Image segmentation method and device, computer readable storage medium and electronic device

Abstract

The present application relates to the field of image processing technology, and specifically to an image segmentation method and an image segmentation device, as well as a computer-readable storage medium and an electronic device, which solve the problem of poor segmentation effect of nodular lesions in medical image sequences. The image segmentation method provided in the embodiment of the present application generates a first segmented image corresponding to the image block of the nodular lesion area based on the feature probability map and the initial probability threshold corresponding to the image block of the nodular lesion area, and then generates a second segmented image corresponding to the image block of the nodular lesion area based on the feature probability map, the first segmented image, the nodular volume condition and the preset probability adjustment parameter, so that the probability threshold can be adjusted according to the nodular volume condition and the preset probability adjustment parameter to generate a second segmented image with better segmentation effect for the nodular lesion area, thereby improving the image segmentation effect.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to an image segmentation method and an image segmentation device, as well as a computer-readable storage medium and an electronic device. Background Art

[0002] The nodule lesions in medical image sequences have complex and diverse morphologies, and the sizes of nodule lesions vary greatly. Therefore, it is difficult to segment the nodule lesions in medical image sequences with high accuracy. That is, the existing image segmentation methods have poor effects on segmenting nodule lesions in medical image sequences. Summary of the invention

[0003] In view of this, embodiments of the present application provide an image segmentation method and an image segmentation device, as well as a computer-readable storage medium and an electronic device, which solve the problem of poor segmentation effect of nodule lesions in medical image sequences.

[0004] In a first aspect, an image segmentation method provided by an embodiment of the present application includes: generating a first segmented image corresponding to a nodule lesion area image block based on a feature probability map corresponding to the nodule lesion area image block and an initial probability threshold, wherein the nodule lesion area image block is generated based on a three-dimensional medical image to be segmented containing a nodule lesion area, and the feature probability map includes the nodule segmentation probability of each pixel in the nodule lesion area image block; generating a second segmented image corresponding to the nodule lesion area image block based on the feature probability map, the first segmented image, the nodule volume condition and a preset probability adjustment parameter.

[0005] In combination with the first aspect of the present application, in some embodiments, based on the feature probability map, the first segmented image, the nodule volume condition and the preset probability adjustment parameter, a second segmented image corresponding to the nodule lesion area image block is generated, including: if the first segmented image does not meet the nodule volume condition, based on the initial probability threshold and the preset probability adjustment parameter, determining the adjusted probability threshold; based on the feature probability map and the adjusted probability threshold, regenerating the first segmented image; if the first segmented image meets the nodule volume condition, determining the first segmented image as the second segmented image.

[0006] In combination with the first aspect of the present application, in some embodiments, before generating a first segmented image corresponding to a nodular lesion area image block based on a feature probability map and an initial probability threshold corresponding to the nodular lesion area image block, the method further includes: determining, based on a medical image sequence to be segmented, at least one three-dimensional detection frame area corresponding to the medical image sequence to be segmented; for each three-dimensional detection frame area in at least one three-dimensional detection frame area, determining a sampling grid corresponding to the three-dimensional detection frame area based on the three-dimensional detection frame area; sampling the three-dimensional detection frame area based on the sampling grid to obtain a nodular lesion area image block corresponding to the three-dimensional detection frame area; determining at least one nodular lesion area image block corresponding to the medical image sequence to be segmented based on the nodular lesion area image block corresponding to each of the at least one three-dimensional detection frame areas; wherein, generating a first segmented image corresponding to the nodular lesion area image block based on the feature probability map and the initial probability threshold corresponding to the nodular lesion area image block includes: for each nodular lesion area image block in at least one nodular lesion area image block, generating a first segmented image corresponding to the nodular lesion area image block based on the feature probability map and the initial preset probability threshold corresponding to the nodular lesion area image block.

[0007] In combination with the first aspect of the present application, in some embodiments, based on the three-dimensional detection frame area, a sampling grid corresponding to the three-dimensional detection frame area is determined, including: based on the image volume and physical spatial resolution corresponding to the three-dimensional detection frame area, determining the physical spatial resolution of the sampling grid corresponding to the three-dimensional detection frame area; performing isotropic processing on the physical spatial resolution of the sampling grid to determine the spatial resolution of the sampling grid; based on the three-dimensional detection frame area, the spatial resolution of the sampling grid, and a preset grid size, determining the sampling grid corresponding to the three-dimensional detection frame area.

[0008] In combination with the first aspect of the present application, in some embodiments, based on the medical image sequence to be segmented, at least one three-dimensional detection frame area corresponding to the medical image sequence to be segmented is determined, including: obtaining an A frame image frame of the two-dimensional detection frame area contained in the medical image sequence to be segmented; based on the A frame image frame, determining at least one three-dimensional detection frame area.

[0009] In combination with the first aspect of the present application, in some embodiments, the medical image sequence to be segmented includes multiple frames of image frames with continuous time sequence numbers superimposed and arranged along the Z-axis direction, the two-dimensional detection frame area includes a nodule area, and based on the A-frame image frame, at least one three-dimensional detection frame area is determined, including: based on the A-frame image frame, determining the nodule attribute information of the nodule area in the two-dimensional detection frame area in the A-frame image frame, wherein the nodule attribute information includes the nodule position, nodule size and nodule type; for each image frame in the A-frame image frame, with the current image frame as the center, determining at least one preceding image frame and at least one subsequent image frame of the current image frame; searching among the at least one preceding image frame and the at least one subsequent image frame to determine at least one image frame whose overlap meets a preset overlap threshold; performing clustering processing based on the nodule attribute information of the nodule area in the two-dimensional detection frame area in the A-frame image frame and at least one image frame whose overlap meets the preset overlap threshold, to determine at least one three-dimensional detection frame area.

[0010] In combination with the first aspect of the present application, in some embodiments, before generating a first segmented image corresponding to the nodule lesion area image block based on the feature probability map corresponding to the nodule lesion area image block and the initial probability threshold, it also includes: performing nodule segmentation on the nodule lesion area image block based on a nodule segmentation model to obtain a feature probability map corresponding to the nodule lesion area image block.

