CN111754480A - Method for retrieving and early warning wafer back defect map, storage medium and computer equipment - Google Patents
Method for retrieving and early warning wafer back defect map, storage medium and computer equipment Download PDFInfo
- Publication number
- CN111754480A CN111754480A CN202010577051.6A CN202010577051A CN111754480A CN 111754480 A CN111754480 A CN 111754480A CN 202010577051 A CN202010577051 A CN 202010577051A CN 111754480 A CN111754480 A CN 111754480A
- Authority
- CN
- China
- Prior art keywords
- defect map
- back defect
- crystal back
- database
- crystal
- 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.)
- Granted
Links
- 230000007547 defect Effects 0.000 title claims abstract description 288
- 238000000034 method Methods 0.000 title claims abstract description 70
- 238000003860 storage Methods 0.000 title claims abstract description 23
- 239000013078 crystal Substances 0.000 claims abstract description 167
- 238000013528 artificial neural network Methods 0.000 claims abstract description 10
- 238000012549 training Methods 0.000 claims description 21
- 238000012545 processing Methods 0.000 claims description 7
- 230000009467 reduction Effects 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 4
- 238000004804 winding Methods 0.000 claims description 2
- 238000010586 diagram Methods 0.000 abstract description 18
- 230000008569 process Effects 0.000 abstract description 16
- 238000001514 detection method Methods 0.000 abstract description 13
- 238000004590 computer program Methods 0.000 description 6
- 238000004519 manufacturing process Methods 0.000 description 6
- 239000004065 semiconductor Substances 0.000 description 6
- 238000011161 development Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000006872 improvement Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000005096 rolling process Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000011109 contamination Methods 0.000 description 1
- 230000006837 decompression Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000001459 lithography Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000000465 moulding Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000001259 photo etching Methods 0.000 description 1
- 229910021420 polycrystalline silicon Inorganic materials 0.000 description 1
- 229920005591 polysilicon Polymers 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30148—Semiconductor; IC; Wafer
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Quality & Reliability (AREA)
- Testing Or Measuring Of Semiconductors Or The Like (AREA)
Abstract
The invention provides a crystal back defect detection and early warning method, a storage medium and computer equipment.A known crystal back defect map in a database is trained through a self-encoder neural network to obtain a crystal back defect map retrieval model, high-dimensional characteristics of the known crystal back defect map are extracted by using the crystal back defect retrieval model, and the crystal back defect map retrieval model is encoded to generate a first encoding database; for the crystal back defect graph to be retrieved, retrieving model second encoding data by using the crystal back defect graph, and searching the second encoding data in the first encoding database by using a nearest neighbor algorithm; and obtaining whether the to-be-retrieved crystal back defect graph is an unknown defect early warning or a known defect type according to the existence of the second coded data in the first coded data base. Therefore, the labor cost is saved, the judgment time is greatly shortened, the efficiency is improved, the retrieval accuracy is improved based on the existing known crystal back defect diagram, the missing rate and the false alarm rate are reduced, and the important guarantee is provided for improving the yield of the subsequent process.
Description
Technical Field
The present invention relates to a semiconductor integrated circuit manufacturing method, and more particularly, to a method for retrieving and warning a defect map of a wafer back, a storage medium, and a computer device.
Background
In the semiconductor chip manufacturing process, with the continuous improvement of the integration level of the semiconductor chip, the size of the semiconductor device is smaller and smaller, and the molding requirement of the device is stricter and stricter. Particularly, the photoetching process requires that the crystal face of the wafer has good uniformity, and the crystal back of the wafer cannot have the defects of pollution, marks and the like. However, it is difficult to avoid a variety of backside defect maps when new technology nodes and/or new products are first produced, because backside defect maps may be generated for a variety of reasons, such as the process tool contacting a contact feature on the backside of the wafer, such as a backside chuck, which tends to mark the backside. The imprint may form a specific pattern residue as the subsequent process proceeds, and the pattern residue may form a ring-shaped polysilicon (Poly) residue, thereby affecting the lithography process on the crystal plane, forming defocus (defocus), and seriously affecting the product yield. Therefore, there is a need to find the contamination or defect of the wafer back accurately in time during the semiconductor manufacturing process.
In the prior art, the most commonly used method for detecting the defect map of the wafer back is as follows: the yield engineer uses a Scanning Electron Microscope (SEM) to distinguish the shape, size, composition, etc. of the backside defect map to find its cause. As can be appreciated, this method has the following drawbacks: firstly, by means of manual judgment, a yield engineer needing to be in charge of detection has rich experience, and the method is time-consuming, labor-consuming and low in efficiency; and secondly, the situation of misjudgment or omission is difficult to avoid by manual operation. And once a misjudgment or omission occurs, the yield may be low, and even the wafer may be scrapped.
