Intelligent processing system of urological medical images based on three-dimensional reconstruction

Abstract

The present invention relates to the field of medical image processing, and is used to solve the problem that when performing three-dimensional reconstruction of medical images, there is a lack of deep processing of fuzzy areas, resulting in poor display effect of three-dimensional reconstructed models. The present invention is a urology medical image intelligent processing system based on three-dimensional reconstruction, comprising a slice analysis unit, a fuzzy processing unit, a three-dimensional reconstruction unit, a simulation derivation unit and an imaging evaluation unit. In the process of reconstructing the three-dimensional model, the present invention performs regional segmentation, clarity analysis and multi-level block labeling processing on two-dimensional images, classifies and manages the clarity of different positions, and performs corresponding picture simulation correction according to the clarity of different areas, so as to scientifically complete the picture through artificial intelligence, improve the display comprehensiveness and display effect of the three-dimensional reconstructed model, and evaluate the three-dimensional reconstructed model according to the credibility of the picture to ensure the overall credibility of the three-dimensional reconstructed model.

Description

Technical Field

[0001] The present invention relates to the field of medical image processing, and in particular to an intelligent processing system for urology medical images based on three-dimensional reconstruction. Background Art

[0002] Medical image 3D reconstruction technology refers to the technology of using a series of 2D slice images to reconstruct a 3D image model and conduct qualitative and quantitative analysis. With the successful development and use of computer tomography and nuclear magnetic resonance equipment, doctors are provided with clear tomographic image information of rich tissues and organs inside the human body, helping doctors to diagnose diseases more intuitively, quickly and accurately. The existing medical image 3D imaging reconstruction system mainly reconstructs human organ tissues based on 2D sequence images obtained by medical imaging methods such as X-ray, computer tomography CT, nuclear magnetic resonance MR and ultrasound, presenting a 3D stereoscopic appearance.

[0003] At present, the existing patent application CN2013105036946 discloses a technical solution, which simulates and restores the processed two-dimensional data into a three-dimensional model by performing multiple operations on the two-dimensional image data, and can also dynamically set its parameters to change the display effect, thereby improving the three-dimensional degree of the medical image. However, in actual application, due to the complexity of the body structure and organization and the limitations of the equipment itself, the two-dimensional image data used as the basic data may contain noise points, blurred areas, etc. The existing patent application lacks a means of processing blurred areas during three-dimensional reconstruction. Therefore, during three-dimensional reconstruction, if the data with noise points and blurred areas is directly reconstructed into three-dimensional, the three-dimensional model may be biased or incompletely displayed, thereby affecting the display effect of the three-dimensional model.

[0004] In view of the above technical problems, this application proposes a solution. Summary of the invention

[0005] In the process of reconstructing the three-dimensional model, the present invention performs regional segmentation, clarity analysis and multi-level block labeling on the two-dimensional image, classifies and manages the clarity of different positions, and performs corresponding image simulation correction according to the clarity of different areas, so as to scientifically complete the image through artificial intelligence, improve the display comprehensiveness and display effect of the 3D reconstructed model, solve the problem of poor display effect of the 3D reconstructed model due to lack of deep processing of blurred areas when performing three-dimensional reconstruction of medical images, and proposes an intelligent processing system for urological medical images based on three-dimensional reconstruction.

[0006] The purpose of the present invention can be achieved by the following technical solutions:

[0007] The intelligent processing system of urological medical images based on three-dimensional reconstruction includes a slice analysis unit, which collects medical images through an image acquisition device and records the collected images as two-dimensional images. The slice analysis unit analyzes the two-dimensional images to generate a complete image and an image to be optimized;

[0008] A fuzzy processing unit, wherein the fuzzy processing unit obtains the image to be optimized through the slice analysis unit, performs region segmentation and region analysis on the image to be optimized, and marks the image into a complete region and a region to be optimized;

[0009] A three-dimensional reconstruction unit, which combines the two-dimensional images to perform three-dimensional reconstruction, performs conventional reconstruction on the complete area during the three-dimensional reconstruction process, and marks adjacent layers of the area to be optimized;

[0010] A simulation derivation unit, wherein the simulation derivation unit obtains adjacent layer marks through the three-dimensional reconstruction unit, performs a reshaping simulation on the area to be optimized according to the adjacent layer marks, and feeds back to the three-dimensional reconstruction unit;

[0011] An imaging evaluation unit analyzes the three-dimensional reconstruction result and performs integrity analysis based on the three-dimensional reconstruction result.

