A wafer box contaminant identification method and device based on machine learning

Abstract

The present invention discloses a wafer box contaminant identification method and device based on machine learning, which is used to determine whether the wafer box to be cleaned meets the cleaning requirements, including the following steps: step S1, a driving mechanism drives the wafer box to swing with a preset axis as the center line; step S2, multiple cameras simultaneously capture the image to be inspected and the point cloud image containing the inner wall of the wafer box; step S3, according to the camera shooting interval and the swing speed of the wafer box, the point cloud image containing the surface of the wafer box is spliced to obtain a complete wafer box point cloud image; step S4, the complete wafer box point cloud image is processed to obtain contaminant information on the inner wall of the wafer box; wherein step S3 also includes step S301, identifying the target object in the image to be inspected, and intercepting the portion of the corresponding point cloud image containing the target object. The present invention combines the point cloud image with the image to be inspected to improve detection efficiency while identifying contaminants on the inside of the wafer box.

Description

Technical Field

[0001] The present invention relates to the technical field of semiconductor detection equipment, and in particular to a method and device for identifying contaminants in a wafer box based on machine learning. Background Art

[0002] Because wafers are generally fragile, they are typically stored in dedicated wafer cassettes during production and transportation to prevent damage. To prevent contamination of the wafers within the cassette by dirt inside, the cassette is cleaned before loading the wafers to ensure the cleanliness level within the cassette meets wafer storage requirements.

[0003] With the advancement of semiconductor manufacturing processes, higher requirements are being placed on the cleanliness of wafer cassettes. Therefore, wafer cassettes need to be sent to more sophisticated wafer cassette cleaning machines for cleaning before loading wafers. During this process, if there are large particles of contaminants or wafer fragments in the wafer cassette, not only will they damage the cleaning machine components during the cleaning of the wafer cassette, but also, when the next batch of wafers is placed in the wafer cassette, the new wafers will be broken again due to the presence of residual wafer fragments. Over time, a large number of wafers will be broken, affecting the operation and work efficiency of the entire machine, and even damaging the equipment, which is a huge loss to the industry.

[0004] In order to overcome the above-mentioned defects, those skilled in the art have actively conducted innovative research in order to create a wafer box contaminant identification method based on machine learning. Summary of the Invention

[0005] In view of this, the present invention provides a method for identifying wafer cassette contaminants based on machine learning to address at least one of the problems in the background art. The method is used to determine whether a wafer cassette to be cleaned meets the cleaning requirements of a fully automatic wafer cassette cleaning machine, and includes the following steps:

[0006] Step S1, the driving mechanism drives the wafer box to swing around a preset axis as the center line;

[0007] Step S2, multiple cameras simultaneously capture the image to be inspected and the point cloud image containing the inner wall of the wafer box;

[0008] Step S3, combining the point cloud images including the surface of the wafer box according to the camera shooting interval and the swing speed of the wafer box to obtain a complete point cloud image of the wafer box;

[0009] Step S4, processing the complete wafer box point cloud image to obtain contaminant information on the inner wall of the wafer box;

[0010] Wherein, step S3 also includes identifying pollutants present in the image to be inspected, and intercepting the portion of the corresponding point cloud image containing the pollutants.

[0011] Optionally, the image to be inspected is binarized according to a first preset threshold value to obtain a binarized image, and based on the binarized image, a portion of the point cloud image containing the image of the inner wall of the wafer box is intercepted.

[0012] Optionally, the first preset threshold is obtained through a convolutional neural network machine learning model.

[0013] Optionally, a second preset threshold is obtained through a convolutional neural network machine learning model, the image to be inspected is binarized according to the second preset threshold, a target area containing possible pollutants is selected, and the area ratio of the pollutants to the target image is greater than a third preset threshold. Based on the positional relationship between the image to be inspected and the point cloud image, a local point cloud map of the selected target area in the point cloud image is intercepted, and the local point cloud map is processed to determine whether there are pollutants in the target area.

[0014] Optionally, in step S4, the pollutant information includes location information, size information, and quantity information.

[0015] Optionally, step S4 further includes step S5, determining the next processing method of the wafer box according to the contaminant information.