Therefore, it is an urgent technical problem to be solved by those skilled in the art to provide a method for searching and early warning a backside defect map so as to automatically and accurately determine the backside defect map and quickly find the cause of the backside defect map.
It is noted that the information disclosed in this background section is only for enhancement of understanding of the general background of the invention and should not be taken as an acknowledgement or any form of suggestion that this information forms the prior art already known to a person skilled in the art.
Disclosure of Invention
The invention aims to provide a crystal back defect map retrieval and early warning method, a storage medium and computer equipment, which are used for solving the technical problems of low retrieval efficiency, missed judgment and wrong judgment in the crystal back defect map retrieval in the prior art.
In order to achieve the purpose, the invention is realized by the following technical scheme: a method for retrieving and early warning a crystal back defect picture comprises the following steps:
s1: taking a known crystal back defect map in a known crystal back defect map database as a training sample set, and training a self-coding neural network until a crystal back defect map retrieval model is obtained; respectively carrying out feature coding on each known crystal back defect map by using the crystal back defect map retrieval model to obtain a first coding database;
s2: performing the characteristic coding on the to-be-retrieved crystal back defect map by using the crystal back defect map model to obtain second coded data;
s3: by utilizing a nearest neighbor algorithm, judging whether the candidate crystal back defect map of the crystal back defect map to be searched exists in the known crystal back defect map database or not by searching the second coded data in the first coded data base;
if so, judging the defect type of the to-be-retrieved crystal back defect map according to the defect type of the candidate crystal back defect map;
if not, giving an early warning that the to-be-retrieved crystal back defect image is an unknown defect;
the crystal back defect map retrieval model meets the following conditions, feature decoding is carried out on each first encoding data in the first encoding database by using the crystal back defect map retrieval model to obtain a reconstructed crystal back defect map data set, and the error of each reconstructed crystal back defect map and the known crystal back defect map corresponding to the reconstructed crystal back defect map is within a first preset threshold range; wherein the feature decoding is an inverse operation of the feature encoding.
Optionally, before step S1, an image normalization process is further performed on each of the original pictures of the known back defect map, so as to obtain normalized pictures with uniform size and the same channel;
in step S1, the step of taking the known wafer back defect map in the known wafer back defect map database as a training sample set includes taking all the normalized pictures as the training sample set.
Optionally, the backside defect map retrieval model comprises an input layer, a residual layer and an output layer, the residual layer is respectively connected with the input layer and the output layer, wherein,
the input layer comprises a plurality of first convolution layers;
the residual layer comprises a plurality of residual sublayers, each residual sublayer is obtained by overlapping at least one convolution block and at least one identification block, and each convolution block comprises a plurality of second convolution layers and a convolution shortcut; each identification block comprises a plurality of third convolution layers;
the output layer comprises a plurality of fully connected layers.
Optionally, the first winding layer is two,
and/or
The number of the residual sub-layers is three, each residual sub-layer comprises a convolution block and an identification block, each convolution block comprises three second convolution layers and a convolution shortcut, and each identification block comprises three third convolution layers;
and/or
The number of the full connecting layers is four.
Optionally, in step S1, the performing feature encoding on each known back defect map by using the back defect map retrieval model to obtain a first encoding database, where the feature encoding method includes the following steps,
s11: the input layer receives the training sample set, reduces the dimension of the training sample set and obtains a first characteristic dimension data set with a characteristic dimension being a first dimension and a depth being a first depth;
s12: the residual error layer receives the first characteristic dimension data set, and performs dimensionality reduction on the first characteristic dimension data set to obtain a second characteristic dimension data set with a characteristic dimension being a second dimension and a depth being a second depth;
s13: and the output layer receives the second characteristic dimension data set, reduces the dimension of the second characteristic dimension data set to obtain a third characteristic dimension data set with a characteristic dimension being a third dimension and a depth being a third depth, and takes the third characteristic dimension data set as the first coding database.
Optionally, the second dimension is one eighth of the first dimension and/or the second depth is eight times the first depth.
Optionally, in step S3, the method for determining whether the candidate back defect map of the back defect map to be retrieved exists in the known back defect map database by retrieving the second encoded data from the first encoded data base by using a nearest neighbor algorithm includes,
calculating Euclidean distance, Hamming distance or cosine similarity between first coded data and second coded data in the first coded data set to obtain similarity between the to-be-retrieved crystal back defect map and the known crystal back defect map, so as to obtain a candidate crystal back defect map list of the to-be-retrieved crystal back defect map;
rearranging the candidate crystal back defect map list according to the size of the similarity, and if a candidate crystal back defect map with the similarity within a second preset threshold range exists, judging that the candidate crystal back defect map of the crystal back defect map to be retrieved exists in the known crystal back defect map database; otherwise, judging that the candidate crystal back defect map of the crystal back defect map to be retrieved does not exist in the known crystal back defect map database.
