CN116861835A - Well region planning method for simulated layout design - Google Patents

Well region planning method for simulated layout design Download PDF

Info

Publication number
CN116861835A
CN116861835A CN202310951773.7A CN202310951773A CN116861835A CN 116861835 A CN116861835 A CN 116861835A CN 202310951773 A CN202310951773 A CN 202310951773A CN 116861835 A CN116861835 A CN 116861835A
Authority
CN
China
Prior art keywords
well region
layout design
planning method
region
simulated layout
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202310951773.7A
Other languages
Chinese (zh)
Inventor
仝明磊
李易婉
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai Electric Power University
Original Assignee
Shanghai Electric Power University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shanghai Electric Power University filed Critical Shanghai Electric Power University
Priority to CN202310951773.7A priority Critical patent/CN116861835A/en
Publication of CN116861835A publication Critical patent/CN116861835A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/30—Circuit design
    • G06F30/39—Circuit design at the physical level
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/30—Circuit design
    • G06F30/39—Circuit design at the physical level
    • G06F30/392—Floor-planning or layout, e.g. partitioning or placement
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Abstract

The invention provides a well region planning method used in simulated layout design, which relates to the field of simulated IC layout design automation, deep learning and computer vision, and comprises the following steps: s1, extracting an initial layout as a picture with a regular size, which comprises a trap covering module and an uncovered module, as an initial data image; s2, segmenting a trap covering module in the initial data image by utilizing an instant segmentation algorithm to generate an approximate contour of a trap region; and S3, optimizing the well region by using a heuristic generation algorithm based on a traditional image processing method, so that the well region accords with the process design rule. The method adopts the flow of the instant segmentation algorithm and the heuristic generation algorithm to plan the well region, so that the automatic generation of the well region in the automatic design of the simulated layout is more rapid and effective.