[0012] As a preferred embodiment of the present invention, after acquiring the two-dimensional image, the slice analysis unit acquires the clarity of the two-dimensional image, and the specific method is:

[0013] The slice analysis unit identifies the shape edges in the two-dimensional image by using a set edge recognition algorithm, and records the shape edges as clear edges and fuzzy edges. The slice analysis unit records the number of pixels of the clear edges and the number of pixels of the fuzzy edges as SH and SL, respectively. The slice analysis unit calculates the clear edge ratio S by using a formula: ;

[0014] The slice analysis unit compares the clear edge ratio with a set threshold value, and marks the two-dimensional image as a complete image or an image to be optimized according to the comparison result.

[0015] As a preferred embodiment of the present invention, the blur processing unit divides the two-dimensional image into i sub-regions according to the area of ​​the two-dimensional image, and counts the clear edges and blurred edges in each sub-region through an edge recognition algorithm, and records them as SHi and SLi respectively, and obtains the regional clarity ratio Si through a formula calculation;

[0016] The fuzzy processing unit compares the region clarity ratio Si with a set ratio threshold, and marks the sub-region as a complete region or a region to be optimized according to the comparison result.

[0017] As a preferred embodiment of the present invention, when the 3D reconstruction unit performs 3D reconstruction, the depth information corresponding to the 2D image is obtained, and the 2D image is arranged according to the depth information, wherein the position of the 2D image is set according to the actual depth information when the 2D image is arranged, and the depth information of the 2D image is the acquisition depth set by the acquisition device when the 2D image is acquired;

[0018] The three-dimensional reconstruction unit performs the first three-dimensional reconstruction on the two-dimensional image after the preliminary arrangement, and records it as an initial three-dimensional image.

[0019] As a preferred embodiment of the present invention, the 3D reconstruction unit records the plane where each 2D image in the initial 3D image is located as a layer, and records each region in the 2D image as a layer region;

[0020] The three-dimensional reconstruction unit performs label processing on each layer area, and the specific processing method is as follows:

[0021] The 3D reconstruction unit records all layer regions corresponding to the 2D image marked as a complete image as normal regions, records the complete region marked as a to-be-optimized image as a normal region, and records the to-be-optimized region as an abnormal region.

[0022] As a preferred embodiment of the present invention, the 3D reconstruction unit selects an abnormal region, and judges the layer regions adjacent to the abnormal region in the longitudinal direction. If the layer regions adjacent to the abnormal region in the longitudinal direction are all normal regions, the adjacent layer regions are simulated and corrected, and the abnormal region is corrected by the adjacent layer regions to improve the clarity of the abnormal region.

[0023] If there are abnormal regions in the layer regions adjacent to the abnormal region in the longitudinal direction, the adjacent abnormal regions are counted and sent to the simulation derivation unit.

[0024] As a preferred embodiment of the present invention, after the simulation derivation unit acquires adjacent abnormal areas, the adjacent abnormal areas are integrated and recorded as abnormal blocks, and the abnormal blocks continue to acquire adjacent layer areas along the longitudinal axis direction. If the adjacent layer areas are all normal areas, the abnormal blocks are simulated and corrected. If there are abnormal areas in the adjacent layer areas, the abnormal areas in the adjacent layer areas are further integrated into the abnormal blocks.

[0025] As a preferred embodiment of the present invention, the three-dimensional reconstruction unit judges the adjacent layer areas of the abnormal area in the same layer. If the adjacent layer areas are all normal areas, the abnormal area is simulated and corrected along the direction of the layer plane. If there is an abnormal area in the adjacent layer area, the abnormal area is integrated through the simulation derivative unit to obtain the abnormal layer area.

[0026] The simulation derivation unit judges the adjacent layer areas of the abnormal layer area. If they are all normal areas, the abnormal layer area is simulated and corrected. If there is an abnormal area, it is further merged into the abnormal layer area.

[0027] As a preferred embodiment of the present invention, after all layer regions are simulated and corrected or merged, the 3D reconstruction unit records the obtained 3D reconstruction model as the final model;

[0028] The imaging evaluation unit performs volume calculation on the number of abnormal regions contained in the abnormal layer and abnormal block in the final model to obtain the abnormal volume, compares the abnormal volume with a set threshold, and generates a reconstructed normal signal or a reconstructed fuzzy signal according to the comparison result.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] 1. In the present invention, in the process of reconstructing the three-dimensional model, the two-dimensional image is segmented into regions, the clarity is analyzed, and multi-level block labeling is performed to classify and manage the clarity of different positions, and the corresponding picture simulation correction is performed according to the clarity of different areas, so that the picture is scientifically supplemented through artificial intelligence, thereby improving the display comprehensiveness and display effect of the three-dimensional reconstructed model.

[0031] 2. In the present invention, when completing the 3D reconstructed image, the credibility of the image during simulation correction is determined by stratifying the blurred areas in the 3D reconstructed image and performing volume statistics in the height direction, and the 3D reconstructed model is evaluated according to the credibility of the image, and the 3D reconstructed model is marked according to the evaluation result, so as to ensure the overall credibility and accuracy of the 3D reconstructed model and avoid the situation in which the volume of the abnormal area during simulation correction is too large, resulting in the distance from the clear image being too far, thereby causing a large anomaly between the 3D reconstructed model and the actual situation. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0033] Figure 1 is a system block diagram of the present invention;

[0034] Figure 2 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0035] The technical scheme of the present invention will be described clearly and completely in conjunction with the embodiments below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] Embodiment 1:

[0037] See also Figure 1 - Figure 2 As shown, the intelligent processing system of urological medical images based on three-dimensional reconstruction includes a slice analysis unit, a fuzzy processing unit, a three-dimensional reconstruction unit, a simulation derivation unit and an imaging evaluation unit;

[0038] The slice analysis unit acquires medical images through an image acquisition device and records the acquired images as two-dimensional images. After acquiring the two-dimensional images, the slice analysis unit acquires the clarity of the two-dimensional images. The specific method is as follows:

[0039] The slice analysis unit identifies the shape edges in the two-dimensional image through the set edge recognition algorithm, and records the shape edges as clear edges and fuzzy edges. The slice analysis unit records the number of pixels of the clear edge and the number of pixels of the fuzzy edge as SH and SL respectively. The slice analysis unit calculates the clear edge ratio S through the formula. ;

[0040] The slice analysis unit compares the clear edge ratio with a set threshold value. If the clear edge ratio is greater than the set threshold value, the two-dimensional image is marked as a complete image. If the clear edge ratio is less than or equal to the set threshold value, the two-dimensional image is marked as an image to be optimized.

[0041] After the fuzzy processing unit obtains the image to be optimized through the slice analysis unit, it divides the image to be optimized into i sub-regions according to the area of ​​the two-dimensional image marked as the image to be optimized, and counts the clear edges and fuzzy edges in each sub-region through the edge recognition algorithm, which are recorded as SHi and SLi respectively. The regional clear ratio Si is calculated by the formula, ;

[0042] The fuzzy processing unit compares the regional clarity ratio Si with the set ratio threshold. If the regional clarity ratio is greater than the set ratio threshold, the sub-region is marked as a complete region. If the regional clarity ratio is less than or equal to the set ratio threshold, the sub-region is marked as a region to be optimized.

[0043] The 3D reconstruction unit combines the 2D images to perform 3D reconstruction. When performing 3D reconstruction, the 3D reconstruction unit obtains depth information corresponding to the 2D images and arranges the 2D images according to the depth information. When arranging, the positions of the 2D images are set according to the actual depth information. The depth information of the 2D images is the acquisition depth set by the acquisition device when the 2D images are acquired.

[0044] The three-dimensional reconstruction unit performs the first three-dimensional reconstruction on the two-dimensional image after the preliminary arrangement, and records it as the initial three-dimensional image;

[0045] The 3D reconstruction unit records the plane where each 2D image in the initial 3D image is located as a layer, and records each sub-region in the 2D image of a layer as a layer region;

[0046] The 3D reconstruction unit performs label processing on each layer area. The specific processing method is as follows:

[0047] The 3D reconstruction unit records all layer regions corresponding to the 2D image marked as a complete image as normal regions, records the complete region in the image marked as to be optimized as normal regions, and records the region to be optimized as an abnormal region.

[0048] The 3D reconstruction unit selects an abnormal region and judges the layer regions adjacent to the abnormal region in the longitudinal direction. If the layer regions adjacent to the abnormal region in the longitudinal direction are all normal regions, the adjacent layer regions are simulated and corrected, and the abnormal region is corrected through the adjacent layer regions to improve the clarity of the abnormal region.

[0049] If there are abnormal areas in the layer area adjacent to the abnormal area in the longitudinal direction, the adjacent abnormal areas are counted and sent to the simulation derivative unit;

[0050] The three-dimensional reconstruction unit judges the adjacent layer areas of the abnormal area in the same layer. If the adjacent layer areas are all normal areas, the abnormal area is simulated and corrected along the direction of the layer plane. If there is an abnormal area in the adjacent layer area, the abnormal area of ​​the adjacent layer area is sent to the simulation derivation unit.

[0051] The simulation derivative unit obtains adjacent layer marks through the three-dimensional reconstruction unit, wherein the adjacent layer marks include adjacent abnormal regions in the longitudinal axis direction and adjacent layer regions in the same layer. After the simulation derivative unit obtains the adjacent abnormal regions, the adjacent abnormal regions are integrated and recorded as abnormal blocks, and the adjacent layer regions are further obtained along the longitudinal axis direction for the abnormal blocks. If the adjacent layer regions are all normal regions, the abnormal blocks are simulated and corrected. If there are abnormal regions in the adjacent layer regions, the abnormal regions in the adjacent layer regions are further integrated into the abnormal blocks.

[0052] The simulation derivation unit integrates the abnormal area to obtain the abnormal layer area. The simulation derivation unit judges the adjacent layer areas of the abnormal layer area. If they are all normal areas, the abnormal layer area is simulated and corrected. If there is an abnormal area, it continues to be merged into the abnormal layer area. After merging the abnormal layer area and the abnormal block and completing the simulation derivation, the simulation derivation unit feeds back to the 3D reconstruction unit.

[0053] Embodiment 2:

[0054] See also Figure 1 - Figure 2 As shown, after all layer regions are simulated and corrected or merged, the 3D reconstruction unit records the obtained 3D reconstruction model as the final model;

[0055] The imaging evaluation unit analyzes the final model of the 3D reconstruction unit. Specifically, the imaging evaluation unit calculates the volume of the abnormal areas contained in the abnormal layers and abnormal blocks in the final model to obtain the abnormal volume, and compares the abnormal volume with the set threshold. If the abnormal volume is less than the set threshold, a normal reconstruction signal is generated. If the abnormal volume is greater than the set threshold, a fuzzy reconstruction signal is generated. The normal reconstruction signal and the fuzzy reconstruction signal are sent to the management platform, and the management personnel understand the credibility of the 3D reconstruction model through the management platform.

[0056] The method for the imaging evaluation unit to calculate the volume is as follows: the imaging evaluation unit selects all abnormal layers and abnormal blocks, and calculates the number of all abnormal areas contained in the abnormal layers and abnormal blocks, and records them as the basic volume X;

[0057] The imaging evaluation unit selects the abnormal layer and abnormal block again, calculates the number of abnormal regions contained in each abnormal block, averages the number of abnormal regions contained in each abnormal block, and records the result of the arithmetic average as the corrected volume Y1, calculates the number of abnormal regions contained in each abnormal layer, averages the number of abnormal regions contained in each abnormal layer, and records the result of the arithmetic average as the corrected volume Y2;

[0058] The imaging evaluation unit calculates the abnormal volume T through the formula, .

[0059] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. An intelligent processing system for urological medical images based on three-dimensional reconstruction, characterized in that: It includes a slice analysis unit, which collects medical images through an image acquisition device and records the collected images as two-dimensional images. The slice analysis unit analyzes the two-dimensional images to generate a complete image and an image to be optimized; A fuzzy processing unit, wherein the fuzzy processing unit obtains the image to be optimized through the slice analysis unit, performs region segmentation and region analysis on the image to be optimized, and marks the image into a complete region and a region to be optimized; A three-dimensional reconstruction unit, which combines the two-dimensional images to perform three-dimensional reconstruction, performs conventional reconstruction on the complete area during the three-dimensional reconstruction process, and marks adjacent layers of the area to be optimized; A simulation derivation unit, wherein the simulation derivation unit obtains adjacent layer marks through the three-dimensional reconstruction unit, performs a reshaping simulation on the area to be optimized according to the adjacent layer marks, and feeds back to the three-dimensional reconstruction unit; An imaging evaluation unit, which analyzes the three-dimensional reconstruction result and performs integrity analysis based on the three-dimensional reconstruction result; The 3D reconstruction unit records the plane where each 2D image in the initial 3D image is located as a layer, and records each region in the 2D image as a layer region; The three-dimensional reconstruction unit performs label processing on each layer area, and the specific processing method is as follows: The 3D reconstruction unit records all layer regions corresponding to the 2D image marked as a complete image as normal regions, records the complete region marked as a to-be-optimized image as a normal region, and records the to-be-optimized region as an abnormal region; The 3D reconstruction unit selects an abnormal region and judges the layer regions adjacent to the abnormal region in the longitudinal direction. If the layer regions adjacent to the abnormal region in the longitudinal direction are all normal regions, the adjacent layer regions are simulated and corrected, and the abnormal region is corrected by the adjacent layer regions to improve the clarity of the abnormal region. If there are abnormal areas in the layer area adjacent to the abnormal area in the longitudinal direction, the adjacent abnormal areas are counted and sent to the simulation derivative unit; After the simulation derivation unit acquires the adjacent abnormal areas, the adjacent abnormal areas are integrated and recorded as abnormal blocks, and the adjacent layer areas are further acquired along the longitudinal axis direction from the abnormal blocks. If the adjacent layer areas are all normal areas, the abnormal blocks are simulated and corrected. If there are abnormal areas in the adjacent layer areas, the abnormal areas in the adjacent layer areas are further integrated into the abnormal blocks. The three-dimensional reconstruction unit judges the adjacent layer areas of the abnormal area in the same layer. If the adjacent layer areas are all normal areas, the abnormal area is simulated and corrected along the layer plane. If there is an abnormal area in the adjacent layer area, the abnormal area is integrated through the simulation derivative unit to obtain the abnormal layer area. The simulation derivation unit judges the adjacent layer areas of the abnormal layer area. If they are all normal areas, the abnormal layer area is simulated and corrected. If an abnormal area exists, it is further merged into the abnormal layer area.

2. The intelligent processing system for urological medical images based on three-dimensional reconstruction according to claim 1 is characterized in that: After acquiring the two-dimensional image, the slice analysis unit acquires the clarity of the two-dimensional image, and the specific method is as follows: The slice analysis unit identifies the shape edges in the two-dimensional image by using a set edge recognition algorithm, and records the shape edges as clear edges and fuzzy edges. The slice analysis unit records the number of pixels of the clear edges and the number of pixels of the fuzzy edges as SH and SL, respectively. The slice analysis unit calculates the clear edge ratio S by using a formula: ; The slice analysis unit compares the clear edge ratio with a set threshold value, and marks the two-dimensional image as a complete image or an image to be optimized according to the comparison result.

3. The intelligent processing system for urological medical images based on three-dimensional reconstruction according to claim 1 is characterized in that: The blur processing unit divides the two-dimensional image into i sub-regions according to the area of ​​the two-dimensional image, and counts the clear edges and blurred edges in each sub-region through an edge recognition algorithm, and records them as SHi and SLi respectively, and obtains the regional clarity ratio Si through a formula calculation; The fuzzy processing unit compares the region clarity ratio Si with a set ratio threshold, and marks the sub-region as a complete region or a region to be optimized according to the comparison result.

4. The intelligent processing system for urological medical images based on three-dimensional reconstruction according to claim 1 is characterized in that: When the 3D reconstruction unit performs 3D reconstruction, it obtains depth information corresponding to the 2D image and arranges the 2D image according to the depth information, wherein the position of the 2D image is set according to the actual depth information when arranging, and the depth information of the 2D image is the acquisition depth set by the acquisition device when the 2D image is acquired; The three-dimensional reconstruction unit performs the first three-dimensional reconstruction on the two-dimensional image after the preliminary arrangement, and records it as an initial three-dimensional image.

5. The intelligent processing system for urological medical images based on three-dimensional reconstruction according to claim 1 is characterized in that: After all the layer regions are simulated and corrected or merged, the 3D reconstruction unit records the obtained 3D reconstruction model as the final model; The imaging evaluation unit performs volume calculation on the number of abnormal regions contained in the abnormal layer and abnormal block in the final model to obtain the abnormal volume, compares the abnormal volume with a set threshold, and generates a reconstructed normal signal or a reconstructed fuzzy signal according to the comparison result.

Images