Alternatively, if it is determined in step S3 that there is no candidate backside defect map corresponding to the known backside defect map database, the following steps are performed,
and amplifying the known crystal back defect map database by taking the crystal back defect map to be retrieved as a known crystal back defect map, and upgrading the crystal back defect map retrieval model and the first encoding database by using the same method as the step S1.
Based on the same inventive concept, the invention further provides a computer-readable storage medium, wherein computer-executable instructions are stored on the computer-readable storage medium, and when the computer-executable instructions are executed, the method for retrieving and early warning the backside defect map is implemented.
Based on the same inventive concept, the present invention further provides a computer device, which includes a processor and a storage device, wherein the processor is adapted to implement each instruction, the storage device is adapted to store a plurality of instructions, and the instructions are adapted to be loaded by the processor and to execute the method for retrieving and warning the backside defect map according to any one of claims 1 to 8.
Compared with the prior art, the crystal back defect detection and early warning method provided by the invention has the following beneficial effects:
according to the crystal back defect detection and early warning method provided by the invention, a large number of existing known crystal back defect maps in a database are trained through a self-encoder neural network, and the generated crystal back defect map retrieval model can quickly and accurately extract high-dimensional features of the known crystal back defect maps and encode the high-dimensional features to generate a first encoding database; for a newly generated crystal back defect graph to be retrieved, generating second coded data with characteristics by using the crystal back defect graph retrieval model, and searching the second coded data in the first coded database by using a nearest neighbor algorithm; whether the second coded data exist in the first coded data base or not is judged, if not, the to-be-retrieved crystal back defect graph is given as an unknown defect early warning, and therefore relevant technicians can track the root of the problem; if the existing defect type exists, the defect type of the to-be-retrieved crystal back defect map is determined according to the corresponding defect type of the known crystal back defect map in the known crystal back defect map database, so that the current manual judgment mode is changed into full-automatic judgment, the labor cost is saved, the judgment time is greatly shortened, the efficiency is improved, in addition, the retrieval accuracy is improved based on the existing known crystal back defect map, the omission ratio and the false alarm ratio are reduced, and the important guarantee is provided for improving the yield of the subsequent process.
Further, according to the method for retrieving and early warning the crystal back defect map provided by the invention, for the crystal back defect map to be retrieved which does not exist in the known crystal back defect map database, the crystal back defect map to be retrieved is further used as the known crystal back defect map to be amplified to the known crystal back defect map database, and the same method as that in step S1 is used for upgrading the crystal back defect map retrieval model and the first encoding database. Therefore, the crystal back defect retrieval model has good openness, can be continuously updated and expanded along with the types of the discovered defects of the crystal back defect map, is further suitable for continuous development of new technical nodes and rapid development of related equipment technologies, and has good adaptability.
Because the readable storage medium and the computer device provided by the invention belong to the same inventive concept as the crystal back defect map retrieval and early warning method, the readable storage medium and the computer device at least have the same beneficial effects and are not repeated.
Drawings
Fig. 1 is a schematic flow chart of a method for retrieving and warning a defect map of a wafer back according to an embodiment of the present invention;
FIG. 2 is a schematic structural diagram of a backside defect map retrieval model shown in FIG. 1;
FIG. 3 is a schematic diagram of a data processing process of the method for retrieving and pre-warning a defect map of a wafer back in FIG. 1;
FIG. 3A is a schematic diagram of the feature encoding process of FIG. 3;
FIG. 3B is a diagram illustrating results of one implementation of the nearest neighbor algorithm of FIG. 3;
FIG. 4 is a schematic diagram of a candidate wafer back defect map list and a picture of the wafer back defect map to be detected corresponding to FIG. 3B;
FIG. 5 is a diagram illustrating a relationship between the number of similar conventional backside defect maps and associated detection indicators in one embodiment of FIG. 3;
wherein the reference numerals are as follows:
100-input layer, 200-residual layer, 300-output layer.
Detailed Description
In order to make the objects, advantages and features of the present invention clearer, the method for retrieving and warning the defect map of the backside of a crystal, the storage medium and the computer device proposed by the present invention will be described in further detail with reference to the accompanying drawings. It will be apparent that the methods described herein comprise a series of steps and that the order of such steps presented herein is not necessarily the only order in which such steps may be performed, and that some of the described steps may be omitted and/or some other steps not described herein may be added to the methods. Further, the described embodiments are only some embodiments of the invention, not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
To facilitate an understanding of the present invention, a brief description of a source of a database of known wafer back defect maps is provided before describing embodiments of the present invention in detail. The Klary Defect system is widely used in the manufacture and production of semiconductors as a Defect monitoring system. After the Defect detection machine detects the Defect signal, the wafer with the excessive Defect number needs to be further inspected on the Defect inspection machine to obtain specific photos of the defects and classify the photos, and after a yield engineer searches the Defect reasons, the original photos of the wafer back Defect map and the Defect reasons (types) are stored in a Klary Defect server through a Klary Defect System, so that the known wafer back Defect map database not only comprises known wafer back Defect maps, but also comprises Defect types corresponding to each known wafer back Defect map.
