Human body back spinous process line extraction method, equipment, medium and product

Abstract

The invention discloses a human body back spinous process line extraction method and device, a medium and a product, and relates to the field of image detection, and the method comprises the steps: obtaining a target human body back RGB-D image and a target human body back RGB image; preprocessing depth data in the target human body back RGB-D image to obtain a processed target human body back RGB-D image; according to the target human body back RGB image, extracting a target human body back frame by using a human body back region extraction model; and according to the target human body back frame and the processed target human body back RGB-D image, extracting a target human body back spinous process line by using a spinous process line extraction model. According to the method, the back spinous process line is extracted based on the RGB-D image of the back of the human body, manual extraction or X-ray film shooting is avoided, the accuracy of extraction of the spinous process line is improved, and radiation to the body is reduced.

Description

Technical Field

[0001] The present application relates to the field of image detection, and in particular to a method, device, medium and product for extracting spinous process lines on the back of a human body. Background Art

[0002] Adolescent idiopathic scoliosis (AIS) is an axial rotation or lateral deviation of the spine that occurs in children around puberty. Traditional methods for extracting spinous process lines on the back of the human body are divided into manual detection and X-ray detection. The manual detection method requires the subject to be in a sitting or side-lying position, combined with neck flexion, pelvic lifting and other actions, and confirm the protrusion position as the spinous process through finger palpation. The manual detection method is cumbersome and inconvenient, and highly dependent on the professionalism of the tester, so the test results are less accurate, and there will be a large fluctuation between the results obtained by different testers; the X-ray detection method is to directly mark the spinous process position on the X-ray film. This method is costly, and the shooting process will bring radiation, which is potentially harmful to younger children. Therefore, they are often only suitable for clinical diagnosis, preoperative planning and postoperative follow-up, and are difficult to apply to large-scale spinous process line extraction. Summary of the invention

[0003] The purpose of this application is to provide a method, device, medium and product for extracting the spinous process line on the back of the human body, so as to improve the accuracy of extracting the spinous process line on the back of the human body.

[0004] To achieve the above objectives, this application provides the following solutions:

[0005] In a first aspect, the present application provides a method for extracting a spinous process line on the back of a human body, comprising:

[0006] Obtain an RGB-D image of the target human back and an RGB image of the target human back;

[0007] Preprocessing the depth data in the target human back RGB-D image to obtain a processed target human back RGB-D image; the preprocessing includes spatial smoothness constraints and depth value constraints;

[0008] According to the target human back RGB image, a target human back border is extracted using a human back region extraction model; wherein the human back region extraction model is obtained by training a YOLO-pose model using a first training data set; the first training data set includes a training human back RGB image and a corresponding human back border label;

[0009] According to the target human back border and the processed target human back RGB-D image, the target human back spinous process lines are extracted using a spinous process line extraction model; wherein the spinous process line extraction model includes a spinous process line extraction sub-model and a spinous process line starting and ending point detection model; the spinous process line extraction sub-model is obtained by training the HRNet network using the second training data set; the second training data set includes a training human back RGB-D image with a human back border marked and the corresponding initial spinous process line labels; the spinous process line starting and ending point detection model is obtained by training the HRNet network using the third training data set; the third training data set includes a training human back RGB-D image with a human back border marked and the corresponding spinous process line starting and ending point labels.

[0010] Optionally, preprocessing the depth data in the target human back RGB-D image to obtain a processed target human back RGB-D image specifically includes:

[0011] Performing a spatial smoothing constraint on the depth data in the RGB-D image of the back of the target human body to obtain depth data after a first constraint;

[0012] Performing a depth value constraint on the depth data in the RGB-D image of the back of the target human body to obtain depth data after a second constraint;

[0013] An optimized depth value is determined according to the first constrained depth data and the second constrained depth data, thereby obtaining a processed RGB-D image of the back of a target human body.