Description

Well region planning method for simulated layout design
Technical Field
The invention relates to the technical fields of automation, deep learning and computer vision of analog IC layout design, in particular to a well region planning method used in analog layout design.
Background
Digital IC designs have now become substantially automated, while automated implementation of analog IC designs remains to be challenged. The analog IC design mainly comprises two parts of a circuit design and a layout design, wherein the layout design work is more biased to manual experience due to the high flexibility and artistry of the layout design work. Under the situation, if the layout design EDA tool can generate a high-quality layout, the number of loop iteration times of an analog circuit engineer in the design and simulation engineering can be greatly reduced, the iteration of a product can be accelerated, the design efficiency is improved, and the success probability of flow sheets is increased.
The existing automatic generation method of the well region mainly comprises two types: firstly, a method of adding rectangular constraint by using simple geometric calculation, such as 'Ou H C, tseng K H, liu JY, et al Layout-dependent-effects-aware analytical analog placement [ C ]// Proceedings ofthe 52ndAnnual DesignAutomation Conference.2015:1-6' ], wherein the method is extremely easy to cause communication errors in complex circuit design, and well regions generated in complex circuits are not smooth and have more rectangular corner boundaries due to the addition of rectangular constraint; secondly, the generated type countermeasure network such as Xu B, lin Y, tang X, et al Wellgan: general-active-differential-network-guided well generation for ana-log/mixed-signal circuit layout [ C ]// Proceedings of the th An-nual Design Automation Conference 2019.2019:1-6 is utilized to simulate the experience layout of a layout design engineer, so that the experience guidance of the design engineer is added in the automatic design, the generated well region is more similar to the manual design scheme, and the method needs to guide the automatic generation of the machine by learning the manual design experience, namely, the prior learning is needed for each process design, and a large amount of work is needed.
Disclosure of Invention
The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to provide a well region planning method for use in a simulated layout design.
The invention provides a well region planning method for simulated layout design, which has the characteristics that the method comprises the following steps: s1, extracting an initial layout as a picture with a regular size, which comprises a trap covering module and an uncovered module, as an initial data image; s2, segmenting a trap covering module in the initial data image by utilizing an instant segmentation algorithm to generate an approximate contour of a trap region; and S3, carrying out optimization treatment on the well region by using a heuristic generation algorithm to enable the well region to accord with the process design rule.
The well region planning method for the simulated layout design provided by the invention can also have the following characteristics: the pictures are RGB pictures, and the trap covering module, the uncovered module and the trap area are distinguished by adopting different colors.
The well region planning method for the simulated layout design provided by the invention can also have the following characteristics: in step S1, the picture is changed to a prescribed size by upsampling or downsampling.
The well region planning method for the simulated layout design provided by the invention can also have the following characteristics: in step S1, a new display file is created at the time of extraction to more quickly distinguish between trap-covered and uncovered modules.
The well region planning method for the simulated layout design provided by the invention can also have the following characteristics: step S2 comprises the steps of: s2-1, marking RoI (Region ofInterest) in an initial data image by using a marking tool Labelme to generate a label data set; and S2-2, training and learning a data set consisting of the initial data image and the tag data set by utilizing an instant segmentation algorithm, and segmenting the rough outline of the RoI, namely the well region.
The well region planning method for the simulated layout design provided by the invention can also have the following characteristics: the instant segmentation algorithm adopts S 2 -a FPN network.
The well region planning method for the simulated layout design provided by the invention can also have the following characteristics: step S3 comprises the steps of: s3-1, obtaining a maximum circumscribed rectangular region R of the RoI generated by the instant segmentation algorithm i And outputs the region R i Is a picture of the image and its vertex coordinates; step S3-2, acquiring region R i Red rectangular block B corresponding to each trap covering module in the array i Upper left vertex coordinates of (2) and length w and width h of the rectangle, and therebyObtaining region R i Each red rectangular block B in i Then for red rectangular block B i Heuristic growth operation is carried out until all red rectangular blocks B are traversed i Obtaining a well region which is closed and regular in shape; step S3-3, the processed region R i Returning to the initial data image, and legalizing the well region in the initial data image.
In step S3-2, the heuristic growth operation includes: setting up, down, left and right growth directions, and marking as (0, 1,2 and 3); for red rectangular block B i Each of the four vertices of (a) is set with the following growth direction: coordinates (x, y) correspond to directions (0, 2), coordinates (x+w, y) correspond to directions (0, 3), coordinates (x, y+h) correspond to directions (1, 2), coordinates (x+w, y+h) correspond to directions (2, 4); each growth direction is performed as set forth below: meeting region R in the growth direction i Other red rectangular blocks B in i A straight line grows, and the red rectangular block B is not touched in the growth direction i Or directly touch the line, and does not grow.
In step S3-3, the legalization process includes: performing margin expansion on the well region according to a process design rule; edge merging is performed on the well region.
Further, edge merging is performed as set forth below: if two adjacent parallel straight lines exist and the distance difference between the two straight lines is smaller than the distance determined by the process, the two straight lines are combined into one straight line according to the straight line with the longer distance.
Effects and effects of the invention
According to the well region planning method for the simulated layout design, the process of 'lightweight instant segmentation algorithm + heuristic generation algorithm' is adopted to carry out well region planning. The method is characterized in that the well region can be obtained faster by the instant segmentation algorithm, the workload is greatly reduced, the heuristic generation algorithm is an accurate region planning algorithm, the generated well region edge is smoother, and more edge and corner boundary problems can not occur. In summary, the well region planning method used in the simulated layout design enables the automatic generation of the well region in the simulated layout automatic design to be quicker and more effective.