[0011] In combination with the first aspect of the present application, in some embodiments, based on the nodule segmentation model, the nodule lesion area image block is subjected to nodule segmentation, and before obtaining the feature probability map corresponding to the nodule lesion area image block, it also includes: sampling the P lesion sampling areas contained in the medical image sequence sample set N times to obtain N combined samples, and training the initial network model based on the N combined samples to obtain the loss results and nodule segmentation models corresponding to the N training times, wherein each combined sample corresponds to a sampling allocation weight, each combined sample includes lesion samples of S types of lesion attributes, the sampling allocation weight is the ratio of the number of lesion samples corresponding to the S types of lesion attributes to the total number of lesion samples included in the combined sample, and S, P, and N are all positive integers; based on the loss results corresponding to the N training times, determine the nodule segmentation model corresponding to the smallest loss result among the N training times.

[0012] In combination with the first aspect of the present application, in some embodiments, based on the nodule segmentation model, the nodule lesion area image block is segmented to obtain a feature probability map corresponding to the nodule lesion area image block, including: based on the nodule segmentation model, the nodule lesion area image block is segmented to obtain a preliminary feature probability map corresponding to the nodule lesion area image block; based on the physical spatial resolution corresponding to the preliminary feature probability map, anisotropic processing is performed to determine the feature probability map corresponding to the nodule lesion area image block.

[0013] In a second aspect, an embodiment of the present application provides an image segmentation device, comprising: a first segmentation module, configured to generate a first segmented image corresponding to a nodule lesion area image block based on a feature probability map corresponding to the nodule lesion area image block and an initial probability threshold, wherein the nodule lesion area image block is generated based on a three-dimensional medical image to be segmented containing a nodule lesion area, and the feature probability map includes the nodule segmentation probability of each pixel in the nodule lesion area image block; a second segmentation module, configured to generate a second segmented image corresponding to the nodule lesion area image block based on the feature probability map, the first segmented image, the nodule volume condition and a preset probability adjustment parameter.

[0014] In a third aspect, an embodiment of the present application provides a computer-readable storage medium, which stores instructions. When the instructions are executed by a processor of an electronic device, the electronic device can execute the image segmentation method mentioned in the first aspect above.

[0015] In a fourth aspect, an embodiment of the present application provides an electronic device, the electronic device comprising: a processor; a memory for storing computer-executable instructions; a processor, for executing computer-executable instructions to implement the image segmentation method mentioned in the first aspect above.

[0016] The image segmentation method provided in the embodiment of the present application generates a first segmented image corresponding to the image block of the nodule lesion area based on the feature probability map and the initial probability threshold corresponding to the image block of the nodule lesion area, and then generates a second segmented image corresponding to the image block of the nodule lesion area based on the feature probability map, the first segmented image, the nodule volume condition and the preset probability adjustment parameter, so that the probability threshold can be adjusted according to the nodule volume condition and the preset probability adjustment parameter to generate a second segmented image with better segmentation effect for the nodule lesion area, thereby improving the image segmentation effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The figure is a schematic diagram of an application scenario of an image segmentation method provided in one embodiment of the present application.

[0018] Figure 2 The figure is a flow chart of an image segmentation method provided in one embodiment of the present application.

[0019] Figure 3 Shown is a flowchart of an image segmentation method provided in another embodiment of the present application.

[0020] Figure 4 Shown is a flowchart of an image segmentation method provided in another embodiment of the present application.

[0021] Figure 5 Shown is a flowchart of an image segmentation method provided in another embodiment of the present application.

[0022] Figure 6 Shown is a flowchart of an image segmentation method provided in another embodiment of the present application.

[0023] Figure 7 Shown is a flowchart of an image segmentation method provided in another embodiment of the present application.

[0024] Figure 8 Shown is a flowchart of an image segmentation method provided in another embodiment of the present application.

[0025] Fig. 9 Shown is a flowchart of an image segmentation method provided in another embodiment of the present application.

[0026] Fig.10 Shown is a flowchart of an image segmentation method provided in another embodiment of the present application.

[0027] Fig.11 Shown is a schematic structural diagram of an image segmentation device provided in one embodiment of the present application.

[0028] Fig.12 Shown is a schematic structural diagram of an image segmentation device provided in another embodiment of the present application.

[0029] Fig.13 Shown is a schematic structural diagram of an image segmentation device provided in another embodiment of the present application.

[0030] Fig.14 Shown is a schematic structural diagram of an image segmentation device provided in another embodiment of the present application.

[0031] Fig.15 Shown is a schematic structural diagram of an image segmentation device provided in another embodiment of the present application.

[0032] Fig.16 Shown is a schematic structural diagram of an image segmentation device provided in another embodiment of the present application.

[0033] Fig.17 Shown is a schematic structural diagram of an image segmentation device provided in another embodiment of the present application.

[0034] Fig.18 Shown is a schematic structural diagram of an image segmentation device provided in another embodiment of the present application.

[0035] Fig.19 Shown is a schematic structural diagram of an image segmentation device provided in another embodiment of the present application.

[0036] Fig. 20 Shown is a schematic structural diagram of an electronic device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0038] Example scenarios

[0039] Figure 1 The figure is a schematic diagram of an application scenario of an image segmentation method provided in one embodiment of the present application. Figure 1 The scene shown includes a server 110 and an image capturing device 120 that is communicatively connected to the server 110. Specifically, the server 110 is used to generate a first segmented image corresponding to the nodular lesion region image block based on a feature probability map corresponding to the nodular lesion region image block and an initial probability threshold, wherein the nodular lesion region image block is generated based on a three-dimensional medical image to be segmented containing a nodular lesion region, and the feature probability map includes a nodule segmentation probability of each pixel in the nodular lesion region image block; based on the feature probability map, the first segmented image, the nodule volume condition and a preset probability adjustment parameter, a second segmented image corresponding to the nodular lesion region image block is generated. The image capturing device 120 is used to obtain the three-dimensional medical image to be segmented, and send the three-dimensional medical image to be segmented to the server 110 so that the server 110 performs the above operations.

[0040] Exemplary Methods

[0041] Figure 2 FIG. 1 is a flow chart of an image segmentation method provided by an embodiment of the present application. Figure 2 As shown, the image segmentation method includes the following steps.

[0042] Step 210: Generate a first segmented image corresponding to the nodule lesion region image block based on the feature probability map corresponding to the nodule lesion region image block and the initial probability threshold.