[0014] Optionally, performing a spatial smoothing constraint on the depth data in the target human back RGB-D image to obtain the depth data after the first constraint specifically includes:

[0015] Using the formula The depth data in the RGB-D image of the back of the target human body is spatially smoothed to obtain the depth data after the first constraint; wherein, E s (i, j) is the depth data after the first constraint; p(i, j) is the coordinate of the pixel point in the RGB-D image of the back of the target human body; (u x ,u y ) is the principal point coordinate of the camera; f x is the focal length of the camera in the x direction; f y is the focal length of the camera in the y direction; D(i,j) is the depth value corresponding to the pixel.

[0016] Optionally, performing a depth value constraint on the depth data in the target human back RGB-D image to obtain the depth data after the second constraint specifically includes:

[0017] Using Formula E p (i,j)=[D(i,j)-D i (i,j)] 2 →0 performs a depth value constraint on the depth data in the RGB-D image of the back of the target human body to obtain the depth data after the second constraint; wherein, E p (i, j) is the depth data after the second constraint; D(i, j) is the depth value corresponding to the pixel.

[0018] Optionally, determining an optimized depth value according to the first constrained depth data and the second constrained depth data specifically includes:

[0019] Using the formula E(D) = ∑ (i,j ) s E s (i,j)+w p E p (i, j) determining an optimized depth value according to the first constrained depth data and the second constrained depth data; wherein E(D) is the optimized depth value; w s is the weight of the spatial smoothness constraint; E s (i, j) is the depth data after the first constraint; w p is the weight of the depth value constraint; E p (i, j) is the depth data after the second constraint.

[0020] Optionally, the YOLO-pose model is trained using the first training data set to obtain the human back region extraction model, specifically including:

[0021] Obtain a first training data set;

[0022] The training human back RGB image in the first training data set is used as input, and the corresponding human back border label is used as output to train the YOLO-pose model to obtain the human back area extraction model.

[0023] Optionally, according to the target human back border and the processed target human back RGB-D image, using a spinous process line extraction model, extracting the target human back spinous process lines specifically includes:

[0024] According to the target human back border and the processed target human back RGB-D image, using the spinous process line extraction sub-model, determine the target human back initial spinous process line;

[0025] According to the target human back border and the processed target human back RGB-D image, using the spinous process line starting and ending point detection model, determine the starting and ending points of the target human back spinous process line;

[0026] The target human back spinous process line is determined according to the target human back initial spinous process line and the starting and ending points of the target human back spinous process line.

[0027] In a second aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the above-described methods for extracting human dorsal spinous process lines.

[0028] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for extracting human dorsal spinous process lines.

[0029] In a fourth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned methods for extracting human dorsal spinous process lines.

[0030] According to the specific embodiments provided in this application, this application has the following technical effects:

[0031] The present application provides a method, device, medium and product for extracting spinous process lines on the back of a human body, which obtains an RGB-D image of the back of a target human body and an RGB image of the back of a target human body; pre-processes the depth data in the RGB-D image of the back of a target human body to obtain a processed RGB-D image of the back of a target human body; extracts a frame of the back of a target human body using a human back region extraction model based on the RGB image of the back of a target human body; extracts spinous process lines on the back of a target human body using a spinous process line extraction model based on the frame of the back of a target human body and the processed RGB-D image of the back of a target human body. The present application extracts spinous process lines on the back based on the RGB-D image of the back of a human body, avoids manual extraction or taking X-rays, improves the accuracy of spinous process line extraction, and reduces radiation to the body. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0033] Figure 1A schematic diagram of a flow chart of a method for extracting spinous process lines on the back of a human body provided in one embodiment of the present application;

[0034] Figure 2 Flowchart for spinous process line extraction provided for this application;

[0035] Figure 3 Schematic diagram of the spinous process line extraction results provided for this application;

[0036] Figure 4 A schematic diagram of the structure of a computer 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] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0039] In an exemplary embodiment, Figure 1 As shown, a method for extracting spinous process lines on the back of a human body is provided, comprising the following steps:

[0040] S1: Obtain the target human back RGB-D image and the target human back RGB image. In practical applications, a Kinect DK camera is used to obtain the human back RGB-D image.

[0041] S2: preprocessing the depth data in the target human back RGB-D image to obtain a processed target human back RGB-D image; the preprocessing includes spatial smoothness constraints and depth value constraints.