In order to shorten the retrieval efficiency of the wafer back defect map and reduce the missing report rate and the missing detection rate of the wafer back defect map, the inventor provides a wafer back defect map retrieval and early warning method based on a known wafer back defect map database through intensive research and continuous practice. The method for retrieving and early warning the crystal back defect map is based on the known crystal back defect map database and the self-coding neural network, so that the retrieval efficiency of the crystal back defects is shortened, and the missing report rate and the missing detection rate of the crystal back defect retrieval in the prior art are reduced. For ease of understanding, the core ideas of the self-coding neural network are briefly described as follows: a neural network model is constructed, the model comprises a model input layer, a model hidden layer and a model output layer, the model collects a picture, the neural network compresses the picture after receiving the picture, and the picture is finally restored from the compressed picture, namely the picture of the self-coding model is a process of decompression after being compressed. When compressing, the quality of the original picture is reduced (dimensionality reduction), and when decompressing, the original picture is restored by a file with small information quantity but containing all information (restoration). Now, it is assumed that information a in the model input layer is decompressed into a hidden layer to obtain a, then a of the model hidden layer is compared with a of the input layer to obtain a prediction error, then reverse transmission is performed, then the accuracy of self-coding is gradually improved, and a part of data a obtained in the middle hidden layer after a period of training is the essence of source data. Self-coding is an unsupervised learning, and the first half (dimension reduction) of self-coding, namely a coding process, can obtain the main characteristics of source data.
Referring to fig. 1 and fig. 3, an embodiment of a method for retrieving and warning a defect map of a wafer back according to the present invention includes the following steps:
s1: training to obtain a crystal back defect map retrieval model and using the model to obtain a first coding database: taking a known crystal back defect map in a known crystal back defect map database as a training sample set, and training a self-coding neural network until a crystal back defect map retrieval model is obtained; and respectively carrying out characteristic coding on each known crystal back defect map by using the crystal back defect map retrieval model to obtain a first coding database.
S2: and performing the characteristic coding on the to-be-retrieved crystal back defect map by using the crystal back defect map model to obtain second coded data.
S3: and judging whether the candidate crystal back defect map of the crystal back defect map to be searched exists in the known crystal back defect map database or not by searching the second coded data in the first coded data base by utilizing a nearest neighbor algorithm. If so, judging the defect type of the to-be-retrieved crystal back defect map according to the defect type of the candidate crystal back defect map; and if not, giving an early warning that the to-be-retrieved crystal back defect image is an unknown defect. The crystal back defect map retrieval model meets the following conditions, feature decoding is carried out on each first encoding data in the first encoding database by using the crystal back defect map retrieval model to obtain a reconstructed crystal back defect map data set, and the error of each reconstructed crystal back defect map and the known crystal back defect map corresponding to the reconstructed crystal back defect map is within a first preset threshold range; wherein the feature decoding is an inverse operation of the feature encoding.
Preferably, in one embodiment, before step S1, an image normalization process is further performed on the original picture of each of the known back defect maps, so as to obtain a normalized picture with a uniform size and the same channel. Therefore, in step S1, the step of taking the known wafer back defect map in the database of known wafer back defect maps as a training sample set includes taking all the normalized pictures as the training sample set. For example, if the original picture of the back defect map is a 1024 × 1024 picture with three RGB channels, in one of the modes, after the image normalization processing is performed, the obtained normalized picture is a 128 × 128 single-channel picture, and then is input into the self-encoding neural network in a tensor form.
Preferably, in an embodiment, referring to fig. 3A, in step S1, the feature coding is performed on each of the known back defect maps by using the back defect map retrieval model to obtain a first coding database, where the feature coding method includes the following steps:
s11: the input layer 100 receives the training samples, and performs dimensionality reduction on the training sample set to obtain a first feature dimension data set with a feature dimension being a first dimension and a depth being a first depth.
S12: the residual layer 200 receives the first feature dimension data set, and performs dimensionality reduction on the first feature dimension data set to obtain a second feature dimension data set with a feature dimension of a second dimension and a depth of a second depth.
S13: the output layer 200 receives the second feature dimension data set, and performs dimensionality reduction on the second feature dimension data set to obtain a third feature dimension data set with a feature dimension being a third dimension. Until the third feature dimension data set is used as the first encoding database. According to the above description, as the back defect map retrieval model needs to satisfy the following conditions, feature decoding is performed on each first encoded data in the first encoded data base by using the back defect map retrieval model to obtain a reconstructed back defect map data set, and an error between each reconstructed back defect map and the known back defect map corresponding to the reconstructed back defect map is within a first preset threshold range; therefore, the above steps S11, S12 and S13 feature encoding process is also a model building process.
Preferably, referring to fig. 2, in one embodiment, the backside defect map retrieval model includes an input layer 100, a residual layer 200, and an output layer 300, wherein the residual layer 200 is respectively connected to the input layer 100 and the output layer 200. Wherein the input layer 100 comprises a plurality of first convolution layers; the residual error layer 200 includes a plurality of residual error sub-layers, each of which is obtained by overlapping at least one convolution block and at least one identification block, wherein each convolution block includes a plurality of second convolution layers and a convolution shortcut; each identification block comprises a plurality of third convolution layers; the output layer 300 includes a number of fully connected layers. Preferably, referring to fig. 2, there are two first convolution layers, there are three residual sub-layers, each of the residual sub-layers includes a convolution block and an identification block, each of the convolution blocks includes three second convolution layers and a convolution shortcut, and each of the identification blocks includes three third convolution layers; the number of the full connecting layers is four. Obviously, the above description is only for the preferred embodiment, and not for the limitation of the present invention, and it should be understood that the number of the first convolutional layers, the number of the residual sub-layers, the number of the convolutional blocks, the number of the second convolutional layers, and the number of the third convolutional layers may be set according to practical situations, but all of them are within the protection scope of the present invention. Further, in one embodiment, after passing through the input layer 100, the data in the training sample set is reduced to a first feature dimension data set with a first dimension of 32 × 32 and a first depth of 64; the residual layer 200 receives the first feature dimension data set, after passing through three residual sublayers, after the rolling blocks of each residual sublayer, each rolling block reduces the feature dimension of the data in the first feature dimension data set to one half of the original dimension, and each identification block does not change the feature dimension of the data in the first feature dimension data set, so that the feature dimension of the data in the first feature dimension data set is reduced to one half of the original dimension and the depth feature becomes twice of the original dimension, so that the feature dimension is 32 × 32 after the data passes through one residual sublayer, the feature dimension is reduced to a second dimension of 4 × 4 after the first feature dimension data set with a first depth of 64 passes through the residual layer 200, and the second feature dimension data set with a second depth of 1024. The output layer 300 performs data feature unidimensional on the second feature dimension data set, and after passing through the four full connection layers, the feature dimension is reduced to 100, that is, the third feature dimension data set with the feature dimension of 100 is the first encoding database.
Preferably, in an embodiment, referring to fig. 3B to fig. 5, a method for determining whether a candidate back defect map of the back defect map to be retrieved exists in the known back defect map database by retrieving the second encoded data in the first encoded data database by using a nearest neighbor algorithm includes calculating a euclidean distance between the first encoded data and the second encoded data in the first encoded data set to obtain a similarity between the back defect map to be retrieved and the known back defect map, so as to obtain a candidate back defect map list of the back defect map to be retrieved; rearranging the candidate crystal back defect map list according to the size of the similarity, and if a candidate crystal back defect map with the similarity within a second preset threshold range exists, judging that the candidate crystal back defect map of the crystal back defect map to be retrieved exists in the known crystal back defect map database; otherwise, judging that the candidate crystal back defect map of the crystal back defect map to be retrieved does not exist in the known crystal back defect map database. As shown in fig. 3B, as a result of the classification in one embodiment, obviously, in other embodiments, the similarity between the back defect map to be retrieved and the known back defect map may also be obtained by calculating a hamming distance or a cosine similarity between the first encoded data and the second encoded data in the first encoded data set, which is not limited in the present invention, but is within the protection scope of the present invention. As shown in fig. 4, it is a schematic diagram of the picture of the to-be-detected back defect map and the list of the most similar 4 candidate back defect maps shown in fig. 3B; fig. 5 is a schematic diagram illustrating a relationship between the number of similar known back defect maps and related detection indexes in one embodiment of fig. 3, and it can be seen from the diagram that by using the back defect detection and warning method provided by the present invention, there is no missing detection (the missing detection rate is 0%), the false alarm rate is much lower than 20%, and the detected defect is an abnormal defect and is recalled by 100%. It should be noted that the foregoing is only a description of the preferred embodiment, and is not a limitation of the present invention, the nearest neighbor algorithm may use a K nearest neighbor basic algorithm, or may use an improved algorithm thereof, and in other embodiments, the similarity between the backside defect map to be retrieved and the backside defect map may also be obtained by calculating a manhattan distance between the first encoded data and the second encoded data in the first encoded data set.