[0016] Optionally, in step S1, the swing angle of the wafer box is not greater than ±60°, and the swing speed is not greater than 5 rpm.

[0017] Optionally, in step S2, the image overlap rate between any two images to be inspected is between a fourth preset threshold and a fifth preset threshold.

[0018] Optionally, the camera includes two depth cameras, and the two depth cameras are mirror-symmetrical about a preset plane.

[0019] The present application also provides a machine learning-based wafer box contaminant identification method, which adopts any of the machine learning-based wafer box contaminant identification methods described above.

[0020] The beneficial effects of the present invention are:

[0021] The present invention uses a depth camera to collect point cloud images of the wafer box surface, and obtains depth information of the wafer box surface through the point cloud image to determine whether there are contaminants such as debris on the wafer box surface, avoiding problems such as misjudgment caused by fading of the wafer box in 2D detection.

[0022] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0024] Figure 1 It is a structural schematic diagram of the present invention;

[0025] Figure 2 is a top view of the present invention;

[0026] Figure 3 is a schematic diagram of the locations of pollutants in an embodiment of the present invention;

[0027] Figure 4 is a graph showing the relationship between pollutant normalized offset and swing speed in an embodiment of the present invention;

[0028] Figure 5 is a graph showing the relationship between pollutant normalized offset and swing angle in an embodiment of the present invention;

[0029] The parts in the accompanying drawings are marked as follows:

[0030] 1. Drive mechanism; 2. Camera assembly; 2a. 2D camera; 2b. Depth camera; 3. Wafer cassette cleaning machine; 301. Loading station;

[0031] F. Wafer box. DETAILED DESCRIPTION

[0032] The exemplary embodiments disclosed herein will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the specific embodiments described herein. Rather, these embodiments are provided to enable a more thorough understanding of the present application and to fully convey the scope of the present application to those skilled in the art.

[0033] In the following description, numerous specific details are provided to provide a more thorough understanding of the present application. However, it will be apparent to those skilled in the art that the present application can be practiced without one or more of these details. In other instances, certain technical features known in the art are not described to avoid confusion with the present application; that is, all features of actual embodiments are not described herein, nor are well-known functions and structures described in detail.

[0034] In the drawings, the sizes of layers, regions, elements and their relative sizes may be exaggerated for clarity. Like reference numerals denote like elements throughout.

[0035] It should be understood that when an element or layer is referred to as being "on," "adjacent to," "connected to," or "coupled to" another element or layer, it may be directly on, adjacent to, connected to, or coupled to the other element or layer, or there may be intervening elements or layers. Conversely, when an element is referred to as being "directly on," "directly adjacent to," "directly connected to," or "directly coupled to" another element or layer, there may be no intervening elements or layers. It should be understood that although the terms first, second, third, etc. may be used to describe various elements, components, regions, layers, and / or parts, these elements, components, regions, layers, and / or parts should not be limited by these terms. These terms are merely used to distinguish one element, component, region, layer, or part from another element, component, region, layer, or part. Therefore, without departing from the teachings of the present application, the first element, component, region, layer, or part discussed below may be represented as a second element, component, region, layer, or part. And when the second element, component, region, layer, or part is discussed, it does not necessarily mean that the first element, component, region, layer, or part is present in the present application.

[0036] Spatially relative terms such as "under," "beneath," "below," "under," "above," "above," etc., may be used herein for convenience of description to describe the relationship of an element or feature shown in the figures to other elements or features. It should be understood that in addition to the orientations shown in the figures, the spatially relative terms are intended to include different orientations of the device in use and operation. For example, if the device in the drawings is turned over, then the elements or features described as "under" or "beneath" or "beneath" the other elements will be oriented "above" the other elements or features. Thus, the exemplary terms "under" and "under" can include both the above and below orientations. The device can be oriented otherwise (rotated 90 degrees or in other orientations) and the spatial descriptors used herein are interpreted accordingly.

[0037] The purpose of the terms used herein is only to describe specific embodiments and is not intended to limit the present application. When used herein, the singular forms "a", "an", and "the" are intended to include the plural forms, unless the context clearly indicates otherwise. It should also be understood that the terms "comprising" and / or "including", when used in this specification, determine the presence of the features, integers, steps, operations, elements and / or parts, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, parts and / or groups. When used herein, the term "and / or" includes any and all combinations of the relevant listed items.

