CN115239719A - Defect detection method, system, electronic device and storage medium - Google Patents
Defect detection method, system, electronic device and storage medium Download PDFInfo
- Publication number
- CN115239719A CN115239719A CN202211154884.7A CN202211154884A CN115239719A CN 115239719 A CN115239719 A CN 115239719A CN 202211154884 A CN202211154884 A CN 202211154884A CN 115239719 A CN115239719 A CN 115239719A
- Authority
- CN
- China
- Prior art keywords
- image
- defect
- model
- image segmentation
- detected
- 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
Links
- 230000007547 defect Effects 0.000 title claims abstract description 158
- 238000001514 detection method Methods 0.000 title claims abstract description 56
- 238000003860 storage Methods 0.000 title claims abstract description 12
- 238000003709 image segmentation Methods 0.000 claims abstract description 109
- 238000012549 training Methods 0.000 claims abstract description 34
- 238000000034 method Methods 0.000 claims abstract description 14
- 230000002950 deficient Effects 0.000 claims description 26
- 238000004088 simulation Methods 0.000 claims description 19
- 238000004364 calculation method Methods 0.000 claims description 15
- 238000002372 labelling Methods 0.000 claims description 15
- 230000006870 function Effects 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 8
- 230000002159 abnormal effect Effects 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 6
- 235000012431 wafers Nutrition 0.000 abstract description 24
- 230000000007 visual effect Effects 0.000 abstract description 3
- 238000004891 communication Methods 0.000 description 10
- 238000004519 manufacturing process Methods 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 239000004065 semiconductor Substances 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 238000009826 distribution Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000005457 optimization Methods 0.000 description 2
- 239000002245 particle Substances 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000011179 visual inspection Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000002146 bilateral effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000012797 qualification Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 230000017105 transposition Effects 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
- 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
-
- 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
- G06N3/08—Learning methods
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/761—Proximity, similarity or dissimilarity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural 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
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Databases & Information Systems (AREA)
- Medical Informatics (AREA)
- Multimedia (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Quality & Reliability (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Testing Or Measuring Of Semiconductors Or The Like (AREA)
Abstract
The invention provides a defect detection method, a system, electronic equipment and a storage medium, belonging to the technical field of visual detection of defects on the surface of a wafer; the method comprises the steps of constructing a defect-free image sample set from a plurality of acquired defect-free images, training a preset image reconstruction model based on the defect-free image sample set to obtain a target image reconstruction model, inputting a plurality of to-be-detected defect images into the target image reconstruction model to reconstruct the images, constructing an image segmentation annotation data set based on the result of image reconstruction, training the preset image segmentation model based on the image segmentation annotation data set to obtain a target image segmentation model, acquiring to-be-detected images of wafers, inputting the to-be-detected images into the target image segmentation model, and outputting a defect binary image in real time through the target image segmentation model. The defect binary image can be output in real time by the aid of the method and the device.
Description
Technical Field
The invention belongs to the technical field of visual detection of defects on the surface of a wafer, and particularly relates to a defect detection method, a defect detection system, electronic equipment and a storage medium.
Background
In order to improve the yield of chips, the whole semiconductor wafer can be scanned by using corresponding equipment in the production process of the semiconductor wafer, and abnormal graphs on the scanned images are analyzed to locate the defect positions on the wafer which can cause abnormal work. During the chip manufacturing process, a plurality of inspection processes are set up to identify defects on the wafer surface in time. The detection of the defects on the surface of the wafer can be realized by a manual mode or a machine vision mode.
With the appearance and development of artificial intelligence and deep learning technology, the object surface defect detection technology based on the vision technology is greatly improved, and the wafer surface defect detection method based on computer vision plays an important role in intelligent manufacturing. In the visual inspection of the surface defects of the wafer, in order to obtain the positioning result of the defects at the pixel level, the method based on the segmentation is gradually the mainstream of the visual inspection of the surface defects of the wafer. More accurate defect descriptions also facilitate subsequent defect qualification and recovery operations.
At present, a segmentation network structure used in the field of wafer surface defect detection is mainly used for designing an image segmentation model from the perspective of multi-scale and light weight. However, before the image segmentation model is used, sample training through pixel-level labeling is usually required, and these pixel-level labeling often have the problems of higher labeling cost, poorer labeling precision and the like, which means that labor and cost are greatly consumed, and the detection efficiency of the wafer surface defects is greatly influenced.
Disclosure of Invention
In order to solve the technical problems, the invention provides a defect detection method, a system, equipment and a storage medium, by optimizing a preset image reconstruction model and a preset image segmentation model, a defect binary image can be output in real time by the optimized preset image segmentation model of an image to be detected only by collecting non-defect samples during training the preset image reconstruction model without complex pixel level labeling, so that the purposes of low-cost labeling and high-precision and high-real-time wafer surface defect detection are achieved.
In a first aspect, the present invention provides a defect detection method for performing defect detection on a wafer, where the defect detection method includes:
constructing a defect-free image sample set from the acquired defect-free images;
training a preset image reconstruction model based on the defect-free image sample set to obtain a target image reconstruction model;
inputting a plurality of to-be-detected defect images into the target image reconstruction model for image reconstruction, and constructing an image segmentation and annotation data set based on the result of image reconstruction;
training a preset image segmentation model based on the image segmentation annotation data set to obtain a target image segmentation model;
acquiring an image to be detected of a wafer;
and inputting the image to be detected into the target image segmentation model, and outputting a defect binary image in real time through the target image segmentation model.