With continued reference to fig. 1, preferably, in one embodiment, if it is determined in step S3 that there is no candidate back defect map corresponding to the known back defect map database, the following steps are performed,
s4: and amplifying the known crystal back defect map database by taking the crystal back defect map to be retrieved as a known crystal back defect map, and upgrading the crystal back defect map retrieval model and the first encoding database by using the same method as the step S1.
Therefore, according to the crystal back defect detection and early warning method provided by the invention, a large number of existing known crystal back defect maps in a database are trained through a self-encoder neural network, and the generated crystal back defect map retrieval model can quickly and accurately extract high-dimensional features of the known crystal back defect maps and encode the high-dimensional features to generate a first encoding database; for a newly generated crystal back defect graph to be retrieved, generating second coded data with characteristics by using the crystal back defect graph retrieval model, and searching the second coded data in the first coded database by using a nearest neighbor algorithm; whether the second coded data exist in the first coded data base or not is judged, if not, the to-be-retrieved crystal back defect graph is given as an unknown defect early warning, and therefore relevant technicians can track the root of the problem; if the existing defect type exists, the defect type of the to-be-retrieved crystal back defect map is determined according to the corresponding defect type of the known crystal back defect map in the known crystal back defect map database, so that the current manual judgment mode is changed into full-automatic judgment, the labor cost is saved, the judgment time is greatly shortened, the efficiency is improved, in addition, the retrieval accuracy is improved based on the existing known crystal back defect map, the omission ratio and the false alarm ratio are reduced, and the important guarantee is provided for improving the yield of the subsequent process.
Further, according to the method for retrieving and early warning the crystal back defect map provided by the invention, for the crystal back defect map to be retrieved which does not exist in the known crystal back defect map database, the crystal back defect map to be retrieved is further used as the known crystal back defect map to be amplified to the known crystal back defect map database, and the same method as that in step S1 is used for upgrading the crystal back defect map retrieval model and the first encoding database. Therefore, the crystal back defect retrieval model has good openness, can be continuously updated and expanded along with the types of the discovered defects of the crystal back defect map, is further suitable for continuous development of new technical nodes and rapid development of related equipment technologies, and has good adaptability.
Based on the same inventive concept, yet another embodiment of the present invention further provides a computer-readable storage medium, where computer-executable instructions are stored on the computer-readable storage medium, and when the computer-executable instructions are executed, the steps of the method for retrieving and warning a wafer back defect map and the method for warning a wafer back defect map are implemented, where the specific steps are already described in detail above, and are not repeated here.
Based on the same inventive concept, yet another embodiment of the present invention further provides a computer device, where the computer device includes a processor and a storage device, the processor is adapted to implement each instruction, the storage device is adapted to store a plurality of instructions, and the instructions are adapted to be loaded by the processor and to execute the method for retrieving and warning a backside defect map according to any one of the foregoing embodiments.
From the above description of embodiments, it should be apparent to those skilled in the art that embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects that is presently preferred. With this understanding in mind, portions of the present solution that contribute to the prior art can be embodied in the form of a computer software product that is stored on a computer-readable storage medium, which includes but is not limited to disk storage, CD-ROM, optical storage, and the like.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Since the readable storage medium and the computer device provided by the embodiment of the invention belong to the same inventive concept as the above-mentioned crystal back defect map retrieval and early warning method, at least the same beneficial effects are achieved, and the details are not repeated.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention in any way, and the present invention includes, but is not limited to, the configurations listed in the above embodiments. Various modifications and alterations to the embodiments of the present invention will become apparent to those skilled in the art from the foregoing description of the embodiments. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A crystal back defect map retrieval and early warning method is characterized by comprising the following steps:
s1: taking a known crystal back defect map in a known crystal back defect map database as a training sample set, and training a self-coding neural network until a crystal back defect map retrieval model is obtained; respectively carrying out feature coding on each known crystal back defect map by using the crystal back defect map retrieval model to obtain a first coding database;
s2: performing the characteristic coding on the to-be-retrieved crystal back defect map by using the crystal back defect map model to obtain second coded data;
s3: by utilizing a nearest neighbor algorithm, judging whether the candidate crystal back defect map of the crystal back defect map to be searched exists in the known crystal back defect map database or not by searching the second coded data in the first coded data base;
if so, judging the defect type of the to-be-retrieved crystal back defect map according to the defect type of the candidate crystal back defect map;
if not, giving an early warning that the to-be-retrieved crystal back defect image is an unknown defect;
the crystal back defect map retrieval model meets the following conditions, feature decoding is carried out on each first encoding data in the first encoding database by using the crystal back defect map retrieval model to obtain a reconstructed crystal back defect map data set, and the error of each reconstructed crystal back defect map and the known crystal back defect map corresponding to the reconstructed crystal back defect map is within a first preset threshold range; wherein the feature decoding is an inverse operation of the feature encoding.
2. The method for retrieving and pre-warning the wafer back defect map according to claim 1, wherein before the step S1, the method further comprises the steps of performing image normalization processing on each of the original pictures of the known wafer back defect map to obtain normalized pictures with uniform size and same channel;
in step S1, the step of taking the known wafer back defect map in the known wafer back defect map database as a training sample set includes taking all the normalized pictures as the training sample set.
3. The method of claim 1, wherein the CRI model comprises an input layer, a residual layer and an output layer, the residual layer is respectively connected to the input layer and the output layer, wherein,
the input layer comprises a plurality of first convolution layers;
the residual layer comprises a plurality of residual sublayers, each residual sublayer is obtained by overlapping at least one convolution block and at least one identification block, and each convolution block comprises a plurality of second convolution layers and a convolution shortcut; each identification block comprises a plurality of third convolution layers;
the output layer comprises a plurality of fully connected layers.
4. The method of claim 3, wherein the defect map of the backside is retrieved and pre-warned,
the number of the first winding layers is two,
and/or
The number of the residual sub-layers is three, each residual sub-layer comprises a convolution block and an identification block, each convolution block comprises three second convolution layers and a convolution shortcut, and each identification block comprises three third convolution layers;
and/or
The number of the full connecting layers is four.
5. The method of claim 3, wherein in step S1, the step of using the back defect map retrieval model to perform feature coding on each of the known back defect maps to obtain a first coding database, wherein the feature coding method comprises the steps of,
s11: the input layer receives the training sample set, reduces the dimension of the training sample set and obtains a first characteristic dimension data set with a characteristic dimension being a first dimension and a depth being a first depth;
s12: the residual error layer receives the first characteristic dimension data set, and performs dimensionality reduction on the first characteristic dimension data set to obtain a second characteristic dimension data set with a characteristic dimension being a second dimension and a depth being a second depth;
s13: and the output layer receives the second characteristic dimension data set, reduces the dimension of the second characteristic dimension data set to obtain a third characteristic dimension data set with a characteristic dimension being a third dimension and a depth being a third depth, and takes the third characteristic dimension data set as the first coding database.
6. The method of claim 5, wherein the second dimension is one eighth of the first dimension and/or the second depth is eight times the first depth.
7. The method for retrieving and warning a wafer back defect map according to claim 1, wherein in step S3, a nearest neighbor algorithm is used to determine whether the candidate wafer back defect map of the wafer back defect map to be retrieved exists in the known wafer back defect map database by retrieving the second encoded data from the first encoded data database, comprising,
calculating Euclidean distance, Hamming distance or cosine similarity between first coded data and second coded data in the first coded data set to obtain similarity between the to-be-retrieved crystal back defect map and the known crystal back defect map, so as to obtain a candidate crystal back defect map list of the to-be-retrieved crystal back defect map;
rearranging the candidate crystal back defect map list according to the size of the similarity, and if a candidate crystal back defect map with the similarity within a second preset threshold range exists, judging that the candidate crystal back defect map of the crystal back defect map to be retrieved exists in the known crystal back defect map database; otherwise, judging that the candidate crystal back defect map of the crystal back defect map to be retrieved does not exist in the known crystal back defect map database.
8. The method of claim 1, wherein if it is determined in step S3 that there is no candidate backside defect map corresponding to the known backside defect map database, the following steps are performed,
and amplifying the known crystal back defect map database by taking the crystal back defect map to be retrieved as a known crystal back defect map, and upgrading the crystal back defect map retrieval model and the first encoding database by using the same method as the step S1.
9. A computer-readable storage medium having computer-executable instructions stored thereon, wherein the computer-executable instructions, when executed, implement the backside defect map retrieval and warning method of any one of claims 1 to 8.
10. A computer device comprising a processor and a storage device, wherein the processor is adapted to implement instructions, and the storage device is adapted to store a plurality of instructions, and the instructions are adapted to be loaded by the processor and to perform the method for retrieving and warning a backside defect map according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010577051.6A CN111754480B (en) | 2020-06-22 | 2020-06-22 | Crystal back defect map retrieval and early warning method, storage medium and computer equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010577051.6A CN111754480B (en) | 2020-06-22 | 2020-06-22 | Crystal back defect map retrieval and early warning method, storage medium and computer equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111754480A true CN111754480A (en) | 2020-10-09 |
| CN111754480B CN111754480B (en) | 2024-04-16 |
Family
ID=72676432
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010577051.6A Active CN111754480B (en) | 2020-06-22 | 2020-06-22 | Crystal back defect map retrieval and early warning method, storage medium and computer equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111754480B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112907542A (en) * | 2021-02-24 | 2021-06-04 | 上海华力集成电路制造有限公司 | Method for detecting defects of wafer back, storage medium and computer device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109885709A (en) * | 2019-01-08 | 2019-06-14 | 五邑大学 | A kind of image search method, device and storage medium based on from the pre- dimensionality reduction of coding |
| CN109977808A (en) * | 2019-03-11 | 2019-07-05 | 北京工业大学 | A kind of wafer surface defects mode detection and analysis method |
| CN110083729A (en) * | 2019-04-26 | 2019-08-02 | 北京金山数字娱乐科技有限公司 | A kind of method and system of picture search |
| US20190273948A1 (en) * | 2019-01-08 | 2019-09-05 | Intel Corporation | Method and system of neural network loop filtering for video coding |
-
2020
- 2020-06-22 CN CN202010577051.6A patent/CN111754480B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109885709A (en) * | 2019-01-08 | 2019-06-14 | 五邑大学 | A kind of image search method, device and storage medium based on from the pre- dimensionality reduction of coding |
| US20190273948A1 (en) * | 2019-01-08 | 2019-09-05 | Intel Corporation | Method and system of neural network loop filtering for video coding |
| CN109977808A (en) * | 2019-03-11 | 2019-07-05 | 北京工业大学 | A kind of wafer surface defects mode detection and analysis method |
| CN110083729A (en) * | 2019-04-26 | 2019-08-02 | 北京金山数字娱乐科技有限公司 | A kind of method and system of picture search |
Non-Patent Citations (2)
| Title |
|---|
| 周纤;邱奕敏;吴振宇;: "基于栈式卷积自编码的哈希图像检索研究", 电视技术, no. 10 * |
| 易晨晖;何蓉;: "北斗电文编码与解码的研究", 电子世界, no. 17 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112907542A (en) * | 2021-02-24 | 2021-06-04 | 上海华力集成电路制造有限公司 | Method for detecting defects of wafer back, storage medium and computer device |
| CN112907542B (en) * | 2021-02-24 | 2024-05-03 | 上海华力集成电路制造有限公司 | Crystal back defect detection method, storage medium and computer equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111754480B (en) | 2024-04-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20200402221A1 (en) | Inspection system, image discrimination system, discrimination system, discriminator generation system, and learning data generation device | |
| CN111179263B (en) | Industrial image surface defect detection model, method, system and device | |
| CN106971953B (en) | Error detection method in manufacturing process | |
| US9142014B2 (en) | System and method for identifying systematic defects in wafer inspection using hierarchical grouping and filtering | |
| CN117173172B (en) | Machine vision-based silica gel molding effect detection method and system | |
| CN112381790A (en) | Abnormal image detection method based on depth self-coding | |
| CN111754480B (en) | Crystal back defect map retrieval and early warning method, storage medium and computer equipment | |
| CN112529109A (en) | Unsupervised multi-model-based anomaly detection method and system | |
| CN114800041B (en) | Cutter state monitoring method and monitoring device thereof | |
| KR101910268B1 (en) | Semiconductor GP Prediction Method and System | |
| CN115905990A (en) | Transformer oil temperature abnormity monitoring method based on density aggregation algorithm | |
| CN116090558A (en) | Method for generating knowledge points based on SPC numerical sequence data | |
| TWI801973B (en) | Device and method for estimating error factors | |
| CN114998216A (en) | Method and device for rapidly detecting surface defects of transparent part | |
| CN117237683B (en) | Chip defect intelligent detection system based on improved neural network | |
| CN114299040A (en) | Ceramic tile flaw detection method and device and electronic equipment | |
| CN113888477A (en) | Network model training method, metal surface defect detection method and electronic equipment | |
| CN117523363A (en) | Wafer map defect mode identification method based on feature pyramid fusion | |
| CN116304963B (en) | Data processing system suitable for geological disaster early warning | |
| CN115082650A (en) | Implementation method of automatic pipeline defect labeling tool based on convolutional neural network | |
| CN115423737A (en) | Wafer map defect detection method based on multiple principal component analysis networks | |
| Matsuoka et al. | Crack detection using spectral clustering based on crack features | |
| US6968280B2 (en) | Method for analyzing wafer test parameters | |
| KR102182678B1 (en) | Method and appratus for predicting fault pattern using multi-classifier based on feature selection method in semiconductor manufacturing process | |
| CN111723851A (en) | Production line fault detection method |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |