail with reference to the above embodiments, it should be understood by those skilled in the art that the technical solutions described in the above embodiments can still be modified, or some of the technical features can be replaced by equivalents, without departing from the spirit and scope defined by the claims of the present application.

Claims

1. A method for processing an ultrasonic spectrum image for ultrasonic diagnosis, characterized in that: The following steps are involved: Step S1, acoustic spectrum image acquisition: acquiring acoustic spectrum image data through an ultrasonic diagnostic instrument; Step S2: Image preprocessing: Step S21, denoising: using median filtering or wavelet transform algorithm to remove noise in the spectrum image; Step S22, image enhancement: using histogram equalization or gamma correction method to enhance the contrast and brightness of the spectrum image; Step S3: feature extraction: Step S31, extracting sound spectrum features: extracting the frequency, amplitude and phase features of the sound spectrum image through a signal processing algorithm; Step S32, morphological processing: using morphological operations to extract shape features of the lesion area in the spectrum image; Step S4: Intelligent analysis: Step S41: Analyze the extracted features using a machine learning algorithm to automatically identify and classify the lesion area in the acoustic spectrum image; Step S42, result evaluation: comparing the analysis results with standard data to verify accuracy; Step S5, result output: the processing results are displayed in the form of graphics and text for the doctor's reference.

2. The method for processing ultrasonic spectrum images for ultrasonic diagnosis according to claim 1, characterized in that: The wavelet transform algorithm in step S21 adopts adaptive threshold denoising to retain the key features of the spectrum image.

3. The method for processing ultrasonic spectrum images for ultrasonic diagnosis according to claim 1, characterized in that: The gamma correction method in step S22 dynamically adjusts the gamma value according to the brightness distribution of the image to optimize the visual effect of the image.

4. The method for processing ultrasonic spectrum images for ultrasonic diagnosis according to claim 1, characterized in that: The morphological processing comprises the following steps: Dilation operation: expanding the edge of the lesion area to fill possible gaps; Erosion operation: Remove noise points to improve the clarity of the lesion area outline.

5. The method for processing ultrasound images for ultrasonic diagnosis according to claim 1, characterized in that: The machine learning algorithm in step S4 uses a convolutional neural network (CNN) or a support vector machine (SVM), and improves the model's ability to recognize specific lesions through a specially designed training data set.

6. The method for processing ultrasonic spectrum images for ultrasonic diagnosis according to claim 1, characterized in that: The step S5 supports the following functions: Multiple visualization forms, including heat maps, curve graphs, and 3D models; User interaction function allows doctors to annotate and correct recognition results.

7. A system for implementing the sound spectrum image processing method according to claim 1, characterized in that: include: An image acquisition module is used to acquire acoustic spectrum image data in real time through an ultrasonic diagnostic instrument; An image processing module, including a denoising unit, an image enhancement unit and a feature extraction unit; Intelligent analysis module, used for machine learning analysis and result evaluation of spectrum images; The user interface module is used to display the processing results of the spectrum image and support user interaction.

8. The ultrasound image processing system for ultrasonic diagnosis according to claim 7, characterized in that: The image enhancement unit in the image processing module further comprises: The local histogram equalization submodule is used to optimize the contrast of the local area of ​​the spectrum image; The adaptive gamma correction submodule is used to adjust the gamma value according to the dynamic range of the spectrum image.

9. The ultrasound image processing system for ultrasonic diagnosis according to claim 7, characterized in that: The intelligent analysis module further comprises: A feature matching unit, used to compare the extracted sound spectrum features with a preset database; Model training unit, used to dynamically update machine learning models based on new data to improve analysis accuracy.

10. The ultrasound image processing system for ultrasonic diagnosis according to claim 7, characterized in that: The user interface module supports real-time voice input and control to improve the doctor's operating efficiency.

Images