[0042] In practical applications, the resolution of the RGB-D image of the human back is unified through the camera intrinsic parameters, and the spatial smoothness constraint and depth value constraint are introduced to optimize the depth data in the image.

[0043] As an optional implementation, S2 specifically includes:

[0044] S21: Performing spatial smoothing constraints on the depth data in the RGB-D image of the back of the target human body to obtain depth data after the first constraint.

[0045] As an optional implementation, S21 specifically includes:

[0046] Using the formula The depth data in the RGB-D image of the back of the target human body is spatially smoothed to obtain the depth data after the first constraint; wherein, E s (i, j) is the depth data after the first constraint; p(i, j) is the coordinate of the pixel point in the RGB-D image of the back of the target human body, which is calculated by the camera intrinsic parameters; (u x ,u y ) is the principal point coordinate of the camera; f x is the focal length of the camera in the x direction; f y is the focal length of the camera in the y direction; D(i,j) is the depth value corresponding to the pixel.

[0047] S22: performing a depth value constraint on the depth data in the target human back RGB-D image to obtain depth data after a second constraint. The depth value constraint ensures that the optimized depth value is close to the original depth value.

[0048] As an optional implementation, S22 specifically includes:

[0049] Using Formula E p (i,j)=[D(i,j)-D i (i,j)] 2 →0 performs a depth value constraint on the depth data in the RGB-D image of the back of the target human body to obtain the depth data after the second constraint; wherein, E p (i, j) is the depth data after the second constraint; D(i, j) is the depth value corresponding to the pixel.

[0050] S23: Determine an optimized depth value according to the first constrained depth data and the second constrained depth data, so as to obtain a processed RGB-D image of the back of the target human body.

[0051] As an optional implementation manner, determining an optimized depth value according to the first constrained depth data and the second constrained depth data specifically includes:

[0052] Using the formula E(D) = ∑ (i,j) w s E s (i,j)+w p E p (i, j) determining an optimized depth value according to the first constrained depth data and the second constrained depth data; wherein E(D) is the optimized depth value; w s is the weight of the spatial smoothness constraint; E s (i, j) is the depth data after the first constraint; wp is the weight of the depth value constraint; E p (i, j) is the depth data after the second constraint.

[0053] S3: According to the target human back RGB image, use the human back area extraction model to extract the target human back border; wherein the human back area extraction model is obtained by training the YOLO-pose model using the first training data set; the first training data set includes the training human back RGB image and the corresponding human back border label.

[0054] As an optional implementation manner, the YOLO-pose model is trained using the first training data set to obtain the human back area extraction model, specifically including:

[0055] Step 1: Get the first training dataset.

[0056] In practical applications, the first training data set is obtained by the shooting system, which consists of a KinectDK camera, a control module, and an ultraviolet light source. First, a specific potion is used to mark all the back boundary key points on the surface of the human back, so that they are visible only under ultraviolet light; then the control module is programmed to coordinate the camera and the ultraviolet light source, and the first shot is taken when the ultraviolet light source is turned off to obtain an unlabeled RGB image of the human back; then the ultraviolet light source is turned on and the second shot is taken to obtain an RGB image of the human back with the back boundary key points marked. The interval between the two shots controlled by the control module is extremely short, so the two images can be regarded as aligned.

[0057] In addition, a specific potion is used to mark all the key points of the spinous processes on the surface of the human back so that they are visible only under ultraviolet light; then the control module is programmed to coordinate the control of the camera and the ultraviolet light source, turn on the ultraviolet light source, and take pictures to obtain an RGB-D image of the human back with the key points of the spinous processes marked.

[0058] Step 2: Taking the training human back RGB image in the first training data set as input and the corresponding human back border label as output, the YOLO-pose model is trained to obtain the human back area extraction model.

[0059] In practical applications, the back area of ​​the human body is extracted from the RGB image of the back through YOLO-pose. Specifically: feature extraction is implemented through the pre-trained Dark-Net53 network (the pre-trained Dark-Net53 network is the backbone network of the YOLO-pose model), and the extracted features are mapped to the coordinates of the key points of the human body using a fully convolutional network; the corresponding back bounding box of the human body is obtained through the coordinates of the six key points of the hip, shoulder and ear, and the back area of ​​the human body is extracted.

[0060] S4: According to the target human back border and the processed target human back RGB-D image, the target human back spinous process lines are extracted using a spinous process line extraction model; wherein the spinous process line extraction model includes a spinous process line extraction sub-model and a spinous process line starting and ending point detection model; the spinous process line extraction sub-model is obtained by training the HRNet network using the second training data set; the second training data set includes a training human back RGB-D image with a human back border marked and the corresponding initial spinous process line labels; the spinous process line starting and ending point detection model is obtained by training the HRNet network using the third training data set; the third training data set includes a training human back RGB-D image with a human back border marked and the corresponding spinous process line starting and ending point labels.

[0061] As an optional implementation, S4 specifically includes:

[0062] S41: According to the target human body back frame and the processed target human body back RGB-D image, the spinous process line extraction sub-model is used to determine the initial spinous process line of the target human body back.

[0063] S42: According to the target human body back frame and the processed target human body back RGB-D image, the spinous process line starting and ending points are determined by using the spinous process line starting and ending point detection model.

[0064] S43: Determine the target human body back spinous process line according to the target human body back initial spinous process line and the starting and ending points of the target human body back spinous process line.

[0065] In practical applications, a spinous process extraction model with HRNet as the backbone network was constructed. The main part consists of 4 stages, with a total of 4 parallel sub-networks. The feature map resolution of the sub-networks from top to bottom is halved, and the number of channels is doubled. The first stage contains 4 residual units, and the second to fourth stages contain 1, 4, and 3 exchange modules respectively. Each exchange module consists of 4 residual units, and each residual unit contains two 5×5 convolutions at each resolution and a exchange unit across all resolutions.

[0066] During the training phase, the tensors input into the HRNet backbone network are multiple RGB-D images organized in batches with dimensions of [B, H, W, C]. To reduce video memory overhead, the original image is input into two consecutive 3×3 convolution modules with a step size of 2 to obtain a 4x downsampled feature map, which is then input into the constructed HRNet network. The final output is a 4x downsampled heat map, which is then restored to the original resolution using the interpolation method. The maximum value position of one dimension in each row of the heat map is taken as the detected target point and converted into coordinate output. Soft-argmax is used to alleviate the problem of quantization accuracy loss. The heat map and position coordinate learning can be directly learned with a high degree of integration, and the network can directly output the required target.

[0067] The spinous process line only occupies a part of the input human back area, so the same HRNet backbone network as above is used to train the spinous process line extraction sub-model and the spinous process line starting and ending point detection model respectively. Figure 2 As shown in FIG, the longitudinal range of the spinous process line in the back region of the human body is determined by the spinous process line starting and ending point detection model, and then the spinous process line output by the spinous process line extraction sub-model is intercepted, and the curve of the corresponding range is obtained as the final human back spinous process line, as shown in FIG. Figure 3 In practical applications, the extracted back spinous process lines can be used to reconstruct the human spine model.

[0068] The human back spinous process line extraction method of the present application inputs refined data into a neural network to extract the spinous process line, thereby improving the accuracy of the spinous process line extraction.

[0069] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the above-mentioned method for extracting spinous process lines on the back of the human body when executing the computer program.

[0070] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, which, when executed by a processor, implements the above-mentioned method for extracting the back spinous process line of the human body.

[0071] In an exemplary embodiment, a computer program product is provided, including a computer program, which implements the above-mentioned method for extracting human back spinous process lines when executed by a processor.

[0072] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 4As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for extracting the back spinous process line of the human body is implemented.