[0043] Specifically, the nodule lesion region image block is generated based on the three-dimensional medical image to be segmented containing the nodule lesion region. The nodule lesion region image block can be obtained by sampling the nodule lesion region in the three-dimensional medical image to be segmented. The feature probability map includes the nodule segmentation probability of each pixel in the nodule lesion region image block.

[0044] Step 220 , generating a second segmented image corresponding to the nodule lesion area image block based on the feature probability map, the first segmented image, the nodule volume condition and the preset probability adjustment parameter.

[0045] Specifically, the nodule volume condition may be the presence of a nodule lesion region in the first segmented image, or the presence of a nodule lesion region that meets a preset nodule size in the first segmented image. The preset nodule size may be 3mm*3mm*3mm, or 5mm*5mm*5mm, which is not specifically limited in this application. Therefore, it is possible to determine whether to adjust the initial probability threshold based on whether the first segmented image meets the nodule volume condition, thereby obtaining a more appropriate probability threshold.

[0046] The image segmentation method provided in the embodiment of the present application generates a first segmented image corresponding to the image block of the nodule lesion area based on the feature probability map and the initial probability threshold corresponding to the image block of the nodule lesion area, and then generates a second segmented image corresponding to the image block of the nodule lesion area based on the feature probability map, the first segmented image, the nodule volume condition and the preset probability adjustment parameter, so that the probability threshold can be adjusted according to the nodule volume condition and the preset probability adjustment parameter to generate a second segmented image with better segmentation effect for the nodule lesion area, thereby improving the image segmentation effect.

[0047] Figure 3 FIG. 2 is a flow chart of an image segmentation method provided by another embodiment of the present application. Figure 2 Based on the embodiment shown, Figure 3 The embodiment shown is described below in detail. Figure 3 The embodiment shown and Figure 2 The differences and similarities of the illustrated embodiments are not described in detail.

[0048] like Figure 3 As shown, in an embodiment of the present application, the step of generating a second segmented image corresponding to the nodule lesion area image block based on the feature probability map, the first segmented image, the nodule volume condition and the preset probability adjustment parameter includes the following steps.

[0049] Step 310: determine whether the first segmented image meets the nodule volume condition.

[0050] Specifically, if the first segmented image does not meet the nodule volume condition, step 320 and step 330 are performed. If the first segmented image meets the nodule volume condition, step 340 is performed.

[0051] Step 320: Determine an adjusted probability threshold based on the initial probability threshold and a preset probability adjustment parameter.

[0052] Step 330 : regenerate the first segmented image based on the feature probability map and the adjusted probability threshold.

[0053] Step 340: determine the first segmented image as the second segmented image.

[0054] Specifically, step 310 to step 330 may be executed cyclically until the first segmented image meets the nodule volume condition, and then step 340 is executed. The probability threshold corresponding to the first segmented image generated last time may be determined as the actual probability threshold.

[0055] Exemplarily, the initial probability threshold may be 0.8, the preset probability adjustment parameter may be -0.05, the adjusted probability threshold determined after the first adjustment of the probability threshold is 0.75, the adjusted probability threshold determined after the second adjustment of the probability threshold is 0.7, the adjusted probability threshold determined after the third adjustment of the probability threshold is 0.65, and so on.

[0056] By repeatedly determining and adjusting the probability threshold, the actual probability threshold finally obtained can be made more appropriate, thereby making the second segmented image more accurate.

[0057] Figure 4 FIG. 2 is a flow chart of an image segmentation method provided by another embodiment of the present application. Figure 2 Based on the embodiment shown, Figure 4 The embodiment shown is described below in detail. Figure 4 The embodiment shown and Figure 2 The differences and similarities of the illustrated embodiments are not described in detail.

[0058] like Figure 4 As shown, in the embodiment of the present application, before the step of generating the first segmented image corresponding to the nodule lesion area image block based on the feature probability map and the initial probability threshold corresponding to the nodule lesion area image block, the following steps are also included.

[0059] Step 410: Determine at least one three-dimensional detection frame region corresponding to the medical image sequence to be segmented based on the medical image sequence to be segmented.

[0060] Specifically, the medical image sequence to be segmented may include multiple two-dimensional detection frame regions, and at least one three-dimensional detection frame region may be obtained by classifying and fusing the multiple two-dimensional detection frame regions.

[0061] Step 420 : for each 3D detection frame area in the at least one 3D detection frame area, based on the 3D detection frame area, determine a sampling grid corresponding to the 3D detection frame area.

[0062] Specifically, each three-dimensional detection box area corresponds to a sampling grid.

[0063] Step 430 , sampling the three-dimensional detection frame area based on the sampling grid to obtain an image block of the nodule lesion area corresponding to the three-dimensional detection frame area.

[0064] Specifically, a point can be selected in the three-dimensional detection frame area as the sampling center point, and then sampling is performed using a sampling grid with the sampling center point as the center to obtain an image block of the nodule lesion area corresponding to the three-dimensional detection frame area.

[0065] Step 440 : Based on the nodule lesion region image block corresponding to each of the at least one three-dimensional detection frame regions, determine at least one nodule lesion region image block corresponding to the medical image sequence to be segmented.

[0066] In an embodiment of the present application, the step of generating a first segmented image corresponding to a nodule lesion area image block based on a feature probability map and an initial probability threshold corresponding to the nodule lesion area image block comprises the following steps: for each nodule lesion area image block in at least one nodule lesion area image block, generating a first segmented image corresponding to the nodule lesion area image block based on the feature probability map and an initial preset probability threshold corresponding to the nodule lesion area image block.

[0067] Since the three-dimensional detection frame area is likely to include the nodule lesion area, the three-dimensional detection frame area is sampled through a sampling grid to obtain an image block of the nodule lesion area corresponding to the three-dimensional detection frame area, thereby increasing the probability that the sampled nodule lesion area image block contains the nodule lesion area.

[0068] Figure 5 FIG. 2 is a flow chart of an image segmentation method provided by another embodiment of the present application. Figure 4 Based on the embodiment shown, Figure 5 The embodiment shown is described below in detail. Figure 5 The embodiment shown and Figure 4 The differences and similarities of the illustrated embodiments are not described in detail.

[0069] like Figure 5 As shown, in an embodiment of the present application, the step of determining a sampling grid corresponding to a three-dimensional detection frame area based on the three-dimensional detection frame area includes the following steps.