[0038] In order to fully understand the present application, detailed steps and detailed structures will be presented in the following description to illustrate the technical solution of the present application. The preferred embodiments of the present application are described in detail below. However, in addition to these detailed descriptions, the present application may also have other implementation methods.

[0039] The wafer cassette F referred to in the embodiments of this application specifically refers to the body portion of the wafer cassette. Generally speaking, a wafer cassette comprises a separable body and lid. During wafer transport, the body and lid are locked together. During wafer cassette cleaning, the body and lid are unlocked and cleaned separately or placed in the same cleaning unit for cleaning. After cleaning, the body and lid are relocked. Several slots are provided on a set of opposing inner walls of the cassette. During substrate transport, a substrate is secured between two slots.

[0040] like Figure 1 and Figure 2 As shown, a wafer box contaminant identification device based on machine learning includes:

[0041] The driving mechanism 1 is configured to fix the wafer box and drive the wafer box to swing at a constant speed.

[0042] The camera assembly 2 is configured to include a 2D camera 2a and two depth cameras 2b. The two depth cameras are symmetrically arranged about a preset plane. The preset plane is the plane where the center line of the wafer box is located during the swinging process.

[0043] The control component is used to process and analyze the images taken by the camera component to determine the contaminants on the inner wall of the wafer box and make decisions on the next processing method of the wafer box.

[0044] In this embodiment, the driving mechanism, the camera assembly, and the control assembly are all arranged at the loading station 301 of the wafer box cleaning machine 3. After the wafer box is loaded, it is inspected and whether it can be cleaned in the wafer box cleaning machine is determined based on the inspection results.

[0045] The present invention provides a machine learning-based wafer cassette contaminant identification method for determining whether a wafer cassette to be cleaned meets the cleaning requirements of a fully automatic wafer cassette cleaning machine, including the following steps:

[0046] Step S1, the driving mechanism drives the wafer box to swing around a preset axis as the center line;

[0047] In this step, the wafer box is transferred to the driving mechanism by the external action mechanism. The driving mechanism fixes the wafer box and drives the wafer box to swing at a preset speed and preset amplitude. The preset axis is on the symmetric center plane of the wafer box.

[0048] Step S2, multiple cameras simultaneously capture the image to be inspected and the point cloud image containing the inner wall of the wafer box;

[0049] In this step, the cameras, including a 2D camera and a depth camera, capture an image of the cassette's interior surface at a preset interval. This interval is set based on the cassette's swing speed. By adjusting the interval between camera captures, multiple images can be stitched together to create a complete image of the cassette's interior surface.

[0050] Step S3, combining the point cloud images including the surface of the wafer box according to the camera shooting interval and the swing speed of the wafer box to obtain a complete point cloud image of the wafer box;

[0051] In this step, multiple images are stitched together to obtain a complete image of the wafer box surface, and the image is processed to detect the residual contaminants on the wafer box surface.

[0052] Step S4 , processing the complete wafer box point cloud image to obtain contaminant information on the inner wall of the wafer box; identifying contaminants in the image to be inspected, and intercepting the portion of the corresponding point cloud image containing the contaminants.

[0053] The inner wall of the wafer box in the image to be inspected is identified through a machine learning algorithm, and the area in the image containing the image of the inner wall of the wafer box is determined. The conversion relationship between the image to be inspected and the point cloud image is obtained based on the spatial position relationship between the 2D camera and the depth camera. The corresponding area in the point cloud image is intercepted, and finally the cropped point cloud images are spliced to obtain a complete point cloud image of the inner wall of the wafer box.

[0054] Those skilled in the art may select a method for stitching point cloud images according to actual conditions, as long as the above-mentioned effects can be achieved. The stitching method is not specifically limited here.

[0055] It is understandable that in order to prevent large particle contaminants remaining in the wafer box from damaging the wafer box cleaning machine during the cleaning process, the wafer box to be cleaned needs to be inspected before it is sent to the wafer box cleaning machine. Common large particle contaminants in the wafer box are wafer fragments. When detecting contaminants in the wafer box through image recognition, due to the reflection of light by wafer fragments and the color difference of the wafer box produced as the wafer box is used, there are often missed detections and qualified wafer boxes are misjudged as unqualified. Similarly, weighing recognition, ultrasonic sensors and photoelectric sensors also have corresponding disadvantages. The applicant found in many experiments that the point cloud image of the inner wall of the wafer box taken by the depth camera, the depth information of the inner wall of the wafer box is obtained through point cloud image processing, and the depth information is used to judge the size information of the contaminants, thereby avoiding the interference of contaminants and the color of the inner wall of the wafer box on the identification of contaminants, and improving the recognition accuracy.

[0056] In this embodiment, by recognizing the image captured by the 2D camera, the background portion of the image to be inspected that is useless for identifying contaminants is removed, and the corresponding area of the point cloud image is simultaneously removed, thereby obtaining a partial area of the point cloud image containing the wafer box image, reducing the task of analyzing the point cloud image and improving the recognition rate.

[0057] In an optional embodiment, the image to be inspected is binarized according to a first preset threshold value to obtain a binarized image, and a portion of the point cloud image containing the image of the inner wall of the wafer box is intercepted based on the binarized image.

[0058] In this embodiment, the wafer box is placed on a light-colored work surface, that is, the grayscale value of the work surface on which the wafer box is placed is smaller than the grayscale value of the wafer box. Through processing, the image to be inspected is processed into a binary image, and then the point cloud image of the corresponding area is intercepted according to the area where the corresponding color block in the binary image is located.

[0059] The specific operation is to convert the image to be inspected into a grayscale image. According to the pre-set threshold and the first preset threshold, the grayscale value of the point with a grayscale value less than the threshold is set to 0, and the grayscale value of the point with a grayscale value greater than the threshold is set to 255. The detection area is selected as the area where the black block is located, and finally the corresponding part of the point cloud image is intercepted and the corresponding part is detected.

[0060] It is understandable that due to the color difference between the wafer box and the workbench on which the wafer box is placed, in the grayscale image, the wafer box area and the background area have a clear boundary, so selecting the area where the corresponding color block is located can include all the inner walls of the wafer box.

[0061] In an optional embodiment, the first preset threshold is obtained through a convolutional neural network machine learning model.

[0062] By establishing a convolutional neural network machine learning model and training the machine learning model, a first preset threshold corresponding to the model of each different type of wafer box is set, and it can be ensured that for wafer boxes of the same type but with different degrees of aging, the grayscale value of the wafer box surface in the grayscale image is always less than the first preset threshold, and the area where the wafer box is located in the grayscale image after binarization is always a black area, so that the method of this embodiment can be applicable to all types of wafer boxes and aged and faded wafer boxes.

[0063] In an optional embodiment, in step S1, the swing angle of the wafer box is no greater than ±60°, and the swing speed is no greater than 5 rpm.

[0064] It is understandable that there are a large number of raised and recessed structures in the slot structure on the surface of the wafer box. If a camera is used to shoot the surface of a stationary wafer box, the raised structures on the surface of the wafer box may block contaminants, thereby causing problems such as missed detection. By shaking the wafer box back and forth, the camera can take photos of the inner wall of the wafer box from multiple angles. Stitching these photos together can minimize the problem of missed detection caused by blind spots in shooting.

[0065] However, as the cassette swings, contaminants adhering to the cassette's inner walls can easily fall off and re-adhere to the cassette's inner walls. Consequently, when capturing images of the cassette's inner walls, a single contaminant can be captured in multiple images. By controlling the cassette's swing amplitude and speed, we can minimize contaminant loss while maintaining inspection and image capture efficiency.

[0066] In an optional embodiment, in step S2, the image overlap rate between any two images to be inspected is between a fourth preset threshold and a fifth preset threshold.

[0067] In this embodiment, the camera's capture frequency is adjusted based on the cassette's swing speed, thereby controlling the overlap ratio between two adjacent images. It's understandable that a higher overlap ratio reduces blind spots caused by the slot structure during subsequent image stitching. However, a higher overlap ratio also increases the number of images required to stitch together a complete cassette image. This increases the image processing workload and difficulty, and, to a certain extent, increases the cost of inspection.

[0068] In an optional embodiment, the fourth preset threshold is 0.10-0.35, and the fifth preset threshold is 0.30-0.65.

[0069] In a more preferred embodiment, the fourth preset threshold is 0.15-0.25, and the fifth preset threshold is 0.45-0.55.

[0070] In an optional embodiment, step S4 further includes step S5, which determines the next processing method of the wafer box based on the contaminant information.

[0071] The contaminant information obtained through deep image processing is used to determine the type of contaminant and the possible damage that the contaminant may cause to the wafer box cleaning machine during wafer box cleaning, and this is used as a basis for determining the next step in wafer box processing.

[0072] In an optional embodiment, different reference values are set for contaminants of different sizes and types, and different weights are assigned according to the damage to the wafer box cleaning machine. Based on this, a threshold value of wafer boxes that can be sent to the wafer box cleaning machine for cleaning is set, and the hazard coefficient of the contaminants in the wafer box to the wafer box cleaning machine is calculated based on the obtained contaminant information. When the hazard coefficient is less than or equal to the threshold value, the wafer box can be sent to the wafer box cleaning machine for cleaning.

[0073] Those skilled in the art may set the reference values, weights, thresholds and other parameters of pollutants according to actual conditions, as long as the above-mentioned effects can be achieved. No specific limitation is imposed on the setting of the reference values, weights, thresholds and other parameters of pollutants.

[0074] In the above embodiment, the subsequent processing method includes sending the wafer to a wafer box cleaning machine for cleaning and the control component issuing an alarm.

[0075] In this embodiment, after receiving the alarm signal, the on-site staff determines whether the wafer box can be sent to the wafer box cleaning machine for cleaning based on the actual situation inside the wafer box.

[0076] In an optional embodiment, the pollutant information includes location information, size information, and quantity information.

[0077] In an optional embodiment, a second preset threshold is obtained through a convolutional neural network machine learning model, the image to be inspected is binarized according to the second preset threshold, a target area containing possible pollutants is selected, and the area ratio of the pollutants to the target image is greater than a third preset threshold. Based on the positional relationship between the image to be inspected and the point cloud image, a local point cloud map of the selected target area in the point cloud image is intercepted, and the local point cloud map is processed to determine whether there are pollutants in the target area.

[0078] In this embodiment, a second threshold value is obtained through a machine learning algorithm based on the grayscale values of the inner wall of the wafer box with different degrees of aging and the grayscale values of pollutants of different types and positions in the grayscale image, so that the area where the pollutants are located in the image after binarization processing is displayed in white, and the area where the inner wall of the wafer box is located is displayed in black, so that the area where the pollutants are located is selected, and the corresponding image in the point cloud image is processed to determine whether there are pollutants in the selected area, and relevant information about the pollutants is obtained through further processing and analysis.

[0079] It's understandable that processing and analyzing the images to be inspected is less difficult than processing and analyzing depth images. By analyzing the images to be inspected, identifying areas where contaminants may be present, and processing the corresponding point cloud images, the amount of point cloud processing can be significantly reduced, improving processing efficiency. Furthermore, analyzing the images to be inspected can screen out a large number of wafer cassettes that are free of contaminants, allowing them to pass inspection more quickly and improving the overall inspection efficiency of the device.

[0080] In the above embodiment, the third preset threshold is between 0.8 and 0.95. By setting the third preset threshold, areas where pollutants may exist are selected, thereby reducing the processing load of the point cloud image.

[0081] Examples 1-9, such as Figure 3 As shown, in this embodiment, a contaminant is placed at wafer boxes A, B, and C respectively, and the wafer boxes are fixed on a driving mechanism, which drives the wafer boxes to swing back and forth with a swing amplitude of 60 degrees and at different swing speeds for 2 minutes.

[0082] Example Swing speed rpm 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9

[0083] For Examples 1-9, the offset of the contaminants from the initial position of the wafer box after the swing is completed is measured, and the normalized offset is obtained after normalization calculation. The relationship between the normalized offset of the contaminants on the surface of the wafer box after the swing is completed and the swing speed is as follows: Figure 4 shown.

[0084] Depend on Figure 4 It can be seen that when the swing speed is greater than 5 rpm, the displacement of the pollutants increases significantly.

[0085] Examples 10-15, such as Figure 3 As shown, in this embodiment, a contaminant is placed at wafer boxes A, B, and C respectively, and the wafer boxes are fixed on a driving mechanism, which drives the wafer boxes to swing back and forth at a speed of 5 rpm and at different swing angles for 2 minutes.

[0086] Example Swing angle 10 30 11 40 12 50 13 60 14 70 15 80

[0087] For Examples 10-15, the offset of the contaminants from the initial position of the wafer box after the swing is completed is measured, and the normalized offset is obtained after normalization calculation. The relationship between the normalized offset of the contaminants on the surface of the wafer box after the swing is completed and the swing angle is as follows: Figure 5 shown.

[0088] Depend on Figure 5 It can be seen that when the swing angle is greater than 60 degrees, the displacement of the pollutants increases significantly.

[0089] It should be understood that the above embodiments are exemplary and are not intended to encompass all possible implementations of the claims. Various modifications and variations may be made to the above embodiments without departing from the scope of the present disclosure. Similarly, the various technical features of the above embodiments may be arbitrarily combined to form additional embodiments of the present application that may not be explicitly described. Therefore, the above embodiments merely illustrate several implementations of the present application and do not limit the scope of protection of the patent application.

Claims

1. A wafer cassette contaminant identification method based on machine learning, used to determine whether a wafer cassette to be cleaned meets the cleaning requirements of a fully automatic wafer cassette cleaning machine, characterized in that: The following steps are involved: Step S1, the driving mechanism drives the wafer box to swing around a preset axis as the center line; Step S2, the 2D camera and the depth camera respectively capture an image to be inspected and a point cloud image containing the inner wall of the wafer box; Step S3, combining the point cloud images including the surface of the wafer box according to the camera shooting interval and the swing speed of the wafer box to obtain a complete point cloud image of the wafer box; Step S4, processing the complete wafer box point cloud image to obtain contaminant information on the inner wall of the wafer box; Identify pollutants present in the image to be inspected, obtain a conversion relationship between the image to be inspected and the point cloud image based on the spatial position relationship between the 2D camera and the depth camera, and intercept the portion of the corresponding point cloud image containing pollutants; Specifically, according to a first preset threshold, the image to be inspected is binarized to obtain a binarized image, and according to the binarized image, the part of the image of the inner wall of the wafer box in the point cloud image is intercepted; the first preset threshold is obtained by a convolutional neural network machine learning model; a second preset threshold is obtained by the convolutional neural network machine learning model, and the image to be inspected is binarized according to the second preset threshold, and a target area containing possible contaminants is selected, and the area ratio of the contaminants to the target area where the contaminants may appear is greater than a third preset threshold, and according to the positional relationship between the image to be inspected and the point cloud image, a local point cloud map of the selected target area in the point cloud image is intercepted, and the local point cloud map is processed to obtain depth information of the inner wall of the wafer box, and the size information of the contaminants is obtained by using the depth information of the inner wall of the wafer box to determine whether there are contaminants in the target area.

2. The method for identifying wafer cassette contaminants based on machine learning according to claim 1, wherein: In step S4, the pollutant information includes location information, size information, and quantity information.

3. The method for identifying wafer cassette contaminants based on machine learning according to claim 2, wherein: Step S4 also includes step S5, which determines the next processing method of the wafer box according to the contaminant information.

4. The method for identifying wafer cassette contaminants based on machine learning according to claim 1, wherein: In step S1 , the wafer box swing angle is no greater than ±60°, and the swing speed is no greater than 5 rpm.

5. The method for identifying wafer cassette contaminants based on machine learning according to claim 1, wherein: In step S2, the image overlap rate between any two images to be inspected is between a fourth preset threshold and a fifth preset threshold.

6. The method for identifying wafer cassette contaminants based on machine learning according to claim 1, wherein: The camera includes two depth cameras, and the two depth cameras are mirror-symmetrical about a preset plane.

7. A wafer box contaminant identification device based on machine learning, characterized in that: A wafer box contaminant identification method based on machine learning as described in any one of claims 1 to 6 is adopted.

Images