Preferably, the step of training a preset image reconstruction model based on the defect-free image sample set to obtain a target image reconstruction model includes:
performing defect simulation on the defect-free image sample set to obtain a defect simulation sample set;
inputting the defect simulation sample set to a preset image reconstruction model to obtain a reconstructed image;
determining a value of a loss function according to the loss of structural similarity calculated by the plurality of non-defective images and the reconstructed image;
and iteratively optimizing the preset image reconstruction model based on the value of the loss function to obtain a target image reconstruction model.
Preferably, the preset image reconstruction model comprises an FCN model or a UNet model.
Preferably, the loss function is specifically:
wherein (x, y) represents a pixel position, μ x Is the average value of x, μ y Is the mean value of y, σ 2 x Is the variance of x, σ 2 y Is the variance of y, σ xy Is the covariance of x and y, c 1 =(k 1 L) 2 、c 2 =(k 2 L) 2 L is the dynamic range of the pixel value, k 1 =0.01、k 2 =0.03。
Preferably, the step of inputting a plurality of defect images to be detected into the target image reconstruction model for image reconstruction, and constructing an image segmentation and annotation data set based on the result of image reconstruction specifically includes:
inputting a plurality of to-be-detected defect images into the target image reconstruction model to obtain a defect-free reconstructed image;
performing one-by-one difference calculation on the pixels of the non-defective reconstructed image and the to-be-detected defective image, and marking the pixel corresponding to the difference value exceeding a preset threshold value as an abnormal pixel to obtain a binary defective image;
and when the calculation efficiency of the target image reconstruction model does not meet the application scene, taking the binary defect image as the labeling result of the preset image segmentation model to obtain an image segmentation labeling data set.
Preferably, the image segmentation annotation data set comprises the defect image to be detected and the binary defect image paired with the defect image to be detected.
Preferably, the step of training a preset image segmentation model based on the image segmentation labeling data set to obtain a target image segmentation model specifically includes:
inputting the image with the defect to be detected into a preset image segmentation model to obtain model output data; wherein the preset image segmentation model comprises a DDRNet model;
calculating pixel-level cross entropy loss according to the image segmentation annotation data set and the model output data;
and iteratively optimizing the preset image segmentation model based on the pixel-level cross entropy loss to obtain a target image segmentation model.
In a second aspect, the invention provides a defect detection system for performing defect detection on a wafer; the defect detection system includes:
the first construction module is used for constructing a defect-free image sample set from the acquired defect-free images;
the first training module is used for training a preset image reconstruction model based on the defect-free image sample set to obtain a target image reconstruction model;
the second construction module is used for inputting a plurality of images with defects to be detected into the target image reconstruction model for image reconstruction and constructing an image segmentation and annotation data set based on the result of the image reconstruction;
the second training module is used for training a preset image segmentation model based on the image segmentation annotation data set to obtain a target image segmentation model;
the image acquisition module is used for acquiring an image to be detected of the wafer;
and the real-time output module is used for inputting the image to be detected to the target image segmentation model and outputting a defect binary image in real time through the target image segmentation model.
Preferably, the first training module comprises:
the simulation unit is used for carrying out defect simulation on the defect-free image sample set to obtain a defect simulation sample set;
the first reconstruction unit is used for inputting the defect simulation sample set into a preset image reconstruction model to obtain a reconstructed image;
a first calculation unit for calculating a loss of structural similarity from the plurality of non-defective images and the reconstructed image to determine a value of a loss function;
and the first iteration unit is used for iteratively optimizing the preset image reconstruction model based on the loss function to obtain a target image reconstruction model.
Preferably, the second building block comprises:
the second reconstruction unit is used for inputting a plurality of to-be-detected defect images into the target image reconstruction model to obtain a defect-free reconstructed image;
the marking unit is used for performing one-by-one difference calculation on the pixels of the non-defective reconstructed image and the to-be-detected defective image, and marking the pixels corresponding to the difference value exceeding a preset threshold value as abnormal pixels so as to obtain a binary defective image;
and the defining unit is used for taking the binary defect image as an annotation result of the preset image segmentation model to obtain an image segmentation annotation data set when the calculation efficiency of the target image reconstruction model does not meet the application scene.
Preferably, the second training module comprises:
the output unit is used for inputting the to-be-detected defect image into a preset image segmentation model to obtain model output data; wherein the preset image segmentation model comprises a DDRNet model;
the second calculation unit is used for calculating pixel-level cross entropy loss according to the image segmentation and labeling data set and the model output data;
and the second iteration unit is used for iteratively optimizing the preset image segmentation model based on the pixel-level cross entropy loss to obtain a target image segmentation model.
In a third aspect, an embodiment of the present application provides an electronic device, including: a memory, a processor and a computer program stored on the memory and executable on the processor, the processor executing the program to implement the defect detection method according to the first aspect.
In a fourth aspect, an embodiment of the present application provides a storage medium storing computer instructions for causing the computer to execute the defect detection method according to the first aspect.
According to the defect detection method, the defect detection system, the defect detection equipment and the storage medium, the defect image to be detected (which may or may not contain defects) is transformed and reconstructed into a reconstructed image which is free of defects, the difference between the reconstructed image and the defect image to be detected is used for obtaining the position of a defect pixel, and an image segmentation annotation data set is constructed based on the position of the defect pixel to be used as an annotation result of image segmentation; training is carried out on the preset image segmentation model based on the image segmentation labeling data set to obtain a target image segmentation model, and the image to be detected outputs a defect binary image in real time through the target image segmentation model. According to the method and the device, through the optimization of the preset image reconstruction model and the preset image segmentation model, only the non-defect sample is collected when the preset image reconstruction model is trained, and complex pixel level marking is not needed, so that the defect binary image can be output in real time by the optimized preset image segmentation model of the image to be detected, and the purposes of low-cost marking and high-precision and high-instantaneity wafer surface defect detection are achieved.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required to be used in the embodiments or the prior art description will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings may be obtained according to these drawings without inventive labor.
Fig. 1 is a flowchart of a defect detection method provided in embodiment 1 of the present invention;
fig. 2 is a detailed flowchart of step S102 of the defect detection method provided in embodiment 1 of the present invention;
FIG. 3 is a schematic illustration image of image reconstruction provided by embodiment 1 of the present invention;
fig. 4 is a detailed flowchart of step S103 of the defect detection method provided in embodiment 1 of the present invention;
fig. 5 is a flowchart illustrating a step S104 of a defect detection method according to embodiment 1 of the present invention;
fig. 6 is a schematic explanatory image of a flow of the defect detection method provided in embodiment 1 of the present invention;
FIG. 7 is a block diagram of a defect detection system corresponding to the method of embodiment 1 in accordance with embodiment 2 of the present invention;
fig. 8 is a schematic diagram of a hardware structure of an electronic device according to embodiment 3 of the present invention.
Description of the reference numerals:
10-a first building block;
20-a first training module, 21-a simulation unit, 22-a first reconstruction unit, 23-a first calculation unit, and 24-a first iteration unit;
30-a second building block, 31-a second building unit, 32-a labeling unit, 33-a defining unit;
40-a second training module, 41-an output unit, 42-a second calculation unit, 43-a second iteration unit;
50-an image acquisition module;
60-a real-time output module;
70-bus, 71-processor, 72-memory, 73-communication interface.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like reference numerals refer to the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the drawings are exemplary and intended to be illustrative of the embodiments of the present invention, and should not be construed as limiting the invention.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of the embodiments of the present invention, "a plurality" means two or more unless specifically limited otherwise.
In the manufacture of semiconductor integrated circuits, particles or defects are introduced into the processes for various reasons, and as the demand for high integration and high performance of very large scale integrated circuits is gradually increased, the semiconductor technology is moving to smaller feature sizes, the effect of the particles or defects on the quality of the integrated circuits is becoming more and more significant, and the necessity and importance of online defect detection is increasing. In the process of semiconductor manufacturing process, the defect detection is carried out on the wafer to locate and analyze the reason causing the defect, and a corresponding countermeasure is found according to the reason to avoid or reduce the defect, so that the yield and the reliability of the product are ensured, the yield and the quality of the production are fundamentally ensured, and higher profit is obtained.
Example 1
Specifically, fig. 1 is a schematic flow chart of the defect detection method provided in this embodiment.
As shown in fig. 1, the defect detection method of the present embodiment includes the following steps:
s101, constructing a defect-free image sample set from the acquired defect-free images.
The non-defective image refers to a picture with no defect in the surface quality of the industrial product. Specifically, photographs are taken on an industrial product production line according to the surface quality of the industrial product, a certain number of pictures with no surface defects are taken from the photographs to serve as a plurality of non-defective images required by the embodiment, and the non-defective images are collected to construct a non-defective image sample set.
And S102, training a preset image reconstruction model based on the defect-free image sample set to obtain a target image reconstruction model.
In order to improve the real-time effect of the image reconstruction model in the prior art, the embodiment continuously trains the adopted existing image reconstruction model through the acquired defect-free image sample set, so as to obtain the target image reconstruction model with excellent real-time effect. The reason why the preset image reconstruction model (namely, the image reconstruction model in the prior art) is not directly adopted as the defect detection model in the embodiment is to solve the defect that the real-time performance of the preset image reconstruction model is poor, so as to meet the requirement for high-speed detection and improve the detection efficiency of the surface defects of the industrial products.
Specifically, as shown in fig. 2, the specific steps of step S102 include:
and S1021, performing defect simulation on the defect-free image sample set to obtain a defect simulation sample set.
The defect simulation in the prior art mainly comprises the steps of generating a defect image based on a CAD model, and generating the defect image by simulating a ray attenuation theorem and using a ray tracking technology; the attenuation path length of each ray is obtained by jointly calculating the attenuation coefficient of the passing voxel point and the number of the voxel points, and the defect image is generated by calculating the energy intensity of the ray reaching each pixel point of the detector, so that various defect images can be generated.
And S1022, inputting the defect simulation sample set to a preset image reconstruction model to obtain a reconstructed image.
The basic principle of image reconstruction is to identify defect information and deduce defect coverage information from information other than defects, such as the image shown with reference to fig. 3. In this embodiment, the adopted preset image reconstruction model is a UNet model, and the specific network structure of the model is symmetrical and shaped like an english letter U, so that the model is called UNet; in general, UNet is an Encoder-Decoder structure (same as FCN), the first half is feature extraction, and the second half is upsampling. Wherein, the Encoder consists of convolution operation and downsampling operation, the convolution structure is unified into a convolution kernel of 3 × 3, padding =0, and striping =1. The Decoder is used to restore the original resolution of the feature map, and besides convolution, the key step of the process is upsampling and layer hopping. The up-sampling is realized by two modes of transposition convolution and interpolation. In an interpolation implementation mode, the comprehensive performance of bilinear interpolation is better and more common; the layer jump connection in UNet fuses the position information of the bottom layer and the semantic information of the deep layer through splicing. Of course, the preset image reconstruction model used in other embodiments may also be an FCN model.
S1023, calculating the loss of structural similarity according to the plurality of non-defective images and the reconstructed image to determine the value of a loss function.
The structural similarity is an index for measuring the similarity of two images, and the higher the similarity between the reconstructed image and the original defect-free image is, the better the similarity is. Natural images have very high structural properties, which means that there are strong correlations between the pixels of the image, which carry important information about the structure of objects in the visual scene. And sensing approximate information of image distortion by detecting whether the structural information is changed or not, and measuring the similarity of the two images.
Specifically, the loss function is specifically:
wherein (x, y) represents a pixel position, μ x Is the average value of x, μ y Is the average value of y, σ 2 x Is the variance of x, σ 2 y Is the variance of y, σ xy Is the covariance of x and y, c 1 =(k 1 L) 2 、c 2 =(k 2 L) 2 L is the dynamic range of the pixel value, k 1 =0.01、k 2 =0.03。
And S1024, iteratively optimizing the preset image reconstruction model based on the value of the loss function to obtain a target image reconstruction model.
S103, inputting a plurality of to-be-detected defect images into the target image reconstruction model for image reconstruction, and constructing an image segmentation and annotation data set based on the result of image reconstruction.
The defect images to be detected in the embodiment are some defect images stored in the earlier stage of the wafers of the same type to be detected and serve as training samples; it should be noted that the defect image to be detected may include a defect image or may include a non-defect image. The image to be detected is transformed and reconstructed into a non-defect perfect image, the position of a defect pixel is obtained through the difference between the result of the reconstructed image and the acquired image to be detected, and the position of the defect pixel is used as the marking result of image segmentation.
Specifically, as shown in fig. 4, the specific steps of step S103 include:
and S1031, inputting the plurality of to-be-detected defect images into the target image reconstruction model to obtain a defect-free reconstruction image.
S1032, performing one-by-one difference calculation on the pixels of the non-defective reconstructed image and the to-be-detected defective image, and marking the pixels corresponding to the difference value exceeding a preset threshold value as abnormal pixels to obtain a binary defective image.
And S1033, when the calculation efficiency of the target image reconstruction model does not meet the application scene, taking the binary defect image as an annotation result of the preset image segmentation model to obtain an image segmentation annotation data set.
The image segmentation and annotation data set comprises the defect image to be detected and the binary defect image matched with the defect image to be detected.
And S104, training a preset image segmentation model based on the image segmentation annotation data set to obtain a target image segmentation model.
Specifically, as shown in fig. 5, the specific steps of step S104 include:
and S1041, inputting the image with the defect to be detected into a preset image segmentation model to obtain model output data.
Wherein the preset image segmentation model comprises a DDRNet model. The DDRNet model used starts with a trunk and then splits into two parallel deep branches with different resolutions; one branch generates a feature map with relatively high resolution, and the other branch extracts rich semantic information through multiple downsampling operations; a plurality of bilateral connections bridge between the two branches to achieve effective information fusion. Specifically, the DDRNet model of the present embodiment employs an encoder-decoder approach, which intuitively reduces computation and inference time compared to extended convolution-based models; the encoder is usually a deep network, extracting context information by repeated spatial reduction, and the decoder restores resolution by interpolation or transposed convolution to achieve dense prediction.
S1042, calculating pixel level cross entropy loss according to the image segmentation annotation data set and the model output data.
The cross entropy is a measure of the difference degree of two different probability distributions in the same random variable, and is expressed as the difference between the true probability distribution and the predicted probability distribution in machine learning. The smaller the value of the cross entropy, the better the model prediction effect. The cross entropy is often matched with softmax in the classification problem, and the softmax processes the output result to enable the sum of the predicted values of a plurality of classifications to be 1, and then calculates the loss through the cross entropy.
And S1043, iteratively optimizing the preset image segmentation model based on the pixel-level cross entropy loss to obtain a target image segmentation model.
S105, acquiring an image to be detected of the wafer.
And S106, inputting the image to be detected into the target image segmentation model, and outputting a defect binary image in real time through the target image segmentation model.
Specifically, through optimization of the preset image reconstruction model and the preset image segmentation model, the defect binary image can be output in real time by the optimized preset image segmentation model of the image to be detected only by collecting non-defect samples during training of the preset image reconstruction model and without complex pixel level marking.
In summary, the image to be detected with defects (which may or may not include defects) is transformed and reconstructed into a "defect-free perfect" reconstructed image, the difference between the reconstructed image and the image to be detected with defects is used to obtain the position of a defective pixel, and an image segmentation annotation data set is constructed based on the position of the defective pixel to be used as an annotation result of image segmentation; training is performed on a preset image segmentation model based on an image segmentation annotation data set to obtain a target image segmentation model, and a defect binary image is output in real time by the image to be detected through the target image segmentation model, which can refer to a schematic explanatory image of the method flow shown in fig. 6. Through the steps, the purposes of marking with low cost and detecting the surface defects of the wafer with high precision and high real-time performance can be achieved.
Example 2
This embodiment provides a block diagram of a system corresponding to the method described in embodiment 1. Fig. 7 is a block diagram of a defect detection system according to an embodiment of the present application, and as shown in fig. 7, the system includes:
the first construction module 10 is used for constructing a defect-free image sample set from the acquired defect-free images;
the first training module 20 is configured to train a preset image reconstruction model based on the defect-free image sample set to obtain a target image reconstruction model;
the second construction module 30 is configured to input a plurality of defect images to be detected into the target image reconstruction model for image reconstruction, and construct an image segmentation and annotation data set based on a result of the image reconstruction;
the second training module 40 is configured to train a preset image segmentation model based on the image segmentation annotation data set to obtain a target image segmentation model;
the image acquisition module 50 is used for acquiring an image to be detected of the wafer;
a real-time output module 60, configured to input the image to be detected into the target image segmentation model, and output a defect binary image in real time through the target image segmentation model;
further, the first training module 20 includes:
the simulation unit 21 is configured to perform defect simulation on the defect-free image sample set to obtain a defect simulation sample set;
the first reconstruction unit 22 is configured to input the defect simulation sample set to a preset image reconstruction model to obtain a reconstructed image;
a first calculating unit 23, configured to calculate a value of a loss function according to the loss of structural similarity of the defect-free images and the reconstructed image;
and the first iteration unit 24 is configured to iteratively optimize the preset image reconstruction model based on the loss function to obtain a target image reconstruction model.
Further, the second building module 30 includes:
the second reconstruction unit 31 is used for inputting a plurality of to-be-detected defect images into the target image reconstruction model to obtain a defect-free reconstructed image;
the marking unit 32 is configured to perform one-by-one difference calculation on the pixels of the non-defective reconstructed image and the to-be-detected defective image, and mark a pixel corresponding to a difference value exceeding a preset threshold as an abnormal pixel to obtain a binary defective image;
and the defining unit 33 is configured to, when the calculation efficiency of the target image reconstruction model does not meet the application scenario, take the binary defect image as an annotation result of the preset image segmentation model to obtain an image segmentation annotation data set.
Further, the second training module 40 includes:
the output unit 41 is configured to input the to-be-detected defect image into a preset image segmentation model to obtain model output data; wherein the preset image segmentation model comprises a DDRNet model;
a second calculation unit 42, configured to calculate a pixel-level cross entropy loss according to the image segmentation annotation data set and the model output data;
a second iteration unit 43, configured to iteratively optimize the preset image segmentation model based on the pixel-level cross entropy loss to obtain a target image segmentation model.
The above modules may be functional modules or program modules, and may be implemented by software or hardware. For a module implemented by hardware, the modules may be located in the same processor; or the modules can be respectively positioned in different processors in any combination.
Example 3
The defect detection method described in connection with fig. 1 may be implemented by an electronic device. Fig. 8 is a schematic diagram of a hardware structure of the electronic device according to the embodiment.
The electronic device may comprise a processor 71 and a memory 72 in which computer program instructions are stored.
Specifically, the processor 71 may include a Central Processing Unit (CPU), or A Specific Integrated Circuit (ASIC), or may be configured to implement one or more Integrated circuits of the embodiments of the present Application.
The memory 72 may be used to store or cache various data files that need to be processed and/or used for communication, as well as possible computer program instructions executed by the processor 71.
The processor 71 realizes the defect detection method of embodiment 1 described above by reading and executing the computer program instructions stored in the memory 72.
In some of these embodiments, the electronic device may also include a communication interface 73 and a bus 70. As shown in fig. 8, the processor 71, the memory 72, and the communication interface 73 are connected via a bus 70 to complete communication therebetween.
The communication interface 73 is used for realizing communication among modules, devices, units and/or equipment in the embodiment of the present application. The communication interface 73 may also enable communication with other components such as: the data communication is carried out among external equipment, image/data acquisition equipment, a database, external storage, an image/data processing workstation and the like.
The bus 70 includes hardware, software, or both to couple the components of the electronic device to one another. Bus 70 includes, but is not limited to, at least one of the following: data Bus (Data Bus), address Bus (Address Bus), control Bus (Control Bus), expansion Bus (Expansion Bus), and Local Bus (Local Bus). By way of example and not limitation, bus 70 may include an Accelerated Graphics Port (AGP) or other Graphics Bus, an Enhanced Industry Standard Architecture (EISA) Bus, a Front-Side Bus (FSB), a Hyper Transport (HT) Interconnect, an ISA (ISA) Bus, an InfiniBand (InfiniBand) Interconnect, a Low Pin Count (LPC) Bus, a memory Bus, a microchannel Architecture (MCA) Bus, a PCI (Peripheral Component Interconnect) Bus, a PCI-Express (PCI-X) Bus, a Serial Advanced Technology Attachment (SATA) Bus, a vlslave Bus, a Video Bus, or a combination of two or more of these suitable electronic buses. Bus 70 may include one or more buses, where appropriate. Although specific buses are described and shown in the embodiments of the application, any suitable buses or interconnects are contemplated by the application.
The electronic device may execute the defect detection method of embodiment 1 of the present application based on the acquired defect detection system.
In addition, in combination with the defect detection method of embodiment 1, the embodiment of the present application can be implemented by providing a storage medium. The storage medium having stored thereon computer program instructions; the computer program instructions, when executed by a processor, implement the defect detection method of embodiment 1 described above.
All possible combinations of the technical features of the above embodiments may not be described for the sake of brevity, but should be considered as within the scope of the present disclosure as long as there is no contradiction between the combinations of the technical features.
The above description is intended to be illustrative of the preferred embodiment of the present invention and should not be taken as limiting the invention, but rather, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention.
Claims (10)
1. A defect detection method is used for detecting defects of a wafer, and is characterized by comprising the following steps:
constructing a defect-free image sample set from the acquired defect-free images;
training a preset image reconstruction model based on the non-defective image sample set to obtain a target image reconstruction model;
inputting a plurality of to-be-detected defect images into the target image reconstruction model for image reconstruction, and constructing an image segmentation and annotation data set based on the result of image reconstruction;
training a preset image segmentation model based on the image segmentation annotation data set to obtain a target image segmentation model;
acquiring an image to be detected of a wafer;
and inputting the image to be detected into the target image segmentation model, and outputting a defect binary image in real time through the target image segmentation model.
2. The method of claim 1, wherein the step of training a preset image reconstruction model based on the defect-free image sample set to obtain a target image reconstruction model comprises:
performing defect simulation on the defect-free image sample set to obtain a defect simulation sample set;
inputting the defect simulation sample set into a preset image reconstruction model to obtain a reconstructed image;
determining a value of a loss function according to the loss of structural similarity calculated by the plurality of non-defective images and the reconstructed image;
and iteratively optimizing the preset image reconstruction model based on the value of the loss function to obtain a target image reconstruction model.
3. The defect detection method of claim 2, wherein the preset image reconstruction model comprises an FCN model or a UNet model.
4. The defect detection method of claim 2, wherein the loss function is specifically:
wherein (x, y) represents a pixel position, μ x Is the average value of x, μ y Is the mean value of y, σ 2 x Is the variance of x, σ 2 y Is the variance of y, σ xy Is the covariance of x and y, c 1 =(k 1 L) 2 、c 2 =(k 2 L) 2 L is the dynamic range of the pixel value, k 1 =0.01、k 2 =0.03。
5. The method according to claim 1, wherein the step of inputting a plurality of images to be detected into the target image reconstruction model for image reconstruction, and constructing an image segmentation annotation dataset based on the result of image reconstruction specifically comprises:
inputting a plurality of to-be-detected defect images into the target image reconstruction model to obtain a defect-free reconstructed image;
performing one-by-one difference calculation on the pixels of the non-defective reconstructed image and the to-be-detected defective image, and marking the pixel corresponding to the difference value exceeding a preset threshold value as an abnormal pixel to obtain a binary defective image;
and when the calculation efficiency of the target image reconstruction model does not meet the application scene, taking the binary defect image as the labeling result of the preset image segmentation model to obtain an image segmentation labeling data set.
6. The defect detection method of claim 5, wherein the image segmentation annotation data set comprises the defect image to be detected and the binary defect image paired therewith.
7. The defect detection method of claim 1, wherein the step of training a preset image segmentation model based on the image segmentation labeling dataset to obtain a target image segmentation model specifically comprises:
inputting the image with the defect to be detected into a preset image segmentation model to obtain model output data; wherein the preset image segmentation model comprises a DDRNet model;
calculating pixel-level cross entropy loss according to the image segmentation annotation data set and the model output data;
and iteratively optimizing the preset image segmentation model based on the pixel-level cross entropy loss to obtain a target image segmentation model.
8. A defect detection system for performing defect detection on a wafer, the defect detection system comprising:
the first construction module is used for constructing a defect-free image sample set from the acquired defect-free images;
the first training module is used for training a preset image reconstruction model based on the defect-free image sample set to obtain a target image reconstruction model;
the second construction module is used for inputting a plurality of to-be-detected defect images into the target image reconstruction model for image reconstruction and constructing an image segmentation annotation data set based on the result of the image reconstruction;
the second training module is used for training a preset image segmentation model based on the image segmentation annotation data set to obtain a target image segmentation model;
the image acquisition module is used for acquiring an image to be detected of the wafer;
and the real-time output module is used for inputting the image to be detected into the target image segmentation model and outputting a defect binary image in real time through the target image segmentation model.
9. An electronic device, comprising: memory, a processor and a computer program stored on the memory and executable on the processor, the processor executing the program to implement the defect detection method of any one of claims 1 to 7.
10. A storage medium having stored thereon a computer program, characterized in that the program is executable by a processor for implementing a defect detection method as claimed in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211154884.7A CN115239719A (en) | 2022-09-22 | 2022-09-22 | Defect detection method, system, electronic device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211154884.7A CN115239719A (en) | 2022-09-22 | 2022-09-22 | Defect detection method, system, electronic device and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115239719A true CN115239719A (en) | 2022-10-25 |
Family
ID=83667404
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211154884.7A Pending CN115239719A (en) | 2022-09-22 | 2022-09-22 | Defect detection method, system, electronic device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115239719A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115861302A (en) * | 2023-02-16 | 2023-03-28 | 华东交通大学 | Pipe joint surface defect detection method and system |
| CN116485779A (en) * | 2023-05-11 | 2023-07-25 | 哈尔滨工业大学重庆研究院 | Adaptive wafer defect detection method and device, electronic equipment and storage medium |
| CN116664988A (en) * | 2023-07-24 | 2023-08-29 | 广立微(上海)技术有限公司 | Picture automatic labeling method, device, computer equipment and storage medium |
| CN116664839A (en) * | 2023-05-31 | 2023-08-29 | 博衍科技(珠海)有限公司 | Weak supervision semi-supervision defect perception segmentation method and device and electronic equipment |
| CN116740728A (en) * | 2023-08-11 | 2023-09-12 | 泓浒(苏州)半导体科技有限公司 | Dynamic acquisition method and system for wafer code reader |
| CN117011290A (en) * | 2023-09-27 | 2023-11-07 | 南昌理工学院 | Defect diagnosis method, system and computer equipment |
| CN117078677A (en) * | 2023-10-16 | 2023-11-17 | 江西天鑫冶金装备技术有限公司 | Defect detection method and system for starting sheet |
| CN117291914A (en) * | 2023-11-24 | 2023-12-26 | 南昌江铃华翔汽车零部件有限公司 | Automobile part defect detection method, system, computer and storage medium |
| CN117351010A (en) * | 2023-12-04 | 2024-01-05 | 中科慧远视觉技术（洛阳）有限公司 | Metal concave structure defect detection method and device based on deep learning |
| CN117876348A (en) * | 2024-01-16 | 2024-04-12 | 深圳市塞外科技有限公司 | Automatic pre-detection method and device for paper cup design defects |
| CN117876348B (en) * | 2024-01-16 | 2024-07-05 | 深圳市塞外科技有限公司 | Automatic pre-detection method and device for paper cup design defects |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110619618A (en) * | 2018-06-04 | 2019-12-27 | 杭州海康威视数字技术股份有限公司 | Surface defect detection method and device and electronic equipment |
| CN111553929A (en) * | 2020-05-12 | 2020-08-18 | 重庆邮电大学 | Mobile phone screen defect segmentation method, device and equipment based on converged network |
| CN112991344A (en) * | 2021-05-11 | 2021-06-18 | 苏州天准科技股份有限公司 | Detection method, storage medium and detection system based on deep transfer learning |
| EP3916635A1 (en) * | 2020-05-26 | 2021-12-01 | Fujitsu Limited | Defect detection method and apparatus |
| CN113781479A (en) * | 2021-11-09 | 2021-12-10 | 聚时科技(江苏)有限公司 | Semiconductor appearance defect quantification method based on image segmentation algorithm |
| CN114119460A (en) * | 2021-09-18 | 2022-03-01 | 上海联麓半导体技术有限公司 | Semiconductor image defect identification method, semiconductor image defect identification device, computer equipment and storage medium |
| CN114155244A (en) * | 2022-02-10 | 2022-03-08 | 北京阿丘科技有限公司 | Defect detection method, device, equipment and storage medium |
| CN114627072A (en) * | 2022-03-14 | 2022-06-14 | 苏州中巨苏智能科技有限公司 | Wafer defect extraction method based on neural network |
| US20220207707A1 (en) * | 2020-12-30 | 2022-06-30 | Hon Hai Precision Industry Co., Ltd. | Image defect detection method, electronic device using the same |
| CN114693685A (en) * | 2022-06-02 | 2022-07-01 | 深圳市华汉伟业科技有限公司 | Unsupervised defect detection model training method and defect detection method |
| CN114841930A (en) * | 2022-04-18 | 2022-08-02 | 南京信息工程大学 | Product defect detection method based on parallel multi-scale fusion network |
| CN114972194A (en) * | 2022-04-27 | 2022-08-30 | 华中科技大学 | Method for detecting defects from inconsistent labels |
| CN114998324A (en) * | 2022-07-19 | 2022-09-02 | 深圳新视智科技术有限公司 | Training method and device for semiconductor wafer defect detection model |
| CN115018857A (en) * | 2022-08-10 | 2022-09-06 | 南昌昂坤半导体设备有限公司 | Image segmentation method, image segmentation device, computer-readable storage medium and computer equipment |
| CN115063385A (en) * | 2022-06-29 | 2022-09-16 | 哈尔滨工业大学重庆研究院 | Machine vision method for wafer detection |
-
2022
- 2022-09-22 CN CN202211154884.7A patent/CN115239719A/en active Pending
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110619618A (en) * | 2018-06-04 | 2019-12-27 | 杭州海康威视数字技术股份有限公司 | Surface defect detection method and device and electronic equipment |
| CN111553929A (en) * | 2020-05-12 | 2020-08-18 | 重庆邮电大学 | Mobile phone screen defect segmentation method, device and equipment based on converged network |
| EP3916635A1 (en) * | 2020-05-26 | 2021-12-01 | Fujitsu Limited | Defect detection method and apparatus |
| US20220207707A1 (en) * | 2020-12-30 | 2022-06-30 | Hon Hai Precision Industry Co., Ltd. | Image defect detection method, electronic device using the same |
| CN112991344A (en) * | 2021-05-11 | 2021-06-18 | 苏州天准科技股份有限公司 | Detection method, storage medium and detection system based on deep transfer learning |
| CN114119460A (en) * | 2021-09-18 | 2022-03-01 | 上海联麓半导体技术有限公司 | Semiconductor image defect identification method, semiconductor image defect identification device, computer equipment and storage medium |
| CN113781479A (en) * | 2021-11-09 | 2021-12-10 | 聚时科技(江苏)有限公司 | Semiconductor appearance defect quantification method based on image segmentation algorithm |
| CN114155244A (en) * | 2022-02-10 | 2022-03-08 | 北京阿丘科技有限公司 | Defect detection method, device, equipment and storage medium |
| CN114627072A (en) * | 2022-03-14 | 2022-06-14 | 苏州中巨苏智能科技有限公司 | Wafer defect extraction method based on neural network |
| CN114841930A (en) * | 2022-04-18 | 2022-08-02 | 南京信息工程大学 | Product defect detection method based on parallel multi-scale fusion network |
| CN114972194A (en) * | 2022-04-27 | 2022-08-30 | 华中科技大学 | Method for detecting defects from inconsistent labels |
| CN114693685A (en) * | 2022-06-02 | 2022-07-01 | 深圳市华汉伟业科技有限公司 | Unsupervised defect detection model training method and defect detection method |
| CN115063385A (en) * | 2022-06-29 | 2022-09-16 | 哈尔滨工业大学重庆研究院 | Machine vision method for wafer detection |
| CN114998324A (en) * | 2022-07-19 | 2022-09-02 | 深圳新视智科技术有限公司 | Training method and device for semiconductor wafer defect detection model |
| CN115018857A (en) * | 2022-08-10 | 2022-09-06 | 南昌昂坤半导体设备有限公司 | Image segmentation method, image segmentation device, computer-readable storage medium and computer equipment |
Non-Patent Citations (2)
| Title |
|---|
| YUANDUO HONG等: "Deep Dual-resolution Networks for Real-time and Accurate Semantic Segmentation of Road Scenes", 《ARXIV》 * |
| 赵朗月等: "基于机器视觉的表面缺陷检测方法研究进展", 《仪器仪表学报》 * |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115861302A (en) * | 2023-02-16 | 2023-03-28 | 华东交通大学 | Pipe joint surface defect detection method and system |
| CN116485779A (en) * | 2023-05-11 | 2023-07-25 | 哈尔滨工业大学重庆研究院 | Adaptive wafer defect detection method and device, electronic equipment and storage medium |
| CN116485779B (en) * | 2023-05-11 | 2024-01-30 | 哈尔滨工业大学重庆研究院 | Adaptive wafer defect detection method and device, electronic equipment and storage medium |
| CN116664839B (en) * | 2023-05-31 | 2024-02-13 | 博衍科技(珠海)有限公司 | Weak supervision semi-supervision defect perception segmentation method and device and electronic equipment |
| CN116664839A (en) * | 2023-05-31 | 2023-08-29 | 博衍科技(珠海)有限公司 | Weak supervision semi-supervision defect perception segmentation method and device and electronic equipment |
| CN116664988B (en) * | 2023-07-24 | 2023-11-21 | 广立微(上海)技术有限公司 | Picture automatic labeling method, device, computer equipment and storage medium |
| CN116664988A (en) * | 2023-07-24 | 2023-08-29 | 广立微(上海)技术有限公司 | Picture automatic labeling method, device, computer equipment and storage medium |
| CN116740728A (en) * | 2023-08-11 | 2023-09-12 | 泓浒(苏州)半导体科技有限公司 | Dynamic acquisition method and system for wafer code reader |
| CN116740728B (en) * | 2023-08-11 | 2023-10-27 | 泓浒(苏州)半导体科技有限公司 | Dynamic acquisition method and system for wafer code reader |
| CN117011290B (en) * | 2023-09-27 | 2023-12-15 | 南昌理工学院 | Defect diagnosis method, system and computer equipment |
| CN117011290A (en) * | 2023-09-27 | 2023-11-07 | 南昌理工学院 | Defect diagnosis method, system and computer equipment |
| CN117078677B (en) * | 2023-10-16 | 2024-01-30 | 江西天鑫冶金装备技术有限公司 | Defect detection method and system for starting sheet |
| CN117078677A (en) * | 2023-10-16 | 2023-11-17 | 江西天鑫冶金装备技术有限公司 | Defect detection method and system for starting sheet |
| CN117291914A (en) * | 2023-11-24 | 2023-12-26 | 南昌江铃华翔汽车零部件有限公司 | Automobile part defect detection method, system, computer and storage medium |
| CN117291914B (en) * | 2023-11-24 | 2024-02-09 | 南昌江铃华翔汽车零部件有限公司 | Automobile part defect detection method, system, computer and storage medium |
| CN117351010A (en) * | 2023-12-04 | 2024-01-05 | 中科慧远视觉技术（洛阳）有限公司 | Metal concave structure defect detection method and device based on deep learning |
| CN117351010B (en) * | 2023-12-04 | 2024-03-01 | 中科慧远视觉技术（洛阳）有限公司 | Metal concave structure defect detection method and device based on deep learning |
| CN117876348A (en) * | 2024-01-16 | 2024-04-12 | 深圳市塞外科技有限公司 | Automatic pre-detection method and device for paper cup design defects |
| CN117876348B (en) * | 2024-01-16 | 2024-07-05 | 深圳市塞外科技有限公司 | Automatic pre-detection method and device for paper cup design defects |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115239719A (en) | Defect detection method, system, electronic device and storage medium | |
| US8826200B2 (en) | Alteration for wafer inspection | |
| CN112861729B (en) | Real-time depth completion method based on pseudo-depth map guidance | |
| US9170211B2 (en) | Design-based inspection using repeating structures | |
| CN112232391A (en) | Dam crack detection method based on U-net network and SC-SAM attention mechanism | |
| CN110826588A (en) | Drainage pipeline defect detection method based on attention mechanism | |
| CN116797533B (en) | Appearance defect detection method and system for power adapter | |
| CN111582309B (en) | Method for generating bad point detection model of design layout and method for detecting bad point | |
| CN110763706A (en) | System and method for analyzing crystal defects | |
| CN112288714A (en) | Hardware Trojan horse detection method based on deep learning | |
| CN115423796A (en) | Chip defect detection method and system based on TensorRT accelerated reasoning | |
| CN116311462A (en) | Facial image restoration and recognition method combining context information and VGG19 | |
| CN116823793A (en) | Device defect detection method, device, electronic device and readable storage medium | |
| CN115661828B (en) | Character direction recognition method based on dynamic hierarchical nested residual error network | |
| CN115690001A (en) | Method for detecting defects in steel pipe welding digital radiographic image | |
| CN114821272A (en) | Image recognition method, image recognition system, image recognition medium, electronic device, and target detection model | |
| Kee et al. | Cracks identification using mask region-based denoised deformable convolutional network | |
| CN113034432A (en) | Product defect detection method, system, device and storage medium | |
| CN113077458A (en) | Cloud and shadow detection method and system in remote sensing image | |
| Wang et al. | IH-ViT: Vision Transformer-based Integrated Circuit Appear-ance Defect Detection | |
| Wang et al. | Frequency and content dual stream network for image dehazing | |
| CN116109627B (en) | Defect detection method, device and medium based on migration learning and small sample learning | |
| CN117440104B (en) | Data compression reconstruction method based on target significance characteristics | |
| CN113657214B (en) | Building damage assessment method based on Mask RCNN | |
| CN112329925B (en) | Model generation method, feature extraction method, device and electronic equipment |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20221025 |