[0070] Step 510 , based on the image volume and the physical space resolution corresponding to the three-dimensional detection frame area, determine the physical space resolution of the sampling grid corresponding to the three-dimensional detection frame area.

[0071] Specifically, the product of the image volume corresponding to the 3D detection frame area and the physical space resolution is the physical volume. The physical volume corresponding to the 3D detection frame area divided by the volume of the sampling grid is the physical space resolution of the sampling grid corresponding to the 3D detection frame area. The volume of the sampling grid can be 64*128*128 or other sizes, which is not specifically limited in this application.

[0072] Step 520: Perform isotropic processing on the physical spatial resolution of the sampling grid to determine the spatial resolution of the sampling grid.

[0073] Specifically, isotropic processing is to adjust the pixel spacing in the three directions of x, y and z to the same spacing. For example, the physical spatial resolution of the sampling grid may be 1*2*2, and the physical spatial resolution of the sampling grid is isotropically processed, and the spatial resolution of the determined sampling grid may be 1*1*1.

[0074] Step 530 , determining a sampling grid corresponding to the three-dimensional detection frame area based on the three-dimensional detection frame area, the spatial resolution of the sampling grid, and a preset grid size.

[0075] By performing isotropic processing on the physical spatial resolution of the sampling grid, the spatial resolution of the sampling grid is made closer to the resolution of the actual nodule lesion, so that the image block of the nodule lesion area obtained by sampling is closer to the actual state of the nodule lesion.

[0076] Figure 6 FIG. 2 is a flow chart of an image segmentation method provided by another embodiment of the present application. Figure 4 Based on the embodiment shown, Figure 6 The embodiment shown is described below in detail. Figure 6 The embodiment shown and Figure 4 The differences and similarities of the illustrated embodiments are not described in detail.

[0077] like Figure 6 As shown, in an embodiment of the present application, based on the medical image sequence to be segmented, the step of determining at least one three-dimensional detection frame area corresponding to the medical image sequence to be segmented includes the following steps.

[0078] Step 610: Acquire an A-frame image frame of a two-dimensional detection frame area contained in the medical image sequence to be segmented.

[0079] Step 620: Determine at least one three-dimensional detection frame area based on the A frame image frame.

[0080] Specifically, at least one three-dimensional detection frame area may be obtained by fusing multiple two-dimensional detection frame areas.

[0081] By fusing multiple two-dimensional detection frame areas into at least one three-dimensional detection frame area, subsequent three-dimensional sampling and image segmentation are facilitated.

[0082] Figure 7 FIG. 2 is a flow chart of an image segmentation method provided by another embodiment of the present application. Figure 6 Based on the embodiment shown, Figure 7 The embodiment shown is described below in detail. Figure 7 The embodiment shown and Figure 6 The differences and similarities of the illustrated embodiments are not described in detail.

[0083] like Figure 7 As shown, in an embodiment of the present application, the step of determining at least one three-dimensional detection frame area based on the A frame image frame includes the following steps.

[0084] Step 710: Based on the A frame image frame, determine the nodule attribute information of the nodule area in the two-dimensional detection frame area in the A frame image frame.

[0085] Specifically, the nodule attribute information includes nodule location, nodule size and nodule type. The medical image sequence to be segmented includes multiple frames of image frames with continuous time sequence numbers that are stacked and arranged along the Z axis. The two-dimensional detection frame area includes the nodule area.

[0086] Step 720 , for each image frame in the A-frame image frames, with the current image frame as the center, determine at least one preceding image frame and at least one succeeding image frame of the current image frame.

[0087] Specifically, the preceding image frame of the current image frame may be an image frame corresponding to a timing sequence number before the timing sequence number corresponding to the current image frame. The succeeding image frame of the current image frame may be an image frame corresponding to a timing sequence number after the timing sequence number corresponding to the current image frame.

[0088] Step 730 , searching at least one preceding image frame and at least one succeeding image frame to determine at least one image frame whose overlap degree satisfies a preset overlap degree threshold.

[0089] Specifically, the preset overlap threshold can be selected according to actual conditions, and this application does not make any specific limitation.

[0090] Step 740, based on the nodule attribute information of the nodule area in the two-dimensional detection frame area in the A frame image frame and at least one image frame whose overlap degree meets a preset overlap degree threshold, clustering processing is performed to determine at least one three-dimensional detection frame area.

[0091] In practical applications, steps 710 to 740 can be implemented through a queue. For example, the candidate queue is first emptied, and then the A-frame image frame is placed in the candidate queue to determine the nodule attribute information of the nodule area in the two-dimensional detection frame area in the A-frame image frame. Then, for each image frame in the A-frame image frame, with the current image frame as the center, at least one preceding image frame and at least one subsequent image frame of the current image frame are determined, and at least one preceding image frame and at least one subsequent image frame are searched to determine at least one image frame whose overlap meets the preset overlap threshold, and at least one image frame that meets the preset overlap threshold is added to the candidate queue. For the candidate queue, the two-dimensional detection frame area of ​​each layer is scanned in the Z direction, and the two-dimensional detection frame area of ​​each layer is clustered according to the nodule category and the degree of aggregation. If there are multiple clusters, it is considered that this candidate queue covers multiple nodule entities, and the queue is split into multiple queues according to different clusters. If there is only one cluster, it is considered that the candidate queue contains only one nodule entity. Finally, all the two-dimensional detection frame areas in the candidate queue are merged, and at least one merged three-dimensional detection frame area is output.

[0092] By searching in at least one preceding image frame and at least one succeeding image frame, at least one image frame whose overlap degree meets a preset overlap degree threshold is determined, and then clustering processing is performed based on the nodule attribute information of the nodule area in the two-dimensional detection frame area in the A frame image frame and at least one image frame whose overlap degree meets the preset overlap degree threshold, so as to determine at least one three-dimensional detection frame area, thereby avoiding missing the three-dimensional detection frame area.

[0093] Figure 8 FIG. 2 is a flow chart of an image segmentation method provided by another embodiment of the present application. Figure 2 Based on the embodiment shown, Figure 8 The embodiment shown is described below in detail. Figure 8 The embodiment shown and Figure 2 The differences and similarities of the illustrated embodiments are not described in detail.

[0094] like Figure 8 As shown, in the embodiment of the present application, before the step of generating the first segmented image corresponding to the nodule lesion area image block based on the feature probability map and the initial probability threshold corresponding to the nodule lesion area image block, the following steps are also included.

[0095] Step 810: Based on the nodule segmentation model, the nodule lesion region image block is segmented to obtain a feature probability map corresponding to the nodule lesion region image block.

[0096] The nodule segmentation model is used to segment the image blocks of the nodule lesion area, which is convenient, fast and efficient.

[0097] Fig. 9FIG. 2 is a flow chart of an image segmentation method provided by another embodiment of the present application. Figure 8 Based on the embodiment shown, Fig. 9 The embodiment shown is described below in detail. Fig. 9 The embodiment shown and Figure 8 The differences and similarities of the illustrated embodiments are not described in detail.

[0098] like Fig. 9 As shown, in an embodiment of the present application, before performing nodule segmentation on the nodule lesion area image block based on the nodule segmentation model and obtaining the feature probability map corresponding to the nodule lesion area image block, the following steps are also included.

[0099] Step 910, sampling the P lesion sampling areas contained in the medical image sequence sample set N times to obtain N combined samples, and training the initial network model based on the N combined samples respectively to obtain the loss results and nodule segmentation models corresponding to the N training times.

[0100] Specifically, the P lesion sampling areas may be areas determined by the detection frame. The area determined by the detection frame may be an area with poor accuracy obtained through preliminary lesion segmentation, that is, the area determined by the detection frame may include the entire nodule or only a part of the nodule. Each combined sample corresponds to a sampling allocation weight, and each combined sample includes lesion samples of S lesion attributes. The sampling allocation weight is the ratio of the number of lesion samples corresponding to each of the S lesion attributes to the total number of lesion samples included in the combined sample. P, N, and S are all positive integers. The initial network model may be a resunet network model. The loss result may be a loss value obtained based on a loss function. The nodule segmentation model may be a model obtained by training the initial network model.

[0101] In one embodiment of the present application, the S types of lesion attributes include: preset nodule size and preset nodule morphology. The preset nodule size can be 5mm*5mm*5mm, or 3mm*3mm*3mm. The preset nodule size can be set according to actual conditions, and this application does not make specific restrictions. The preset nodule morphology can be an irregular shape such as a burr shape, or a regular shape such as a sphere or an ellipse.

[0102] Exemplarily, N may be 3. The three combined samples each contain 1000 lesion samples. The first combined sample may include: 500 lesion samples containing large nodules, and 500 lesion samples containing small nodules. The second combined sample may include: 300 lesion samples containing large nodules, and 700 lesion samples containing small nodules. The third combined sample may include: 800 lesion samples containing large nodules, and 200 lesion samples containing small nodules.

[0103] Step 920: Based on the loss results corresponding to the N training times, determine the nodule segmentation model corresponding to the smallest loss result among the loss results corresponding to the N training times.

[0104] Specifically, the specific value of N can be set according to actual needs. For example, N can be 60. That is, based on the loss results corresponding to each of the 60 trainings, determine the nodule segmentation model corresponding to the smallest loss result among the loss results corresponding to each of the 60 trainings. The specific value of N can also be set according to the convergence of the loss function during the model training process. For example, when the model is trained for the 40th time, the convergence of the loss function has met the convergence condition, then N can be 40. The convergence condition can be a convergence threshold or a convergence speed, which is not specifically limited in this application.

[0105] In practical applications, the nodule segmentation model corresponding to the smallest loss result among the loss results corresponding to N training times can be determined as the final trained nodule segmentation model, or the nodule segmentation model with the best convergence effect of the loss results corresponding to N training times can be determined as the final trained nodule segmentation model.

[0106] By making each combined sample correspond to a sampling allocation weight, the combined samples obtained according to a plurality of different sampling allocation weights can be used to train the initial network model, thereby selecting the nodule segmentation model with the best nodule segmentation effect. In addition, the sampling allocation weight is the ratio of the number of lesion samples corresponding to each of the S lesion attributes to the total number of lesion samples included in the combined sample, thereby balancing the influence of the various lesion attributes on the initial network model training, achieving differentiated sampling, and thus improving the segmentation effect of the nodule segmentation model.

[0107] Fig.10 FIG. 2 is a flow chart of an image segmentation method provided by another embodiment of the present application. Figure 8 Based on the embodiment shown, Fig.10 The embodiment shown is described below in detail. Fig.10 The embodiment shown and Figure 8 The differences and similarities of the illustrated embodiments are not described in detail.

[0108] like Fig.10 As shown, in an embodiment of the present application, based on the nodule segmentation model, the nodule lesion area image block is segmented to obtain a feature probability map corresponding to the nodule lesion area image block, including the following steps.

[0109] Step 1010, based on the nodule segmentation model, perform nodule segmentation on the nodule lesion region image block to obtain a preliminary feature probability map corresponding to the nodule lesion region image block.

[0110] Step 1020, based on the physical space resolution corresponding to the preliminary feature probability map, anisotropic processing is performed to determine the feature probability map corresponding to the image block of the nodule lesion area.

[0111] Combination of the above Figures 1 to 10 , describes the method embodiment of the present application in detail, and the following is combined with Figures 11 to 19 , describes the device embodiment of the present application in detail. It should be understood that the description of the method embodiment corresponds to the description of the device embodiment, so the parts not described in detail can refer to the previous method embodiment.

[0112] Exemplary Devices

[0113] Fig.11 FIG. 1 is a schematic diagram of the structure of an image segmentation device provided by an embodiment of the present application. Fig.11 As shown, the image segmentation device 1100 of the embodiment of the present application includes a first segmentation module 1110 and a second segmentation module 1120 .

[0114] Specifically, the first segmentation module 1100 is configured to generate a first segmented image corresponding to the nodular lesion region image block based on a feature probability map corresponding to the nodular lesion region image block and an initial probability threshold, wherein the nodular lesion region image block is generated based on a three-dimensional medical image to be segmented containing a nodular lesion region, and the feature probability map includes a nodule segmentation probability for each pixel in the nodular lesion region image block. The second segmentation module 1120 is configured to generate a second segmented image corresponding to the nodular lesion region image block based on the feature probability map, the first segmented image, the nodule volume condition, and a preset probability adjustment parameter.

[0115] Fig.12 FIG. 1 is a schematic diagram of the structure of an image segmentation device provided by another embodiment of the present application. Fig.11 Based on the embodiment shown, Fig.12 The embodiment shown is described below in detail. Fig.12 The embodiment shown and Fig.11 The differences and similarities of the illustrated embodiments are not described in detail.

[0116] like Fig.12 As shown, the second segmentation module 1120 of the embodiment of the present application includes: a first judgment result determination unit 1121, a first determination unit 1122 and a second judgment result determination unit 1123.

[0117] Specifically, the first judgment result determination unit 1121 is configured to determine the adjusted probability threshold based on the initial probability threshold and the preset probability adjustment parameter if the first segmented image does not meet the nodule volume condition. The first determination unit 1122 is configured to regenerate the first segmented image based on the feature probability map and the adjusted probability threshold. The second judgment result determination unit 1123 is configured to determine the first segmented image as the second segmented image if the first segmented image meets the nodule volume condition.

[0118] Fig.13 FIG. 1 is a schematic diagram of the structure of an image segmentation device provided by another embodiment of the present application. Fig.11 Based on the embodiment shown, Fig.13 The embodiment shown is described below in detail. Fig.13 The embodiment shown and Fig.11 The differences and similarities of the illustrated embodiments are not described in detail.

[0119] like Fig.13 As shown, the image segmentation device 1100 of the embodiment of the present application further includes: a three-dimensional detection frame area determination module 1130, a sampling grid determination module 1140, a sampling module 1150, and a nodule lesion area image block determination module 1160.

[0120] Specifically, the three-dimensional detection frame area determination module 1130 is configured to determine at least one three-dimensional detection frame area corresponding to the medical image sequence to be segmented based on the medical image sequence to be segmented. The sampling grid determination module 1140 is configured to determine the sampling grid corresponding to the three-dimensional detection frame area based on the three-dimensional detection frame area for each three-dimensional detection frame area in the at least one three-dimensional detection frame area. The sampling module 1150 is configured to sample the three-dimensional detection frame area based on the sampling grid to obtain the nodule lesion area image block corresponding to the three-dimensional detection frame area. The nodule lesion area image block determination module 1160 is configured to determine at least one nodule lesion area image block corresponding to the medical image sequence to be segmented based on the nodule lesion area image block corresponding to each of the at least one three-dimensional detection frame areas.

[0121] In one embodiment of the present application, the first segmentation module 1100 is further configured to generate a first segmented image corresponding to the nodule lesion area image block for each nodule lesion area image block in at least one nodule lesion area image block based on the feature probability map corresponding to the nodule lesion area image block and an initial preset probability threshold.

[0122] Fig.14 FIG. 1 is a schematic diagram of the structure of an image segmentation device provided by another embodiment of the present application. Fig.13 Based on the embodiment shown, Fig.14 The embodiment shown is described below in detail. Fig.14 The embodiment shown and Fig.13 The differences and similarities of the illustrated embodiments are not described in detail.

[0123] like Fig.14 As shown, the sampling grid determination module 1140 of the embodiment of the present application includes: a sampling grid physical space resolution determination unit 1141, an isotropy processing unit 1142 and a sampling grid determination unit 1143.

[0124] Specifically, the physical spatial resolution determination unit 1141 of the sampling grid is configured to determine the physical spatial resolution of the sampling grid corresponding to the three-dimensional detection frame area based on the image volume and the physical spatial resolution corresponding to the three-dimensional detection frame area. The isotropic processing unit 1142 is configured to perform isotropic processing on the physical spatial resolution of the sampling grid to determine the spatial resolution of the sampling grid. The sampling grid determination unit 1143 is configured to determine the sampling grid corresponding to the three-dimensional detection frame area based on the three-dimensional detection frame area, the spatial resolution of the sampling grid, and the preset grid size.

[0125] Fig.15 FIG. 1 is a schematic diagram of the structure of an image segmentation device provided by another embodiment of the present application. Fig.13 Based on the embodiment shown, Fig.15 The embodiment shown is described below in detail. Fig.15 The embodiment shown and Fig.13 The differences and similarities of the illustrated embodiments are not described in detail.

[0126] like Fig.15 As shown, the three-dimensional detection frame area determination module 1130 of the embodiment of the present application includes: an A frame image acquisition unit 1131 and a three-dimensional detection frame area determination unit 1132.

[0127] Specifically, the A frame image acquisition unit 1131 is configured to acquire an A frame image frame of a 2D detection frame region contained in the medical image sequence to be segmented. The 3D detection frame region determination unit 1132 is configured to determine at least one 3D detection frame region based on the A frame image frame.

[0128] Fig.16 FIG. 1 is a schematic diagram of the structure of an image segmentation device provided by another embodiment of the present application. Fig.15 Based on the embodiment shown, Fig.16 The embodiment shown is described below in detail. Fig.16 The embodiment shown and Fig.15 The differences and similarities of the illustrated embodiments are not described in detail.

[0129] like Fig.16As shown, the 3D detection frame area determination unit 1132 of the embodiment of the present application includes: a nodule attribute determination subunit 1610 , a preceding image and a succeeding image determination subunit 1620 , a retrieval subunit 1630 and a clustering processing subunit 1640 .

[0130] Specifically, the sequence of medical images to be segmented includes multiple frames of image frames with continuous time sequence numbers that are superimposed and arranged along the Z-axis direction, and the two-dimensional detection frame area includes a nodule area. The nodule attribute determination subunit 1610 is configured to determine the nodule attribute information of the nodule area in the two-dimensional detection frame area in the A-frame image frame based on the A-frame image frame, wherein the nodule attribute information includes the nodule position, nodule size and nodule type. The previous image and subsequent image determination subunit 1620 is configured to determine, for each image frame in the A-frame image frame, at least one previous image frame and at least one subsequent image frame of the current image frame with the current image frame as the center. The retrieval subunit 1630 is configured to search in at least one previous image frame and at least one subsequent image frame to determine at least one image frame whose overlap meets a preset overlap threshold. The clustering processing subunit 1640 is configured to perform clustering processing based on the nodule attribute information of the nodule area in the two-dimensional detection frame area in the A frame image frame and at least one image frame whose overlap meets a preset overlap threshold, to determine at least one three-dimensional detection frame area.

[0131] Fig.17 FIG. 1 is a schematic diagram of the structure of an image segmentation device provided by another embodiment of the present application. Fig.11 Based on the embodiment shown, Fig.17 The embodiment shown is described below in detail. Fig.17 The embodiment shown and Fig.11 The differences and similarities of the illustrated embodiments are not described in detail.

[0132] like Fig.17 As shown, the image segmentation device 1100 of the embodiment of the present application further includes: a nodule segmentation module 1170 .

[0133] Specifically, the nodule segmentation module 1170 is configured to perform nodule segmentation on the nodule lesion area image block based on the nodule segmentation model to obtain a feature probability map corresponding to the nodule lesion area image block.

[0134] Fig.18 FIG. 1 is a schematic diagram of the structure of an image segmentation device provided by another embodiment of the present application. Fig.17 Based on the embodiment shown, Fig.18 The embodiment shown is described below in detail. Fig.18 The embodiment shown and Fig.17 The differences and similarities of the illustrated embodiments are not described in detail.

[0135] like Fig.18 As shown, the image segmentation device 1100 of the embodiment of the present application further includes: a model training module 1180 and a model determination module 1190.

[0136] Specifically, the model training module 1180 is configured to sample the P lesion sampling areas contained in the medical image sequence sample set N times to obtain N combined samples, and train the initial network model based on the N combined samples to obtain the loss results and nodule segmentation models corresponding to the N trainings, wherein each combined sample corresponds to a sampling allocation weight, each combined sample includes lesion samples of S lesion attributes, and the sampling allocation weight is the ratio of the number of lesion samples corresponding to the S lesion attributes to the total number of lesion samples included in the combined sample, and N is a positive integer. The model determination module 1190 is configured to determine the nodule segmentation model corresponding to the minimum loss result among the N trainings based on the loss results corresponding to the N trainings.

[0137] Fig.19 FIG. 1 is a schematic diagram of the structure of an image segmentation device provided by another embodiment of the present application. Fig.17 Based on the embodiment shown, Fig.19 The embodiment shown is described below in detail. Fig.19 The embodiment shown and Fig.17 The differences and similarities of the illustrated embodiments are not described in detail.

[0138] like Fig.19 As shown, the nodule segmentation module 1170 of the embodiment of the present application includes: a preliminary feature probability map determining unit 1171 and a feature probability map determining unit 1172 .

[0139] Specifically, the preliminary feature probability map determining unit 1171 is configured to perform nodule segmentation on the nodule lesion region image block based on the nodule segmentation model to obtain a preliminary feature probability map corresponding to the nodule lesion region image block. The feature probability map determining unit 1172 is configured to perform anisotropic processing based on the physical space resolution corresponding to the preliminary feature probability map to determine the feature probability map corresponding to the nodule lesion region image block.

[0140] Exemplary Electronic Devices

[0141] Fig. 20 FIG. 1 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present application. Fig. 20 As shown, the electronic device 2000 includes: one or more processors 2001 and a memory 2002; and computer program instructions stored in the memory 2002, and when the computer program instructions are executed by the processor 2001, the processor 2001 executes the image segmentation method as any of the above embodiments.

[0142] The processor 2001 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions.

[0143] The memory 2002 may include one or more computer program products, and the computer program products may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 2001 may execute the program instructions to implement the steps in the image segmentation method of each embodiment of the present application described above and / or other desired functions.

[0144] In one example, the electronic device 2000 may further include: an input device 2003 and an output device 2004, these components are connected via a bus system and / or other forms of connection mechanisms ( Fig. 20 ) interconnected.

[0145] In addition, the input device 2003 may also include, for example, a keyboard, a mouse, a microphone, and the like.

[0146] The output device 2004 can output various information to the outside, and can include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto.

[0147] Of course, to simplify, Fig. 20 Only some of the components related to the present application in the electronic device 2000 are shown, and components such as a bus, an input device / output interface, etc. are omitted. In addition, according to specific application conditions, the electronic device 2000 may also include any other appropriate components.

[0148] Exemplary computer-readable storage media

[0149] In addition to the above methods and devices, the embodiments of the present application may also be a computer program product, including computer program instructions, which, when executed by a processor, enable the processor to execute the steps in the image segmentation method of any of the above embodiments.

[0150] The computer program product may be written in any combination of one or more programming languages ​​to write program codes for performing the operations of the embodiments of the present application, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" language or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0151] In addition, an embodiment of the present application may also be a computer-readable storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the processor executes the steps of the image segmentation method according to various embodiments of the present application described in the above "Exemplary Method" section of this specification.

[0152] Computer readable storage media can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0153] The basic principles of the present application are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present application are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present application. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, not for limitation, and the above details do not limit the present application to being implemented by adopting the above specific details.

[0154] The block diagrams of the devices, apparatuses, equipment, and systems involved in this application are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and / or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The words "such as" used here refer to the phrase "such as but not limited to", and can be used interchangeably with them.

[0155] It should also be noted that in the apparatus, device and method of the present application, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0156] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0157] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

[0158] The above are only preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent substitutions, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. An image segmentation method, characterized in that: include: Based on a feature probability map corresponding to a nodule lesion region image block and an initial probability threshold, a first segmented image corresponding to the nodule lesion region image block is generated, wherein the nodule lesion region image block is generated based on a three-dimensional medical image to be segmented containing a nodule lesion region, and the feature probability map includes a nodule segmentation probability of each pixel in the nodule lesion region image block; Based on the feature probability map, the first segmented image, the nodule volume condition and the preset probability adjustment parameter, generating a second segmented image corresponding to the nodule lesion area image block; The step of generating a second segmented image corresponding to the nodule lesion area image block based on the feature probability map, the first segmented image, the nodule volume condition and the preset probability adjustment parameter comprises: The initial probability threshold is adjusted according to the nodule volume condition and a preset probability adjustment parameter, and a second segmented image is generated according to the adjusted probability threshold, the feature probability map and the first segmented image.

2. The image segmentation method according to claim 1, characterized in that: The step of generating a second segmented image corresponding to the nodule lesion area image block based on the feature probability map, the first segmented image, the nodule volume condition and the preset probability adjustment parameter comprises: If the first segmented image does not satisfy the nodule volume condition, determining an adjusted probability threshold based on the initial probability threshold and the preset probability adjustment parameter; regenerating the first segmented image based on the feature probability map and the adjusted probability threshold; If the first segmented image meets the nodule volume condition, the first segmented image is determined as the second segmented image.

3. The image segmentation method according to claim 1 or 2, characterized in that: Before generating the first segmented image corresponding to the nodular lesion area image block based on the feature probability map corresponding to the nodular lesion area image block and the initial probability threshold, the method further includes: Based on the sequence of medical images to be segmented, determining at least one three-dimensional detection frame area corresponding to the sequence of medical images to be segmented; For each three-dimensional detection frame area in the at least one three-dimensional detection frame area, Based on the three-dimensional detection frame area, determining a sampling grid corresponding to the three-dimensional detection frame area; Sampling the three-dimensional detection frame area based on the sampling grid to obtain a nodule lesion area image block corresponding to the three-dimensional detection frame area; Based on the nodule lesion region image blocks corresponding to each of the at least one three-dimensional detection frame regions, determining at least one nodule lesion region image block corresponding to the medical image sequence to be segmented; The step of generating a first segmented image corresponding to the nodule lesion region image block based on the nodule lesion region image block, the feature probability map corresponding to the nodule lesion region image block and the initial probability threshold comprises: For each nodule lesion area image block in the at least one nodule lesion area image block, a first segmented image corresponding to the nodule lesion area image block is generated based on a feature probability map corresponding to the nodule lesion area image block and an initial preset probability threshold.

4. The image segmentation method according to claim 3, characterized in that: The determining, based on the three-dimensional detection frame area, a sampling grid corresponding to the three-dimensional detection frame area includes: Determining a physical spatial resolution of a sampling grid corresponding to the three-dimensional detection frame area based on an image volume and a physical spatial resolution corresponding to the three-dimensional detection frame area; Performing isotropic processing on the physical spatial resolution of the sampling grid to determine the spatial resolution of the sampling grid; A sampling grid corresponding to the three-dimensional detection frame area is determined based on the three-dimensional detection frame area, the spatial resolution of the sampling grid, and a preset grid size.

5. The image segmentation method according to claim 3, characterized in that: The step of determining at least one three-dimensional detection frame region corresponding to the medical image sequence to be segmented based on the medical image sequence to be segmented includes: Acquire an A-frame image frame of a two-dimensional detection frame area contained in the medical image sequence to be segmented; Based on the A frame image frame, determine the at least one three-dimensional detection frame area.

6. The image segmentation method according to claim 5, characterized in that: The sequence of medical images to be segmented includes a plurality of image frames with consecutive time sequence numbers superimposed and arranged along the Z-axis direction, the two-dimensional detection frame area includes a nodule area, and determining the at least one three-dimensional detection frame area based on the A-frame image frame includes: Based on the A-frame image frame, determining nodule attribute information of a nodule area in a two-dimensional detection frame area in the A-frame image frame, wherein the nodule attribute information includes a nodule position, a nodule size, and a nodule type; For each image frame in the A-frame image frames, Taking the current image frame as the center, determining at least one preceding image frame and at least one succeeding image frame of the current image frame; Searching among the at least one preceding image frame and the at least one succeeding image frame to determine at least one image frame whose overlap degree satisfies a preset overlap degree threshold; Based on the nodule attribute information of the nodule area in the two-dimensional detection frame area in the A frame image frame and at least one image frame whose overlap degree meets a preset overlap degree threshold, clustering processing is performed to determine the at least one three-dimensional detection frame area.

7. The image segmentation method according to claim 1 or 2, characterized in that: Before generating the first segmented image corresponding to the nodular lesion area image block based on the feature probability map corresponding to the nodular lesion area image block and the initial probability threshold, the method further includes: Based on the nodule segmentation model, the nodule lesion area image block is segmented to obtain a feature probability map corresponding to the nodule lesion area image block.

8. The image segmentation method according to claim 7, characterized in that: Before performing nodule segmentation on the nodule lesion region image block based on the nodule segmentation model to obtain a feature probability map corresponding to the nodule lesion region image block, the method further includes: Perform N sampling on P lesion sampling areas included in the medical image sequence sample set to obtain N combined samples, and train the initial network model based on the N combined samples to obtain the loss results and nodule segmentation models corresponding to the N trainings, wherein each of the combined samples corresponds to a sampling allocation weight, each of the combined samples includes lesion samples with S lesion attributes, and the sampling allocation weight is the ratio of the number of lesion samples corresponding to each of the S lesion attributes to the total number of lesion samples included in the combined sample, and S, P, and N are all positive integers; Based on the loss results corresponding to the N training times, a nodule segmentation model corresponding to the smallest loss result among the loss results corresponding to the N training times is determined.

9. The image segmentation method according to claim 7, characterized in that: The step of performing nodule segmentation on the nodule lesion region image block based on the nodule segmentation model to obtain a feature probability map corresponding to the nodule lesion region image block includes: Based on the nodule segmentation model, performing the nodule segmentation on the nodule lesion area image block to obtain a preliminary feature probability map corresponding to the nodule lesion area image block; Based on the physical space resolution corresponding to the preliminary feature probability map, anisotropic processing is performed to determine the feature probability map corresponding to the image block of the nodule lesion area.

10. An image segmentation device, characterized in that: include: A first segmentation module is configured to generate a first segmented image corresponding to the nodule lesion region image block based on a feature probability map corresponding to the nodule lesion region image block and an initial probability threshold, wherein the nodule lesion region image block is generated based on a three-dimensional medical image to be segmented containing a nodule lesion region, and the feature probability map includes a nodule segmentation probability of each pixel in the nodule lesion region image block; The second segmentation module is configured to generate a second segmented image corresponding to the nodule lesion area image block based on the characteristic probability map, the first segmented image, the nodule volume condition and the preset probability adjustment parameter; the generating the second segmented image corresponding to the nodule lesion area image block based on the characteristic probability map, the first segmented image, the nodule volume condition and the preset probability adjustment parameter includes: adjusting the initial probability threshold according to the nodule volume condition and the preset probability adjustment parameter, and generating the second segmented image according to the adjusted probability threshold, the characteristic probability map and the first segmented image.

11. A computer-readable storage medium, characterized in that: The storage medium stores instructions, and when the instructions are executed by a processor of an electronic device, the electronic device is enabled to execute the image segmentation method described in any one of claims 1 to 9.

12. An electronic device, characterized in that: The electronic device comprises: processor; memory for storing computer executable instructions; The processor is used to execute the computer executable instructions to implement the image segmentation method according to any one of claims 1 to 9.

Images