Feature enhancement system and method for thyroid and breast surgery CT image

Abstract

The invention provides a feature enhancement system and method for a thyroid and breast surgery CT image, and the method comprises the steps: segmenting a target enhancement region of a target thyroid and breast surgery patient through a tissue recognition model, and extracting a target region CT image; a thermal imaging image of the target patient in the thyroid and breast surgery department is collected, and a soft tissue recognition boundary and a focus communication area of a target enhancement area of the target patient in the thyroid and breast surgery department are extracted based on the thermal imaging image; performing soft tissue boundary enhancement on the CT image of the target area according to the soft tissue identification boundary to obtain a CT enhanced image, and extracting image texture features of the CT enhanced image; and performing consistency enhancement on the image texture features of the CT enhanced image according to the focus connected region to obtain a CT visual image of the target patient in the thyroid and breast surgery, and performing image enhancement on the CT image in the thyroid and breast surgery according to the soft tissue identification boundary and the focus connected region of the thermal imaging image of the target patient. And the texture coherence of the CT image in the focus area is improved.

Description

Technical Field

[0001] The present application relates to the technical field of image feature enhancement, and more specifically, to a feature enhancement system and method for thyroid and breast surgery CT images. Background Art

[0002] CT (computed tomography) is a commonly used medical imaging technique that uses X-ray beams to scan the human body layer by layer and reconstruct the images with the help of a computer, producing clear images of internal structures. With its advantages of fast imaging speed and high detail resolution, CT has become widely used in the diagnosis of clinical diseases. In the field of thyroid and breast surgery, CT images can be used to visually demonstrate the tissue structure and lesion characteristics of the breast and thyroid regions, assisting doctors in determining the specific location, morphology, density, and spatial relationship of tumors with surrounding tissues, providing important evidence for differentiating benign and malignant lesions, preoperative evaluation, and postoperative follow-up.

[0003] However, existing technologies mainly rely on doctors' direct observation and empirical judgment of CT images for diagnosis. CT images have certain ambiguity when displaying soft tissue boundaries, and the texture features in key lesion areas are often incoherent and unclear, making identification difficult. This situation not only increases the degree of dependence on doctors' professional experience, but also limits the further development and application of thyroid and breast surgery CT in high-precision diagnosis and intelligent analysis. Summary of the Invention

[0004] The present application provides a feature enhancement system and method for thyroid and breast surgery CT images, which can enhance thyroid and breast surgery CT images based on the soft tissue identification boundaries and lesion connectivity areas of the target patient's thermal imaging image, thereby improving the texture continuity of the CT image in the lesion area.

[0005] In a first aspect, the present application provides a method for feature enhancement of thyroid and breast surgical CT images. The method can be executed by a network device, or by a chip configured in the network device, and the present application does not limit this.

[0006] Specifically, the method includes: Acquiring an initial CT image of a target thyroid and breast surgery patient, and preprocessing the initial CT image; Acquire a preprocessed initial CT image, segment the target enhancement area of ​​the target thyroid and breast surgery patient using a tissue recognition model, and extract a CT image of the target area; Acquiring a thermal imaging image of the target thyroid and breast surgery patient, and extracting a soft tissue identification boundary and a lesion connectivity area of ​​a target enhancement area of ​​the target thyroid and breast surgery patient based on the thermal imaging image; performing soft tissue boundary enhancement on the CT image of the target area according to the soft tissue identification boundary to obtain a CT enhanced image, and extracting image texture features of the CT enhanced image; The image texture features of the CT enhanced image are consistency enhanced according to the lesion connected area to obtain a CT visualization image of the target thyroid and breast surgery patient.

[0007] In conjunction with the first aspect, in certain implementations of the first aspect, extracting the soft tissue identification boundary and the lesion connectivity area of ​​the target enhancement area of ​​the target thyroid and breast surgery patient based on the thermal imaging image specifically includes: Extracting a temperature matrix based on the thermal imaging image to obtain a lesion temperature matrix; Performing gradient feature extraction on the lesion temperature matrix using a temperature difference operator to obtain a lesion temperature gradient map; determining a boundary segmentation threshold based on the temperature gradient distribution of the lesion temperature gradient map, and performing boundary feature extraction on the lesion temperature gradient map based on the boundary segmentation threshold to obtain a soft tissue identification boundary of the target enhancement area; A regional clustering algorithm is used to extract lesion connected regions from the lesion temperature gradient map, and a four-neighborhood analysis algorithm is used to perform feature screening on multiple lesion connected regions, retaining multiple connected regions with an area greater than a set threshold as the final lesion connected regions.

[0008] In conjunction with the first aspect, in certain implementations of the first aspect, performing soft tissue boundary enhancement on the CT image of the target area according to the soft tissue identification boundary to obtain the CT enhanced image specifically includes: Acquiring a soft tissue identification boundary of the thermal imaging image, and mapping the soft tissue identification boundary to a CT image coordinate system of the target area CT image based on spatial registration coordinate parameters to obtain a boundary mask in the CT image coordinate system; Performing local area extraction on the CT image based on the boundary mask, extracting the boundary structure area covered by the mask, obtaining image grayscale features of the boundary structure area, and determining the recognition boundary matching degree according to the image grayscale features of the boundary structure area; Dynamic boundary enhancement is performed on the boundary structure region based on the identified boundary matching degree to obtain the CT enhanced image.

[0009] In combination with the first aspect, in certain implementations of the first aspect, in the process of dynamically enhancing the boundary structure area based on the identified boundary matching degree, rule mapping is performed according to the identified boundary matching degree, corresponding boundary enhancement parameters are determined, and dynamic boundary enhancement operations are performed on the boundary structure area based on the boundary enhancement parameters.

[0010] In conjunction with the first aspect, in certain implementations of the first aspect, performing consistency enhancement on the image texture features of the CT enhanced image according to the lesion connected area to obtain a CT visualization image of thyroid and breast surgery specifically includes: Acquire multiple lesion connected regions in the thermal imaging image, and map the multiple lesion connected regions to the coordinate system of the CT enhanced image based on preset spatial registration parameters to form corresponding multiple connected region masks; In the CT enhanced image, for each connected region mask, image texture features within the connected region mask are extracted, and a connected region matching degree corresponding to the connected region mask is determined based on the image texture features; According to the connected region matching degree corresponding to each connected region mask, the pixel grayscale within each connected region mask is enhanced in consistency to obtain a CT visualization image of thyroid and breast surgery.

[0011] In combination with the first aspect, in certain implementations of the first aspect, a spiral CT scanner is used to acquire an initial CT image of the target thyroid and breast surgery patient.

[0012] In combination with the first aspect, in certain implementations of the first aspect, a convolutional neural network is used as a tissue recognition model to segment the target enhancement area of ​​the target thyroid and breast surgery patient.

[0013] In a second aspect, the present application provides a feature enhancement system for thyroid and breast surgery CT images, which includes an image processing unit, wherein the image processing unit includes: An image acquisition module, configured to acquire an initial CT image of a target thyroid and breast surgery patient and pre-process the initial CT image; An image segmentation module is used to obtain a pre-processed initial CT image, segment the target enhancement area of ​​the target thyroid and breast surgery patient using a tissue recognition model, and extract a CT image of the target area; a feature extraction module, configured to acquire a thermal imaging image of the target thyroid and breast surgery patient, and extract a soft tissue identification boundary and a lesion connectivity area of ​​a target enhancement area of ​​the target thyroid and breast surgery patient based on the thermal imaging image; an image enhancement module, configured to perform soft tissue boundary enhancement on the CT image of the target area according to the soft tissue identification boundary to obtain a CT enhanced image, and extract image texture features of the CT enhanced image; The image enhancement module is further used to perform consistency enhancement on the image texture features of the CT enhanced image according to the lesion connected area to obtain a CT visualization image of the target thyroid and breast surgery patient.

[0014] In a third aspect, the present application provides a computer terminal device, which includes a memory and a processor, wherein the memory stores a code, and the processor is configured to obtain the code and execute the above-mentioned feature enhancement method for thyroid and breast surgical CT images.

[0015] In a fourth aspect, the present application provides a computer-readable storage medium, which stores at least one computer program, and the computer program is loaded and executed by a processor to implement the operations performed by the above-mentioned method for feature enhancement of thyroid and breast surgical CT images.

[0016] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: The present application provides a feature enhancement system and method for thyroid and breast surgery CT images, which first collect an initial CT image of a target thyroid and breast surgery patient and preprocess the initial CT image; obtain the preprocessed initial CT image, use a tissue recognition model to segment the target enhancement area of ​​the target thyroid and breast surgery patient, and extract the CT image of the target area; collect a thermal imaging image of the target thyroid and breast surgery patient, and extract the soft tissue recognition boundary and the lesion connectivity area of ​​the target enhancement area of ​​the target thyroid and breast surgery patient based on the thermal imaging image; enhance the soft tissue boundary of the CT image of the target area according to the soft tissue recognition boundary to obtain a CT enhanced image, and extract the image texture features of the CT enhanced image; and perform consistency enhancement on the image texture features of the CT enhanced image according to the lesion connectivity area to obtain a CT visualization image of the target thyroid and breast surgery patient.

[0017] Therefore, it can be seen that the present application can reflect the metabolic activity and blood flow changes of tissues by capturing the temperature distribution on the surface of tissues through infrared thermal imaging, and adopts the temperature gradient mutation at the boundary to correspond to the natural boundary of tissue structure, and forms a stable connected area through the thermal anomaly associated with the lesion area (such as nodules and masses). The soft tissue boundary and the lesion connected area extracted according to thermal imaging can make up for the problem of unclear grayscale transition of soft tissues and insufficient expression of weak texture areas at the lesion boundary in CT images, and use the thermal imaging boundary to guide CT edge enhancement, improve the grayscale gradient at the junction of soft tissues, and solve the problem of blurred soft tissue boundaries. In the lesion connected area, the thermal map texture consistency constraint is used to ensure the overall coherence of the texture in the area and the uniformity of local details.

[0018] In summary, this application achieves structural guidance and regional consistency enhancement of CT images by introducing soft tissue identification boundaries and lesion connectivity areas in thermal imaging images, which makes up for the inherent weaknesses of CT in soft tissue imaging, significantly improves the texture continuity and visualization expression effect of the lesion area, and thus enhances the image diagnosis accuracy and reliability of thyroid and breast surgical diseases. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is an exemplary flow chart of a method for enhancing features of thyroid and breast surgical CT images according to some embodiments of the present application; Figure 2 is a structural diagram of an image processing unit according to some embodiments of the present application; Figure 3 It is a structural diagram of a computer terminal device for implementing a feature enhancement method for thyroid and breast surgical CT images according to some embodiments of the present application. DETAILED DESCRIPTION

[0020] The present application collects an initial CT image of a target thyroid and breast surgery patient and preprocesses the initial CT image; obtains the preprocessed initial CT image, uses a tissue recognition model to segment the target enhancement area of ​​the target thyroid and breast surgery patient, and extracts the CT image of the target area; collects a thermal imaging image of the target thyroid and breast surgery patient, and extracts the soft tissue recognition boundary and the lesion connectivity area of ​​the target enhancement area of ​​the target thyroid and breast surgery patient based on the thermal imaging image; performs soft tissue boundary enhancement on the CT image of the target area according to the soft tissue recognition boundary to obtain a CT enhanced image, and extracts image texture features of the CT enhanced image; performs consistency enhancement on the image texture features of the CT enhanced image according to the lesion connectivity area to obtain a CT visualization image of the target thyroid and breast surgery patient, and can perform image enhancement on the thyroid and breast surgery CT image according to the soft tissue recognition boundary and the lesion connectivity area of ​​the thermal imaging image of the target patient, thereby improving the texture consistency of the CT image in the lesion area.

[0021] In order to better understand the above technical solution, the following will be combined with the accompanying drawings and specific implementation methods to describe the above technical solution in detail. Figure 1 This figure is an exemplary flow chart of a method for enhancing features of CT images of thyroid and breast surgery according to some embodiments of the present application. The method 100 for enhancing features of CT images of thyroid and breast surgery mainly includes the following steps: In step S101, an initial CT image of a target thyroid and breast surgery patient is acquired and preprocessed.

[0022] Preferably, in some embodiments, a spiral CT scanner is used to acquire the initial CT image of the target thyroid and breast surgery patient. In some other embodiments, other devices or equipment capable of acquiring CT images may also be used, and this application does not limit this.

[0023] Optionally, in some embodiments, the scanning layer thickness of the initial CT image is set to 0.1 mm, the scanning position of the initial CT image is from the mandible to the lower edge of the chest of the thyroid and breast surgery patient, and whether to inject contrast agent for enhancement is determined as needed.

[0024] Optionally, in some embodiments, preprocessing the initial CT image includes performing image denoising and grayscale normalization on the initial CT image to improve image quality and enhance the accuracy of subsequent image processing and fusion.

[0025] In a specific implementation, Gaussian filtering can be used to denoise the initial CT image. By weighted summing each pixel and its neighborhood, this effectively reduces speckle noise and improves image smoothness and edge preservation. In some embodiments, non-local means filtering can also be used to perform detail-preserving denoising on the image, thereby preserving texture information in thyroid and breast tissue.

[0026] Among them, in terms of grayscale normalization in the preprocessing of the initial CT image, the grayscale value of the CT image can be standardized to the range of [0, 255] by a linear normalization method to reduce the impact of image brightness differences between devices; in other embodiments, the Z-score normalization method can also be used to perform standard deviation normalization on the image pixel intensity.

[0027] Furthermore, in some preferred embodiments, the preprocessing also includes spatial registration of the initial CT image and the thermal imaging image, specifically including achieving precise spatial alignment of different modal images through key point matching and transformation matrix calculation based on a rigid affine or non-rigid B-spline registration algorithm, and obtaining corresponding spatial registration parameters, for example, using the transformation matrix of the spatial coordinate system as the spatial registration parameter to ensure that the thermal imaging information can be mapped to the corresponding CT anatomical structure area.

[0028] Among them, in terms of image resolution in the preprocessing of the initial CT image, in order to ensure the uniformity of subsequent model processing, in some embodiments of the present application, the preprocessed CT image can be resized and the pixel spacing of the image can be standardized to a preset voxel spacing (such as 1.0mm×1.0mm×1.0mm), and the image can be reconstructed using bilinear interpolation or cubic spline interpolation to improve the consistency of image processing for different patients.

[0029] In step S102, the pre-processed initial CT image is acquired, the target enhancement area of ​​the target thyroid and breast surgery patient is segmented using a tissue recognition model, and a CT image of the target area is extracted.

[0030] It should be noted that the target enhancement area refers to a specific image area in the thyroid and breast surgery image that is selected for subsequent image processing after recognition model or rule analysis. In specific implementation, a fixed scanning range can be selected as the target enhancement area and image segmentation can be performed based on the scanning position of the initial CT image. After obtaining the pre-processed initial CT image, the target enhancement area of ​​the target thyroid and breast surgery patient can be segmented using a tissue recognition model to extract the corresponding thyroid or breast area CT image.

[0031] In specific implementation, a tissue recognition model can be constructed based on a convolutional neural network structure commonly used in the field of medical image segmentation. Preferably, in some embodiments, the tissue recognition model is a deep segmentation network based on a U-Net architecture, which has skip connections and an encoding-decoding structure, and can effectively identify soft tissue boundaries and lesion areas in CT images.

[0032] When training the tissue recognition model, a professionally annotated thyroid and breast surgical CT image dataset can be used. The training samples contain pixel-level masks of the thyroid, breast and their typical lesion areas, and the model is optimized and trained in a supervised manner. Preferably, a Dice loss function or a cross entropy loss function is introduced during the training process to improve the segmentation accuracy.

[0033] In certain embodiments, in order to enhance the model's adaptability to different patients and image modalities, the tissue recognition model may also introduce an attention mechanism, a multi-scale feature extraction structure (such as an ASPP module), or a multi-branch fusion network structure to improve the model's ability to recognize areas with blurred boundaries and similar density.

[0034] During the actual inference process, the preprocessed initial CT image is input into the trained tissue recognition model, and a binary segmentation image of the target enhanced area is output; the foreground area in the binary image corresponds to anatomical structures related to thyroid and breast surgery (such as thyroid tissue, breast parenchyma or lesions), and the corresponding target area CT image is subsequently extracted from the original CT image based on the mask of this area.

[0035] Optionally, to further improve the segmentation accuracy, in some embodiments, the tissue recognition model can also be used in conjunction with traditional medical image segmentation algorithms based on graph cuts, region growing, etc., to form a combination strategy of coarse segmentation + fine restoration to achieve higher quality image region extraction.

[0036] In step S103, a thermal imaging image of the target thyroid and breast surgery patient is acquired, and the soft tissue recognition boundary and the lesion connected area of ​​the target enhancement area of ​​the target thyroid and breast surgery patient are extracted based on the thermal imaging image.

[0037] It should be noted that infrared thermal imaging, by capturing the temperature distribution on the tissue surface, can reflect tissue metabolic activity and blood flow changes, and possesses highly sensitive temperature contrast characteristics. In thyroid and breast surgery, tissue boundaries often manifest as areas of abrupt temperature gradient changes, while lesions (such as nodules and masses) are accompanied by stable thermal anomalies, forming distinct connected structures. Therefore, thermal imaging can effectively extract soft tissue boundaries and lesion connectivity, compensating for CT imaging's shortcomings in soft tissue grayscale transitions and insufficient lesion texture representation, thereby improving structural visibility and diagnostic accuracy in lesion areas.

[0038] Optionally, in some embodiments, a thermal imaging image of the target breast and thyroid surgery patient is collected by an infrared thermal imaging camera to obtain temperature distribution information of the body surface and shallow tissues, and the target enhancement area is further identified and analyzed in combination with the thermal imaging image. In specific implementation, the resolution of the thermal imaging image is 640×480 or higher, the image grayscale value corresponds to different temperature gradients, and the color mapping can adopt the JET pseudo-color mapping method to enhance the visual expression of temperature differences.

[0039] Optionally, in some embodiments, before extracting the soft tissue identification boundary and lesion connectivity area of ​​the target enhancement area of ​​the target breast and thyroid surgery patient based on the thermal imaging image, the method further includes: spatially registering the initial CT image with the thermal imaging image, specifically based on a rigid affine or non-rigid B-spline registration algorithm, through key point matching and transformation matrix calculation, to achieve precise spatial alignment of the initial CT image and the thermal imaging image, thereby ensuring that the thermal imaging information is mapped to the corresponding CT anatomical structure area.

[0040] Preferably, in some embodiments, extracting the soft tissue identification boundary and the lesion connected area of ​​the target enhancement area of ​​the target thyroid and breast surgery patient based on the thermal imaging image specifically includes: Extracting a temperature matrix based on the thermal imaging image to obtain a lesion temperature matrix; Performing gradient feature extraction on the lesion temperature matrix using a temperature difference operator to obtain a lesion temperature gradient map; determining a boundary segmentation threshold based on the temperature gradient distribution of the lesion temperature gradient map, and performing boundary feature extraction on the lesion temperature gradient map based on the boundary segmentation threshold to obtain a soft tissue identification boundary of the target enhancement area; A regional clustering algorithm is used to extract lesion connected regions from the lesion temperature gradient map, and a four-neighborhood analysis algorithm is used to perform feature screening on multiple lesion connected regions, retaining multiple connected regions with an area greater than a set threshold as the final lesion connected regions.

[0041] In specific implementation, the temperature matrix extraction can perform temperature inversion based on the pixel grayscale value of the thermal imaging image. Through a lookup table or a thermal imaging device calibration model, the grayscale value of each pixel is converted into a corresponding temperature value to construct a two-dimensional lesion temperature matrix to reflect the actual temperature distribution of the relevant areas of thyroid and breast surgery.

[0042] Furthermore, in the temperature gradient feature extraction process, it is preferred to use the Sobel operator or a custom temperature difference convolution kernel to perform local gradient calculation on the lesion temperature matrix, extract the edge features of the temperature mutation area, and construct a lesion temperature gradient map, thereby identifying the soft tissue boundary and the thermal abnormality change area.

[0043] In some embodiments, the boundary segmentation threshold can be determined using an adaptive method, specifically including calculating the global mean, standard deviation, or Gaussian distribution characteristics of the temperature gradient map and setting a dynamic threshold related to local background contrast to enhance compatibility across different patients. In other embodiments, a histogram analysis of the temperature gradient map G(x,y) can be performed, and a boundary segmentation threshold can be set based on the histogram, such as selecting the mean or using the Otsu method for automatic segmentation. Based on this threshold, a set of pixels whose temperature gradients exceed the boundary segmentation threshold is extracted to form a preliminary boundary mask. Alternatively, the gradient map can be binarized using the Otsu automatic thresholding method to extract edge candidate regions. Finally, the coherence and closure of the soft tissue identification boundary are further optimized through edge tracing and morphological operations (such as closing and erosion) to obtain a clearly structured soft tissue identification boundary of the target enhancement region. In a specific implementation, morphological dilation and closing operations are performed on the preliminary boundary mask to remove isolated noise points and broken boundaries. Contour detection (such as based on connected domains or Canny edge tracing) is then combined to extract coherent curves, ultimately generating the soft tissue identification boundary of the target enhancement region.

[0044] It should be noted that the present invention can extract the soft tissue identification boundaries of thyroid and breast surgery from thermal imaging images, mainly based on the significant differences in metabolic activity and blood perfusion of soft tissues, which in turn appear as local temperature mutations in the thermal map, especially at the junction of tissues, forming obvious temperature gradient changes, forming natural thermal dividing lines. By performing gradient operations on the temperature matrix, the temperature change intensity of the edge area can be effectively identified to form a temperature gradient map, and then combined with the boundary segmentation threshold and connected area analysis, the tissue boundary position can be accurately located. This method realizes the stable extraction of soft tissue structure boundaries with physiological significance from the thermal map, providing a reliable structural prior for the subsequent enhancement of CT images.

[0045] In the process of extracting the connected areas of the lesions, K-means, Mean-Shift or density-based spatial clustering algorithms (such as DBSCAN) are used to cluster the temperature gradient map of the lesions based on temperature similarity, and a plurality of clustering areas with local temperature consistency are preliminarily divided.

[0046] Furthermore, the connectivity of the preliminary clustering results is determined based on the four-neighborhood analysis algorithm (or eight-neighborhood analysis algorithm) to identify the connected areas that actually constitute the anatomical structure. In some embodiments, a minimum area threshold (such as greater than 100 pixels or an actual area greater than 5mm) can be set. 2 ), isolated points, small patches and background misjudgment areas are eliminated, and only multiple connected blocks with continuous structure and reasonable morphology are retained as the final target lesion connected area for enhancement.

[0047] Optionally, in some preferred embodiments, the spatial relationship between the soft tissue identification boundary and the lesion connected area can also be logically verified, such as determining whether the boundary is closed and whether the connected area falls within the boundary range, and a topological structure correction algorithm can be introduced to further improve the accuracy of region extraction and the consistency of anatomical structure.

[0048] Preferably, in some embodiments, after extracting the soft tissue identification boundary and the lesion connectivity area of ​​the target enhancement area of ​​the target thyroid and breast surgery patient based on the thermal imaging image, it also includes: mapping the soft tissue identification boundary and the lesion connectivity area to the CT image coordinate system based on the spatial registration relationship between the completed thermal imaging image and the initial CT image.

[0049] In step S104, the soft tissue boundary of the target area CT image is enhanced according to the soft tissue identification boundary to obtain a CT enhanced image, and image texture features of the CT enhanced image are extracted.

[0050] Optionally, in some embodiments, performing soft tissue boundary enhancement on the target area CT image according to the soft tissue identification boundary to obtain the CT enhanced image specifically includes: Acquiring a soft tissue identification boundary of the thermal imaging image, and mapping the soft tissue identification boundary to a CT image coordinate system of the target area CT image based on spatial registration coordinate parameters to obtain a boundary mask in the CT image coordinate system; Performing local area extraction on the CT image based on the boundary mask, extracting the boundary structure area covered by the mask, obtaining image grayscale features of the boundary structure area, and determining the recognition boundary matching degree according to the image grayscale features of the boundary structure area; Dynamic boundary enhancement is performed on the boundary structure region based on the identified boundary matching degree to obtain the CT enhanced image.

[0051] It should be noted that this solution spatially registers thermal images with CT images, mapping the boundaries and lesion regions in the thermal image to the CT image coordinate system to generate a structural constraint mask. The boundary mask is used to limit the local enhancement range to prevent blind processing, while the lesion connectivity mask provides a reference for regional coherence to avoid texture breaks. These masks serve as structural priors, guiding the enhancement process in a targeted manner, thereby improving image enhancement accuracy and structural consistency.

[0052] In a specific implementation, performing soft tissue boundary enhancement on the target area CT image according to the soft tissue identification boundary to obtain a CT enhanced image specifically includes: First, the soft tissue identification boundary extracted from the thermal imaging image is obtained, and based on the spatial registration coordinate parameters established in the early stage, the soft tissue identification boundary is mapped to the CT image coordinate system of the CT image of the target area to obtain the corresponding boundary position and its spatial coordinate contour.

[0053] Furthermore, a boundary mask in the CT image coordinate system is generated based on the mapped soft tissue recognition boundary, and the mask is used to indicate the structure boundary area that should be processed emphatically in the target CT image.

[0054] Next, the boundary mask is used to perform local area extraction on the target area CT image to extract the boundary structure area covered by the mask; on this basis, the image grayscale features of the boundary structure area are obtained, including but not limited to grayscale average value, edge gradient intensity distribution, texture contrast, etc.

[0055] In some embodiments, the recognition boundary matching degree of the boundary structure area is further calculated, which is defined as the overall grayscale gradient average value of the area in the boundary normal (i.e., vertical) direction. The recognition boundary matching degree described in this application is used to measure the degree of alignment between the infrared boundary position and the actual structure boundary in the CT image, and based on the recognition boundary matching degree, the boundary structure area of ​​the target area CT image is dynamically enhanced, thereby strengthening the edge response of the image by enhancing the influence of the directional filter output in the low matching area. It should be noted that the edge response described in this application can be equivalently replaced by the grayscale change rate of the image at the edge of the selected structure area, and this application will not go into details about this.

[0056] Preferably, in some embodiments, in the process of dynamically enhancing the boundary structure region based on the identified boundary matching degree, rule mapping can be performed according to the identified boundary matching degree to determine corresponding boundary enhancement parameters, and the boundary structure region is dynamically enhanced based on the boundary enhancement parameters. In specific implementation, the boundary enhancement parameters can be determined according to the following formula:

[0057] in is the boundary enhancement parameter, is the preset maximum boundary enhancement parameter, exp is the logarithmic function, is the response coefficient, which can be set in the range of 0.05–0.2 based on experience. For the identification boundary matching degree, a directional filter can be used to enhance the edge response in the normal direction based on the boundary enhancement parameter, wherein the boundary enhancement parameter is used to proportionally enhance the edge information of the directional filter. For example, the vertical edge response of the boundary structure area extracted by the Sobel directional filter is R(x, y), and the boundary enhancement parameter is 1.2. The final enhanced image result obtained is an overlapping image of the initial target area CT image and 1.2 times the vertical edge response.

[0058] In step S105, consistency enhancement is performed on the image texture features of the CT enhanced image according to the lesion connected area to obtain a thyroid and breast surgery CT visualization image.

[0059] It should be noted that this application proposes a two-level collaborative enhancement mechanism of "boundary enhancement + texture consistency enhancement", which guides the edge enhancement of CT images through thermal imaging boundaries, effectively improves the grayscale gradient at the junction of soft tissues, improves boundary blur problems, and enhances structural recognition capabilities; at the same time, the consistency characteristics of the thermal map are used in the connected area of ​​the lesion to perform gradient normalization, histogram standardization or low-frequency texture compensation on the weak or discontinuous texture areas in the CT image, so as to achieve overall coherence and detail unification of the texture in the area, thereby ensuring clear image structure and stable and reliable lesion expression.

[0060] Preferably, in some embodiments, performing consistency enhancement on the image texture features of the CT enhanced image according to the lesion connected area to obtain a CT visualization image of thyroid and breast surgery specifically includes: Acquire multiple lesion connected regions in the thermal imaging image, and map the multiple lesion connected regions to the coordinate system of the CT enhanced image based on preset spatial registration parameters to form corresponding multiple connected region masks; In the CT enhanced image, for each connected region mask, image texture features within the connected region mask are extracted, and a connected region matching degree corresponding to the connected region mask is determined based on the image texture features; According to the connected region matching degree corresponding to each connected region mask, the pixel grayscale within each connected region mask is enhanced in consistency to obtain a CT visualization image of thyroid and breast surgery.

[0061] In a specific implementation, first, multiple lesion connected areas in the thermal imaging image are obtained, and based on the established spatial registration parameters (such as affine transformation matrix or nonlinear B-spline grid), the multiple lesion connected areas are respectively mapped to the coordinate system of the CT enhanced image to form corresponding multiple connected area masks for indicating the structural position of the thermal imaging information in the CT image.

[0062] Subsequently, for each connected region mask in the CT-enhanced image, image texture features within the connected region mask are extracted. Preferably, the image texture features include, but are not limited to, the grayscale mean and standard deviation of pixels within the region, gradient histogram distribution, grayscale co-occurrence matrix, etc., and then, based on the image texture features, the connected region matching degree corresponding to the connected region mask is calculated. Preferably, the matching degree is defined as the ratio of the grayscale average gradient change ∇I of the connected region in the infrared thermal imaging image to the grayscale average gradient change ∇CT of the connected region in the CT-enhanced image. The connected region matching degree is used to reflect the image texture clarity of different connected region masks.

[0063] Preferably, in some embodiments, in the process of performing consistency enhancement on the pixel grayscales within each connected region mask according to the connected region matching degree corresponding to each connected region mask, the pixel grayscales within the connected region mask can be determined according to the following formula:

[0064] in, is the grayscale value of the pixel at (x, y) coordinate in the connected region mask after consistency enhancement, is the mean grayscale value of the image in the connected region mask, is the connected area matching degree, To enhance consistency, the grayscale value of the (x, y) coordinate pixel in the connected region mask mentioned above is is the standard deviation of the image grayscale within the connected region mask, is the preset target image grayscale standard deviation.

[0065] This application introduces soft tissue identification boundaries and lesion connectivity areas in thermal imaging images to achieve structural guidance and regional consistency enhancement of CT images, making up for the inherent weaknesses of CT in soft tissue imaging, significantly improving the texture coherence and visualization expression of the lesion area, thereby enhancing the image diagnosis accuracy and reliability of thyroid and breast surgical diseases.

[0066] In addition, in another aspect of the present application, in some embodiments, the present application provides a feature enhancement system for thyroid and breast surgery CT images, the system comprising an image processing unit, referring to Figure 2, which is a schematic diagram of exemplary hardware and / or software structure of an image processing unit according to some embodiments of the present application. The image processing unit 200 includes: an image acquisition module 201, an image segmentation module 202, a feature extraction module 203, and an image enhancement module 204, which are described as follows: An image acquisition module 201 is used to acquire an initial CT image of a target thyroid and breast surgery patient and pre-process the initial CT image; An image segmentation module 202 is configured to obtain a pre-processed initial CT image, segment the target enhancement region of the target thyroid and breast surgery patient using a tissue recognition model, and extract a CT image of the target region; A feature extraction module 203 is configured to acquire a thermal imaging image of the target thyroid and breast surgery patient, and extract a soft tissue identification boundary and a lesion connectivity area of ​​a target enhancement area of ​​the target thyroid and breast surgery patient based on the thermal imaging image; An image enhancement module 204 is configured to perform soft tissue boundary enhancement on the target area CT image according to the soft tissue identification boundary to obtain a CT enhanced image, and extract image texture features of the CT enhanced image; The image enhancement module 204 is further configured to perform consistency enhancement on the image texture features of the enhanced CT image based on the connected lesion region, thereby obtaining a CT visualization image of the target thyroid and breast surgery patient. The above details an example of a system and method for enhancing CT images of thyroid and breast surgery provided in an embodiment of the present application. It is understood that, in order to implement the aforementioned functions, the corresponding apparatus includes hardware structures and / or software modules for performing the respective functions.

[0067] Those skilled in the art should easily realize that, in combination with the units and algorithm steps of each example described in the embodiments disclosed herein, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a certain function in the application is executed in hardware or in a computer software-driven hardware manner depends on the specific application and design constraints of the technical solution. Therefore, professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0068] In addition, the present application also provides a computer terminal device, which includes a memory and a processor, the memory stores code, and the processor is configured to obtain the code and execute the above-mentioned feature enhancement method for thyroid and breast surgical CT images.

[0069] In some embodiments, reference Figure 3, which is a schematic diagram of the structure of a computer terminal device for implementing a method for enhancing the features of CT images of thyroid and breast surgery according to some embodiments of the present application. Figure 3 The computer terminal device 300 shown in FIG. 1 is implemented as shown in FIG. 1 , and the computer terminal device 300 includes at least one communication bus 301 , a communication interface 302 , a processor 303 and a memory 304 .

[0070] The processor 303 may be a general-purpose central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more processors for controlling the execution of a feature enhancement method for thyroid and breast surgical CT images in the present application.

[0071] The communication bus 301 may include a path for transmitting information between the aforementioned components.

[0072] Memory 304 may be, but is not limited to, a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, a random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, an optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), a magnetic disk or other magnetic storage device, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer. Memory 304 may be independent and connected to processor 303 via communication bus 301. Memory 304 may also be integrated with processor 303.

[0073] Memory 304 is used to store program code for executing the solution of the present application, and is controlled by processor 303 for execution. Processor 303 is used to execute the program code stored in memory 304. The program code may include one or more software modules. In the above embodiment, the determination of the soft tissue identification boundary can be implemented by processor 303 and one or more software modules in the program code in memory 304.

[0074] The communication interface 302 uses any transceiver or other device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area network (WLAN), etc.

[0075] Optionally, the computer terminal device 300 may further include a power supply 305 for providing power to various devices or circuits in the real-time computer terminal device.

[0076] In a specific implementation, as an example, a computer terminal device may include multiple processors, each of which may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. A processor herein may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0077] The aforementioned computer terminal device can be a general-purpose computer terminal device or a dedicated computer terminal device. In a specific implementation, the computer terminal device can be a desktop computer, a portable computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. The embodiments of this application do not limit the type of computer terminal device.

[0078] In addition, other aspects of the present application further provide a computer-readable storage medium, which stores at least one computer program, and the computer program is loaded and executed by a processor to implement the operations performed by the above-mentioned method for feature enhancement of thyroid and breast surgical CT images.

[0079] In summary, in a feature enhancement system and method for thyroid and breast surgery CT images disclosed in an embodiment of the present application, the initial CT image of the target thyroid and breast surgery patient is first collected and preprocessed; the preprocessed initial CT image is obtained, and the target enhancement area of ​​the target thyroid and breast surgery patient is segmented using a tissue recognition model to extract the CT image of the target area; the thermal imaging image of the target thyroid and breast surgery patient is collected, and the soft tissue recognition boundary and the lesion connectivity area of ​​the target enhancement area of ​​the target thyroid and breast surgery patient are extracted based on the thermal imaging image; the soft tissue boundary of the target area CT image is enhanced according to the soft tissue recognition boundary to obtain a CT enhanced image, and the image texture features of the CT enhanced image are extracted; the image texture features of the CT enhanced image are consistency enhanced according to the lesion connectivity area to obtain a CT visualization image of the target thyroid and breast surgery patient, and the thyroid and breast surgery CT image can be enhanced according to the soft tissue recognition boundary and the lesion connectivity area of ​​the target patient's thermal imaging image, thereby improving the texture consistency of the CT image in the lesion area.

[0080] The above description is merely an embodiment of the present application. Common knowledge such as the specific technical solutions or features of the solutions is not described in detail herein. It should be noted that those skilled in the art may make various modifications and improvements without departing from the technical solution of the present application, and these modifications and improvements should also be considered within the scope of protection of the present application. These modifications and improvements will not affect the effectiveness of the implementation of the present application or the practical application of the patent.

[0081] The scope of protection claimed by this application shall be determined by the content of the claims. The specific embodiments and other descriptions in the specification may be used to interpret the content of the claims. Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of the invention. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application is intended to include such modifications and variations.

Claims

1. A feature enhancement method for thyroid and breast surgery CT images, characterized in that: include: Acquiring an initial CT image of a target thyroid and breast surgery patient, and preprocessing the initial CT image; Acquire a preprocessed initial CT image, segment the target enhancement area of ​​the target thyroid and breast surgery patient using a tissue recognition model, and extract a CT image of the target area; Acquiring a thermal imaging image of the target thyroid and breast surgery patient, and extracting a soft tissue identification boundary and a lesion connectivity area of ​​a target enhancement area of ​​the target thyroid and breast surgery patient based on the thermal imaging image; performing soft tissue boundary enhancement on the CT image of the target area according to the soft tissue identification boundary to obtain a CT enhanced image, and extracting image texture features of the CT enhanced image; The image texture features of the CT enhanced image are consistency enhanced according to the lesion connected area to obtain a CT visualization image of the target thyroid and breast surgery patient.

2. The method according to claim 1, wherein Extracting the soft tissue identification boundary and the lesion connected area of ​​the target enhancement area of ​​the target thyroid and breast surgery patient based on the thermal imaging image specifically includes: Extracting a temperature matrix based on the thermal imaging image to obtain a lesion temperature matrix; Performing gradient feature extraction on the lesion temperature matrix using a temperature difference operator to obtain a lesion temperature gradient map; determining a boundary segmentation threshold based on the temperature gradient distribution of the lesion temperature gradient map, and performing boundary feature extraction on the lesion temperature gradient map based on the boundary segmentation threshold to obtain a soft tissue identification boundary of the target enhancement area; A regional clustering algorithm is used to extract lesion connected regions from the lesion temperature gradient map, and a four-neighborhood analysis algorithm is used to perform feature screening on multiple lesion connected regions, retaining multiple connected regions with an area greater than a set threshold as the final lesion connected regions.

3. The method according to claim 1, wherein Performing soft tissue boundary enhancement on the target area CT image according to the soft tissue identification boundary to obtain a CT enhanced image specifically includes: Acquiring a soft tissue identification boundary of the thermal imaging image, and mapping the soft tissue identification boundary to a CT image coordinate system of the target area CT image based on spatial registration coordinate parameters to obtain a boundary mask in the CT image coordinate system; Performing local area extraction on the CT image based on the boundary mask, extracting the boundary structure area covered by the mask, obtaining image grayscale features of the boundary structure area, and determining the recognition boundary matching degree according to the image grayscale features of the boundary structure area; Dynamic boundary enhancement is performed on the boundary structure region based on the identified boundary matching degree to obtain the CT enhanced image.

4. The method according to claim 3, wherein In the process of dynamically enhancing the boundary structure region based on the identified boundary matching degree, rule mapping is performed according to the identified boundary matching degree to determine corresponding boundary enhancement parameters, and a dynamic boundary enhancement operation is performed on the boundary structure region based on the boundary enhancement parameters.

5. The method according to claim 1, wherein Consistency enhancement of the image texture features of the CT enhanced image is performed according to the lesion connected area to obtain a CT visualization image of thyroid and breast surgery, specifically comprising: Acquire multiple lesion connected regions in the thermal imaging image, and map the multiple lesion connected regions to the coordinate system of the CT enhanced image based on preset spatial registration parameters to form corresponding multiple connected region masks; In the CT enhanced image, for each connected region mask, image texture features within the connected region mask are extracted, and a connected region matching degree corresponding to the connected region mask is determined based on the image texture features; According to the connected region matching degree corresponding to each connected region mask, the pixel grayscale within each connected region mask is enhanced in consistency to obtain a CT visualization image of thyroid and breast surgery.

6. The method according to claim 1, wherein A spiral CT scanner was used to acquire initial CT images of the target thyroid and breast surgery patients.

7. The method according to claim 1, wherein A convolutional neural network is used as a tissue recognition model to segment the target enhancement area of ​​the target thyroid and breast surgery patient.

8. A feature enhancement system for CT images of thyroid and breast surgery, comprising an image processing unit, wherein the image processing unit is configured to execute the feature enhancement method for CT images of thyroid and breast surgery according to any one of claims 1 to 7, characterized in that: The image processing unit includes: An image acquisition module, configured to acquire an initial CT image of a target thyroid and breast surgery patient and pre-process the initial CT image; An image segmentation module is used to obtain a pre-processed initial CT image, segment the target enhancement area of ​​the target thyroid and breast surgery patient using a tissue recognition model, and extract a CT image of the target area; a feature extraction module, configured to acquire a thermal imaging image of the target thyroid and breast surgery patient, and extract a soft tissue identification boundary and a lesion connectivity area of ​​a target enhancement area of ​​the target thyroid and breast surgery patient based on the thermal imaging image; an image enhancement module, configured to perform soft tissue boundary enhancement on the CT image of the target area according to the soft tissue identification boundary to obtain a CT enhanced image, and extract image texture features of the CT enhanced image; The image enhancement module is further used to perform consistency enhancement on the image texture features of the CT enhanced image according to the lesion connected area to obtain a CT visualization image of the target thyroid and breast surgery patient.

9. A computer terminal device, characterized in that: The computer terminal device includes a memory and a processor, the memory stores codes, and the processor is configured to obtain the codes and execute the feature enhancement method for thyroid and breast surgical CT images according to any one of claims 1 to 7.

10. A computer-readable storage medium storing at least one computer program, characterized in that: The computer program is loaded and executed by a processor to implement the operations performed by the method for feature enhancement of thyroid and breast surgical CT images according to any one of claims 1 to 7.

Images