Drawings
FIG. 1 is a flow diagram of a well region planning method for use in a simulated layout design in an embodiment of the present invention;
FIG. 2 is a schematic diagram of an initial data image in an embodiment of the invention
FIG. 3 is a schematic diagram of an initial data image after image segmentation in an embodiment of the present invention;
FIG. 4 is a schematic illustration of an image of a well region optimized for each RoI in an embodiment of the invention;
FIG. 5 is a schematic illustration of an optimized global area image in an embodiment of the invention;
fig. 6 is a schematic diagram of an overall area image after the legalization process in an embodiment of the present invention.
Detailed Description
In order to make the technical means, the creation features, the achievement of the purpose and the effect of the present invention easy to understand, the present invention is specifically described below with reference to the accompanying drawings.
Examples
FIG. 1 is a flow diagram of a well region planning method for use in a simulated layout design.
As shown in fig. 1, the present embodiment provides a well region planning method for use in a simulated layout design, including the following steps:
step S1, extracting an initial layout as an RGB picture with a regular size, which comprises a trap covering module and an uncovered module, as an initial data image. The method comprises the step of preprocessing data of an initial layout.
Wherein upon extraction, the trap overlay module and the uncovered module can be distinguished more quickly by creating a new display file.
In this embodiment, as shown in the schematic diagram of the initial data image of fig. 2, the well covering module is represented as a rectangular block without small square marks, and the uncovered module is represented as a rectangular block with small square marks; the image size is changed to a prescribed input size by upsampling or downsampling. Note that the small square marks in the figure are only used for distinguishing the trap covered module and the uncovered module after black-and-white processing in the specification drawing, and the trap covered module and the uncovered module in the actual picture are distinguished by different colors, such as red and green, and do not contain the small square marks; when outputting, it is necessary to confirm whether the image size and the image content have only trap covering modules and uncovered modules except the background.
The initial data image can be used as a verification set of a subsequent instant segmentation algorithm for verifying the result of the instant segmentation algorithm.
And S2, segmenting the trap covering module in the initial data image by utilizing an instant segmentation algorithm to generate the approximate outline of the trap region.
Specifically, step S2 includes the following subdivision steps:
step S2-1, data set preparation is carried out: and (3) performing RoI marking in the initial data image by using a marking tool Labelme to generate a label part in the current data set. It should be noted that the dataset is divided into two parts in total: the image dataset (referring to the initial data image) and the label dataset both include a training set and a test set, the training set and the test set being distinguished by randomly selecting 80% of the initial dataset for training and the remaining 20% for testing.
Step S2-2, using an instant segmentation algorithm (S 2 FPN), training learning the dataset, segmenting the rough outline of the RoI, i.e. well region, in the initial data image.
In this embodiment, as shown in the schematic diagram of the initial data image after image segmentation in fig. 3, the RoI is denoted as a line bounding region, and the RoI in the actual image is denoted as a blue line bounding region.
And S3, optimizing the well region by using a heuristic generation algorithm based on a traditional image processing method, so that the well region accords with the process design rule. The well region in the step S2 is precisely segmented. The process design rules comprise a minimum area rule, a minimum interval rule and the like.
Specifically, step S3 includes the following subdivision steps:
step S3-1, obtaining the most significant RoILarge circumscribed rectangular region R i And outputs the region R i See fig. 4 for a schematic representation of the well region image after each RoI optimization. Wherein the output picture and coordinates are prepared for subsequent processing back to the original data image.
Step S3-2, acquiring region R i Red rectangular block B corresponding to each trap covering module in the array i Upper left vertex coordinates of (2) and length and width (w, h) of the rectangle, and thereby obtaining region R i Each red rectangular block B in i Then for these red rectangular blocks B i Heuristic growth operation is carried out until all red rectangular blocks B are traversed i And obtaining a closed well region with a regular shape, which is shown in a schematic diagram of the whole region image after the optimization of fig. 5.
The heuristic growth operation specifically comprises the following steps: setting up, down, left and right growth directions, and marking as (0, 1,2 and 3); for red rectangular block B i Each of the four vertices of (a) is set with the following growth direction: coordinates (x, y) correspond to directions (0, 2), coordinates (x+w, y) correspond to directions (0, 3), coordinates (x, y+h) correspond to directions (1, 2), coordinates (x+w, y+h) correspond to directions (2, 4); each growth direction is performed as set forth below: meeting region R in the growth direction i Other red rectangular blocks B in i A straight line is grown, and the region R is not touched in the growth direction i Other red rectangular blocks B in i Or directly touch the line, and does not grow.
Step S3-3, the region R processed in step S3-2 i Returning to the initial data image, and performing legal treatment on the well region so that the generated well region conforms to the process design rule, see the schematic diagram of the whole region image after legal treatment in fig. 6.
Wherein, the legalization process includes: performing margin expansion on the well region according to a process design rule; if two adjacent parallel straight lines exist and the distance between the two straight lines is smaller than the distance a (a is determined by the process), the two straight lines are combined into one straight line according to the straight line with the longer distance, namely, the edge combination is carried out on the well region.
Effects and effects of the examples
According to the well region planning method for the simulated layout design, the process of 'lightweight instant segmentation algorithm + heuristic generation algorithm' is adopted to carry out well region planning. The method is characterized in that the well region can be obtained faster by the instant segmentation algorithm, the workload is greatly reduced, the heuristic generation algorithm is an accurate region planning algorithm, the generated well region edge is smoother, and more edge and corner boundary problems can not occur. In summary, the well region planning method used in the simulated layout design enables the automatic generation of the well region in the simulated layout automatic design to be quicker and more effective.
The above embodiments are preferred examples of the present invention, and are not intended to limit the scope of the present invention.