[0073] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0074] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0075] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0076] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0077] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0078] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for extracting spinous process lines on the back of a human body, characterized in that: include: Obtain an RGB-D image of the target human back and an RGB image of the target human back; Preprocessing the depth data in the target human back RGB-D image to obtain a processed target human back RGB-D image; the preprocessing includes spatial smoothness constraints and depth value constraints; According to the target human back RGB image, a target human back border is extracted using a human back region extraction model; wherein the human back region extraction model is obtained by training a YOLO-pose model using a first training data set; the first training data set includes a training human back RGB image and a corresponding human back border label; According to the target human back border and the processed target human back RGB-D image, the target human back spinous process lines are extracted using a spinous process line extraction model; wherein the spinous process line extraction model includes a spinous process line extraction sub-model and a spinous process line starting and ending point detection model; the spinous process line extraction sub-model is obtained by training the HRNet network using the second training data set; the second training data set includes a training human back RGB-D image with a human back border marked and the corresponding initial spinous process line labels; the spinous process line starting and ending point detection model is obtained by training the HRNet network using the third training data set; the third training data set includes a training human back RGB-D image with a human back border marked and the corresponding spinous process line starting and ending point labels.

2. The method for extracting spinous process lines on the back of a human body according to claim 1, characterized in that: Preprocessing the depth data in the target human back RGB-D image to obtain a processed target human back RGB-D image specifically includes: Performing a spatial smoothing constraint on the depth data in the RGB-D image of the back of the target human body to obtain depth data after a first constraint; Performing a depth value constraint on the depth data in the RGB-D image of the back of the target human body to obtain depth data after a second constraint; An optimized depth value is determined according to the first constrained depth data and the second constrained depth data, thereby obtaining a processed RGB-D image of the back of a target human body.

3. The method for extracting spinous process lines on the back of a human body according to claim 2, characterized in that: Performing spatial smoothing constraints on the depth data in the RGB-D image of the back of the target human body to obtain the depth data after the first constraint specifically includes: Using the formula The depth data in the RGB-D image of the back of the target human body is spatially smoothed to obtain the depth data after the first constraint; wherein, E s (i, j) is the depth data after the first constraint; p(i, j) is the coordinate of the pixel point in the RGB-D image of the back of the target human body; (u x ,u y ) is the principal point coordinate of the camera; f x is the focal length of the camera in the x direction; f y is the focal length of the camera in the y direction; D(i,j) is the depth value corresponding to the pixel.

4. The method for extracting spinous process lines on the back of a human body according to claim 2, characterized in that: The depth data in the RGB-D image of the back of the target human body is subjected to a depth value constraint to obtain the depth data after the second constraint, specifically including: Using Formula E p (i,j)=[D(i,j)-D i (i,j)] 2 →0 performs a depth value constraint on the depth data in the RGB-D image of the back of the target human body to obtain the depth data after the second constraint; wherein, E p (i, j) is the depth data after the second constraint; D(i, j) is the depth value corresponding to the pixel.

5. The method for extracting spinous process lines on the back of a human body according to claim 2, characterized in that: Determining an optimized depth value according to the first constrained depth data and the second constrained depth data specifically includes: Using the formula E(D) = ∑ (i,j) w s E s (i,j)+w p E p (i, j) determining an optimized depth value according to the first constrained depth data and the second constrained depth data; wherein E(D) is the optimized depth value; w s is the weight of the spatial smoothness constraint; E s (i, j) is the depth data after the first constraint; w p is the weight of the depth value constraint; E p (i, j) is the depth data after the second constraint.

6. The method for extracting spinous process lines on the back of a human body according to claim 1, characterized in that: The YOLO-pose model is trained using the first training data set to obtain the human back area extraction model, specifically including: Obtain a first training data set; The training human back RGB image in the first training data set is used as input, and the corresponding human back border label is used as output to train the YOLO-pose model to obtain the human back area extraction model.

7. The method for extracting spinous process lines on the back of a human body according to claim 1, characterized in that: According to the target human back border and the processed target human back RGB-D image, the spinous process line of the target human back is extracted using a spinous process line extraction model, specifically including: According to the target human back border and the processed target human back RGB-D image, using the spinous process line extraction sub-model, determine the target human back initial spinous process line; According to the target human back border and the processed target human back RGB-D image, using the spinous process line starting and ending point detection model, determine the starting and ending points of the target human back spinous process line; The target human back spinous process line is determined according to the target human back initial spinous process line and the starting and ending points of the target human back spinous process line.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for extracting the back spinous process line of the human body according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for extracting the back spinous process line of the human body according to any one of claims 1 to 7 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for extracting the back spinous process line of the human body according to any one of claims 1 to 7 is implemented.

Images