Claims (10)

1. The well region planning method for the simulated layout design is characterized by comprising the following steps of:
s1, extracting an initial layout as a picture with a regular size, which comprises a trap covering module and an uncovered module, as an initial data image;
s2, segmenting the trap covering module in the initial data image by utilizing an instant segmentation algorithm to generate an approximate contour of a trap region;
and S3, optimizing the well region by using a heuristic generation algorithm to enable the well region to conform to a process design rule.
2. The well region planning method for use in simulated layout design as claimed in claim 1, wherein:
the pictures are RGB pictures, and the well covering module, the uncovered module and the well region are distinguished by adopting different colors.
3. The well region planning method for use in simulated layout design as claimed in claim 1, wherein:
in step S1, the picture is changed to a specified size by upsampling or downsampling.
4. The well region planning method for use in simulated layout design as claimed in claim 1, wherein:
wherein in step S1, a new display file is created at the time of extraction to more rapidly distinguish the well covered module from the uncovered module.
5. The well region planning method for use in simulated layout design as claimed in claim 1, wherein:
wherein, the step S2 includes the following steps:
s2-1, performing RoI marking in the initial data image by using a marking tool Labelme to generate a tag data set;
and S2-2, training and learning a data set consisting of the initial data image and the tag data set by utilizing the instant segmentation algorithm, and segmenting out the rough outline of the RoI, namely the well region.
6. The well region planning method for use in simulated layout design as claimed in claim 1, wherein:
wherein the instant segmentation algorithm is S 2 -a FPN network.
7. The well region planning method for use in simulated layout design as claimed in claim 1, wherein:
wherein, the step S3 includes the following steps:
s3-1, obtaining a maximum circumscribed rectangular region R of the RoI generated by the instant segmentation algorithm i And outputs the region R i Is a picture of the image and its vertex coordinates;
step S3-2, obtaining the region R i Red rectangular block B corresponding to each trap covering module in the array i Upper left vertex coordinates of (2) and length w and width h of the rectangle, and therebyObtaining the region R i Each of the red rectangular blocks B in i Then for the red rectangular block B i Heuristic growing operation is carried out until all the red rectangular blocks B are traversed i Obtaining the well region which is closed and regular in shape;
step S3-3, the processed region R i And returning to the initial data image, and performing legal processing on the well region in the initial data image.
8. The well region planning method for use in simulated layout design as claimed in claim 7, wherein:
wherein, in the step S3-2, the heuristic growth operation comprises:
setting up, down, left and right growth directions, and marking as (0, 1,2 and 3);
for the red rectangular block B i Each of the four vertices of (a) is set with the following growth direction: coordinates (x, y) correspond to directions (0, 2), coordinates (x+w, y) correspond to directions (0, 3), coordinates (x, y+h) correspond to directions (1, 2), coordinates (x+w, y+h) correspond to directions (2, 4);
each of the growth directions is performed according to the following settings:
meeting the region R in the growth direction i Other of the red rectangular blocks B i A straight line is grown,
the red rectangular block B is not touched in the growth direction i Or directly touch the line, and does not grow.
9. The well region planning method for use in simulated layout design as claimed in claim 7, wherein:
in the step S3-3, the legalization process includes:
performing margin expansion on the well region according to a process design rule;
and carrying out edge combination on the well region.
10. The well region planning method for use in simulated layout design as claimed in claim 9, wherein:
wherein, in the step S3-3, the edge merging is performed according to the following settings:
if two adjacent parallel straight lines exist and the distance difference between the two straight lines is smaller than the distance determined by the process, the two straight lines are combined into one straight line according to the straight line with the longer distance.
CN202310951773.7A 2023-07-31 2023-07-31 Well region planning method for simulated layout design Pending CN116861835A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310951773.7A CN116861835A (en) 2023-07-31 2023-07-31 Well region planning method for simulated layout design

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310951773.7A CN116861835A (en) 2023-07-31 2023-07-31 Well region planning method for simulated layout design

Publications (1)

Publication Number Publication Date
CN116861835A true CN116861835A (en) 2023-10-10

Family

ID=88230519

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310951773.7A Pending CN116861835A (en) 2023-07-31 2023-07-31 Well region planning method for simulated layout design

Country Status (1)

Country Link
CN (1) CN116861835A (en)

Similar Documents

Publication Publication Date Title
CN108428229B (en) Lung texture recognition method based on appearance and geometric features extracted by deep neural network
CN104134234B (en) A kind of full automatic three-dimensional scene construction method based on single image
US7925084B1 (en) Merge and removal in a planar map of an image
CN103544711B (en) The autoegistration method of remote sensing image
CN104299263B (en) A kind of method that cloud scene is modeled based on single image
CN107833213A (en) A kind of Weakly supervised object detecting method based on pseudo- true value adaptive method
CN112733486B (en) Intelligent wiring method and system for chip design
CN110059697A (en) A kind of Lung neoplasm automatic division method based on deep learning
CN105574527A (en) Quick object detection method based on local feature learning
CA2358729C (en) Computer aided method of circuit extraction
CN110473174B (en) Method for calculating accurate number of pencils based on image
CN106484692A (en) A kind of method for searching three-dimension model
CN110263794B (en) Training method of target recognition model based on data enhancement
CN110598634B (en) Machine room sketch identification method and device based on graph example library
CN108197567B (en) Method, apparatus and computer readable medium for image processing
US7929755B1 (en) Planar map to process a raster image
CN112883926B (en) Identification method and device for form medical images
CN106447673A (en) Chip pin extraction method under non-uniform illumination condition
CN112446262A (en) Text analysis method, text analysis device, text analysis terminal and computer-readable storage medium
CN105068918A (en) Page test method and device
CN103500469B (en) A kind of urban afforestation modeling method
CN107145643A (en) The intelligent design system and design method of a kind of metal plate sculpture type face steelframe
CN107145888A (en) Video caption real time translating method
CN108197627A (en) A kind of method for extracting two-value graph outline along clockwise direction
CN116861835A (en) Well region planning method for simulated layout design

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination