KR102402194B1 - Deep learning based end-to-end o-ring defect inspection method - Google Patents

Abstract

A deep learning-based object detection method performed by a computing device according to one embodiment of the present disclosure is disclosed. The method includes the steps of: inputting an image to a first neural network model and extracting location information of an object included in the image; performing pre-processing on the image based on the location information of the object extracted through the first neural network model; and inputting the pre-processed image to a second neural network model to determine a defect of the object. The first neural network model and the second neural network model may be different neural network models.

Description

DEEP LEARNING BASED END-TO-END O-RING DEFECT INSPECTION METHOD

The present invention relates to a method for detecting an object, and more particularly, to a technology for detecting a defect in an object based on an artificial intelligence technology.

In the production process, the computer vision inspection area has a very high rate of manual re-judgment of inspection results by operators due to the limitations of inspection algorithms. Accordingly, the number of workers input to re-determine a good product and a bad product is increased, and there is a fear that the cost competitiveness of the product may be weakened.

Recently, deep learning technology is leading the rapid development in the field of computer vision and is positioned as a core technology. In order to overcome the limitations of visual inspection and to increase productivity by improving the reliability and speed of inspection, computer vision-based inspection automation system development is actively underway. However, in industry, there is a problem of high cost due to the purchase of hardware and software and a lack of expertise in computer vision. In addition, a significant data set must be built to utilize deep learning techniques, but in reality, the number of data sets for poor products is very scarce. As such, it is difficult to practically apply computer vision-based deep learning technology in the field to determine defective products.

Therefore, in order to increase the detection efficiency of defective products despite the insufficient data set, an automated defective product detection system based on deep learning technology is required.

US Patent Publication No. 10-2021-0150700 (May 20, 2021. 5.) discloses a defect detection apparatus and method.

The present disclosure has been devised in response to the above-described background technology, and aims to determine a defect of an object based on deep learning.

Disclosed is a method for detecting a defect on an object performed by a computing device according to an embodiment of the present disclosure for realizing the above-described problem. The method may include: inputting an image into a first neural network model and extracting location information of an object included in the image; performing pre-processing on the image based on the location information of the object extracted through the first neural network model; and inputting the pre-processed image into a second neural network model to determine a defect in the object, wherein the first neural network model and the second neural network model are different from each other. can be

In an alternative embodiment, the object may include a seal, which is a component in the manufacturing process made of at least one of synthetic rubber or heat-resistant plastic.

In an alternative embodiment, the object may include a ring made of synthetic rubber or synthetic resin and having a circular cross-section.

In an alternative embodiment, performing pre-processing on the image may include non-linearly transforming the intensity of light on the image.

In an alternative embodiment, performing pre-processing on the image may include performing gamma correction on the image.

In an alternative embodiment, the first neural network model performs model scaling related to at least one of width scaling to control the number of filters, depth scaling to control the number of layers, or resolution scaling to control the resolution of an input image , and the first neural network model may be transfer-trained after pre-training.

In an alternative embodiment, the step of determining the defect of the object by inputting the image on which the pre-processing has been performed into the second neural network model includes inputting the image on which the pre-processing has been performed into the second neural network model to determine the defect of the object. extracting distinguishable features; and classifying the object as good or defective through the second neural network model based on the extracted feature.

In an alternative embodiment, the second neural network model performs model scaling related to at least one of width scaling to control the number of filters, depth scaling to control the number of layers, or resolution scaling to control the resolution of an input image , and the second neural network model may be transfer-trained after pre-training.

In an alternative embodiment, the first neural network model may include: obtaining first result data including a bounding box through the first neural network model; generating first training data by cropping the result data to correspond to a bounding box of the first result data; generating a second training data set by applying a data augmentation algorithm to the generated first training data; and performing additional training based on the generated second training data set.

In an alternative embodiment, the data augmentation algorithm comprises: zooming in first training data comprising the object; sampling the first training data including good objects; and by distorting the first training data according to a predetermined criterion, sampling the first training data including the defective object.

In an alternative embodiment, the predetermined criterion is determined according to a defect type of the object, wherein the defect type of the object is protruding, dent, short, filled, scratch, Or it may include at least one of a breakage.

In an alternative embodiment, the first neural network model may include an EfficientDet-D2 model, and the second neural network model may include an EfficientNet-B0 model.

A computer program stored in a computer-readable storage medium is disclosed according to an embodiment of the present disclosure for realizing the above-described problems. When the computer program is executed in one or more processors, the computer program performs the following operations for detecting a defect in an object, the operations comprising: inputting an image into a first neural network model to provide location information of an object included in the image the operation of extracting ; performing pre-processing on the image based on the location information of the object extracted through the first neural network model; and inputting the preprocessed image into a second neural network model to determine a defect in the object, wherein the first neural network model and the second neural network model may be different neural network models. have.

Disclosed is a computing device for detecting a defect in an object according to an embodiment of the present disclosure for realizing the above-described problem. The apparatus includes: a processor including at least one core; a memory containing program codes executable by the processor; and a network unit, wherein the processor inputs an image into a first neural network model to extract location information of an object included in the image, and based on the location information of the object extracted through the first neural network model , performing pre-processing on the image, inputting the pre-processed image to a second neural network model to determine a defect in the object, and the first neural network model and the second neural network model are each other It may be a different neural network model.

According to an embodiment of the present disclosure, a computer-readable recording medium storing a data structure for storing data related to operation of a neural network is disclosed. The data is obtained through the following operations, and the operations include: inputting an image into a first neural network model to extract location information of an object included in the image; performing pre-processing on the image based on the location information of the object extracted through the first neural network model; and inputting the preprocessed image into a second neural network model to determine a defect in the object, wherein the first neural network model and the second neural network model may be different neural network models. have.

Through the method according to an embodiment of the present disclosure, a defect of an object may be determined regardless of the size of the object based on deep learning.

1 is a block diagram of a computing device for detecting a defect in an object according to an embodiment of the present disclosure.
2 is a schematic diagram illustrating a network function according to an embodiment of the present disclosure;
3 is a flowchart illustrating an object defect detection process of a deep learning model according to an embodiment of the present disclosure.
4 is a diagram illustrating a process of determining an object defect according to an embodiment of the present disclosure.
5 is a diagram illustrating pre-processing of an image according to an embodiment of the present disclosure.
6 is a flowchart illustrating a pre-learning process of a neural network model according to an embodiment of the present disclosure.
7 is a diagram illustrating a defect type of an object according to an embodiment of the present disclosure.
8 is a schematic diagram of a computing environment according to an embodiment of the present disclosure;

Various embodiments are now described with reference to the drawings. In this specification, various descriptions are presented to provide an understanding of the present disclosure. However, it is apparent that these embodiments may be practiced without these specific descriptions.

The terms “component,” “module,” “system,” and the like, as used herein, refer to a computer-related entity, hardware, firmware, software, a combination of software and hardware, or execution of software. For example, a component can be, but is not limited to being, a process running on a processor, a processor, an object, a thread of execution, a program, and/or a computer. For example, both an application running on a computing device and the computing device may be a component. One or more components may reside within a processor and/or thread of execution. A component may be localized within one computer. A component may be distributed between two or more computers. In addition, these components can execute from various computer readable media having various data structures stored therein. Components may communicate via a network such as the Internet with another system, for example, via a signal having one or more data packets (eg, data and/or signals from one component interacting with another component in a local system, distributed system, etc.) may communicate via local and/or remote processes depending on the data being transmitted).

In addition, the term “or” is intended to mean an inclusive “or” rather than an exclusive “or.” That is, unless otherwise specified or clear from context, "X employs A or B" is intended to mean one of the natural implicit substitutions. That is, X employs A; X employs B; or when X employs both A and B, "X employs A or B" may apply to either of these cases. It should also be understood that the term “and/or” as used herein refers to and includes all possible combinations of one or more of the listed related items.

Also, the terms "comprises" and/or "comprising" should be understood to mean that the feature and/or element in question is present. However, it should be understood that the terms "comprises" and/or "comprising" do not exclude the presence or addition of one or more other features, elements and/or groups thereof. Also, unless otherwise specified or unless it is clear from context to refer to a singular form, the singular in the specification and claims should generally be construed to mean “one or more”.

And, the term "at least one of A or B" should be interpreted as meaning "when only A is included, when only B is included, "when combined with the configuration of A and B".

Those skilled in the art will further appreciate that the various illustrative logical blocks, configurations, modules, circuits, means, logics, and algorithm steps described in connection with the embodiments disclosed herein may be implemented in electronic hardware, computer software, or combinations of both. It should be recognized that they can be implemented with To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, configurations, means, logics, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application. However, such implementation decisions should not be interpreted as causing a departure from the scope of the present disclosure.

Descriptions of the presented embodiments are provided to enable those skilled in the art to use or practice the present invention. Various modifications to these embodiments will be apparent to those skilled in the art of the present disclosure. The generic principles defined herein may be applied to other embodiments without departing from the scope of the present disclosure. Therefore, the present invention is not limited to the embodiments presented herein. This invention is to be interpreted in its widest scope consistent with the principles and novel features presented herein.

In the present disclosure, a network function, an artificial neural network, a neural network, a neural network model, and a deep learning model may be used interchangeably.

The term “Nth (N is a natural number)” expressed throughout the detailed description and claims of the present disclosure may be understood as an expression used to distinguish elements of the invention from a functional or structural point of view. The term “Nth” may be understood as a term used to distinguish and express components according to a predetermined criterion. For example, in the present disclosure, two models that are based on the same neural network structure but perform different functions due to differences in input/output data, etc. can be distinguished through an expression called an N-th configuration, like the first model and the second model. In addition, two models that perform the same function but are based on different neural network structures can be distinguished through the expression of the N-th configuration like the first model and the second model. However, this is only an example, and the term “Nth” may be used to distinguish components according to various criteria from various viewpoints.

As used throughout this description, the terms "image", "image", "product shot image", or "camera image" may be used interchangeably and refer to discrete image elements (eg, two-dimensional image). It refers to multidimensional data composed of pixels in the case of a 3D image and voxels in the case of a 3D image). For example, "image" means an AOI device, VRS device, ROS device, any device in which a camera is installed to photograph a product, or any other product photographing imaging system known in the art, that is, a subject collected by an imaging system. It may be an image obtained by detecting a subject. Also, the image may be provided in a context other than product defect detection, for example, a remote sensing system, electron microscopy, and the like.

Throughout the description of the present invention, "image" or "image", used interchangeably with each other, refers to a visual image (eg, displayed on a video screen) or an image (eg, AOI device, VRS device, ROS device, etc.) A term referring to a digital representation of an image (such as a file corresponding to the pixel output of

1 is a block diagram of a computing device for processing an oral image according to an embodiment of the present disclosure.

The configuration of the computing device 100 shown in FIG. 1 is only a simplified example. In an embodiment of the present disclosure, the computing device 100 may include other components for performing the computing environment of the computing device 100 , and only some of the disclosed components may configure the computing device 100 .

The computing device 100 may include a processor 110 , a memory 130 , and a network unit 150 .

The processor 110 may include one or more cores, and a central processing unit (CPU) of a computing device, a general purpose graphics processing unit (GPGPU), and a tensor processing unit (TPU). unit) and the like, and may include a processor for data analysis and deep learning. The processor 110 may read a computer program stored in the memory 130 to perform data processing for machine learning according to an embodiment of the present disclosure. According to an embodiment of the present disclosure, the processor 110 may perform an operation for learning the neural network. The processor 110 for learning of the neural network, such as processing input data for learning in deep learning (DL), extracting features from the input data, calculating an error, updating the weight of the neural network using backpropagation calculations can be performed. At least one of a CPU, a GPGPU, and a TPU of the processor 110 may process learning of a network function. For example, the CPU and the GPGPU can process learning of a network function and data classification using the network function. Also, in an embodiment of the present disclosure, learning of a network function and data classification using the network function may be processed by using the processors of a plurality of computing devices together. In addition, the computer program executed in the computing device according to an embodiment of the present disclosure may be a CPU, GPGPU or TPU executable program.

According to an embodiment of the present disclosure, the processor 110 may extract location information included in the image by inputting the image into the first neural network model. The image may include an image obtained from an image capturing apparatus (eg, a camera) of any type for capturing an object in an industry. In the present disclosure, an image may include an object used for repair or maintenance in the field of aerospace, automobile, or general engineering. For example, the object may include a seal, which is a component in a manufacturing process made of at least one of synthetic rubber and heat-resistant plastic. In this case, the object may include a ring having a circular cross-section. As an example, an object in the present disclosure may include an O-ring. For convenience of explanation, the O-ring will be described below for such an object as an example. The first neural network model may be any type of deep learning model that performs object detection. In the present disclosure, the first neural network model may include a deep learning model that performs localization in object detection. Localization may mean outputting location information about where an object in an image is located in the image. For example, the processor 110 may output location information included in the image using a bounding box through the first neural network model.

According to an embodiment of the present disclosure, the processor 110 may perform pre-processing on the image based on the location information of the object extracted through the first neural network model. Pre-processing of an image may refer to any form of processing for an image, including loading an image, saving an image, changing an image size, cropping an image, changing an image tone, removing a background, and the like. For example, the preprocessing of the image is performed by changing, enlarging, reducing, modifying, deleting, synthesizing, and adding the image for the purpose of improving the quality of the result of the deep learning model before inputting the image to the deep learning model. This may include processing.

By performing preprocessing on the image, the image can be converted into features that can be used in the deep learning model. In the present disclosure, preprocessing for an image is 'image cropping' that reduces the dimension by removing the periphery of the image, or 'image cropping' that crops the image through array slicing, and 'image that visually changes the image based on the kernel color change'. Here, the kernel determines the degree of blur and softness of the image, and may mean an operation performed by neighboring pixels. In one embodiment, the image tone change may be performed by non-linearly transforming the intensity of light with respect to the image through a non-linear transfer function. For example, the intensity of light for an image may be modified through gamma correction. According to an embodiment of the present disclosure, the processor 110 performs gamma correction on an image, so that defects of an object to be described later can be clearly identified.

According to an embodiment of the present disclosure, the processor 110 may input the image on which the preprocessing has been performed into the second neural network model to determine a defect in the object. The second neural network model may be any deep learning model that performs classification on the image on which the above-described preprocessing has been performed. In the present disclosure, the second neural network model may include a deep learning model that performs classification in object detection. Classification may mean classifying the type of an object in an image given as an input. For example, the processor 110 may distinguish whether the state of the object included in the image is bad (NG) (or defective) or good (OK) through the second neural network model.

According to an embodiment of the present disclosure, the memory 130 may store any type of information generated or determined by the processor 110 and any type of information received by the network unit 150 .

According to an embodiment of the present disclosure, the memory 130 is a flash memory type, a hard disk type, a multimedia card micro type, or a card type memory (eg, SD or XD memory, etc.), Random Access Memory (RAM), Static Random Access Memory (SRAM), Read-Only Memory (ROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Programmable Read (PROM) -Only Memory), a magnetic memory, a magnetic disk, and an optical disk may include at least one type of storage medium. The computing device 100 may operate in relation to a web storage that performs a storage function of the memory 130 on the Internet. The description of the above-described memory is only an example, and the present disclosure is not limited thereto.

The network unit 150 according to an embodiment of the present disclosure may use any type of known wired/wireless communication system.

The network unit 150 may receive an image obtained from an object detection and/or management system for photographing an object in an industry. For example, an image including an object is a detection image obtained by photographing a structure of an object, such as a two-dimensional or three-dimensional shape, and may be data for training or inference of a neural network model. In this case, the image including the object may include all images representing the structural shape of the object obtained through photographing using an AOI device, a VRS device, an ROS device, or the like.

In addition, the network unit 150 may transmit and receive information processed by the processor 110 , a user interface, and the like through communication with other terminals. For example, the network unit 150 may provide a user interface generated by the processor 110 to a client (e.g. a user terminal). Also, the network unit 150 may receive an external input of a user authorized as a client and transmit it to the processor 110 . In this case, the processor 110 may process an operation of outputting, modifying, changing, or adding information provided through the user interface based on the user's external input received from the network unit 150 .

Meanwhile, the computing device 100 according to an embodiment of the present disclosure may include a server as a computing system that transmits and receives information through communication with a client. In this case, the client may be any type of terminal that can access the server. For example, the computing device 100 as a server may receive object images from an object defect detection system, detect a defect included in the image, and provide the detection result to a server or a user terminal.

In an additional embodiment, the computing device 100 may include any type of terminal that receives a data resource generated from an arbitrary server and performs additional information processing.

2 is a schematic diagram illustrating a network function according to an embodiment of the present disclosure;

A deep learning model according to an embodiment of the present disclosure may include a neural network for detecting an object. Throughout this specification, the terms neural network, network function, and neural network may be used interchangeably. A neural network may be composed of a set of interconnected computational units, which may generally be referred to as nodes. These nodes may also be referred to as neurons. A neural network is configured to include at least one or more nodes. Nodes (or neurons) constituting the neural networks may be interconnected by one or more links.

In the neural network, one or more nodes connected through a link may relatively form a relationship between an input node and an output node. The concepts of an input node and an output node are relative, and any node in an output node relationship with respect to one node may be in an input node relationship in a relationship with another node, and vice versa. As described above, an input node-to-output node relationship may be created around a link. One or more output nodes may be connected to one input node through a link, and vice versa.

In the relationship between the input node and the output node connected through one link, the value of the data of the output node may be determined based on data input to the input node. Here, a link interconnecting the input node and the output node may have a weight.

Throughout this specification, weights and parameters may be used interchangeably. The weight may be variable, and may be changed by a user or an algorithm in order for the neural network to perform a desired function. For example, when one or more input nodes are interconnected to one output node by respective links, the output node sets values input to input nodes connected to the output node and links corresponding to the respective input nodes. An output node value may be determined based on the weight.

As described above, in a neural network, one or more nodes are interconnected through one or more links to form an input node and an output node relationship in the neural network. The characteristics of the neural network may be determined according to the number of nodes and links in the neural network, the correlation between the nodes and the links, and the value of a weight assigned to each of the links. For example, when the same number of nodes and links exist and there are two neural networks having different weight values of the links, the two neural networks may be recognized as different from each other.

A neural network may consist of a set of one or more nodes. A subset of nodes constituting the neural network may constitute a layer. Some of the nodes constituting the neural network may configure one layer based on distances from the initial input node. For example, a set of nodes having a distance of n from the initial input node may constitute n layers. The distance from the initial input node may be defined by the minimum number of links that must be passed to reach the corresponding node from the initial input node. However, the definition of such a layer is arbitrary for description, and the order of the layer in the neural network may be defined in a different way from the above. For example, a layer of nodes may be defined by a distance from the final output node.

The initial input node may mean one or more nodes to which data is directly input without going through a link in a relationship with other nodes among nodes in the neural network. Alternatively, in a relationship between nodes based on a link in a neural network, it may mean nodes that do not have other input nodes connected by a link. Similarly, the final output node may refer to one or more nodes that do not have an output node in relation to other nodes among nodes in the neural network. In addition, the hidden node may mean nodes constituting the neural network other than the first input node and the last output node.

The neural network according to an embodiment of the present disclosure may be a neural network in which the number of nodes in the input layer may be the same as the number of nodes in the output layer, and the number of nodes decreases and then increases again as progresses from the input layer to the hidden layer. can Also, in the neural network according to another embodiment of the present disclosure, the number of nodes in the input layer may be less than the number of nodes in the output layer, and the number of nodes may be reduced as the number of nodes progresses from the input layer to the hidden layer. have. In addition, the neural network according to another embodiment of the present disclosure may be a neural network in which the number of nodes in the input layer may be greater than the number of nodes in the output layer, and the number of nodes increases as the number of nodes progresses from the input layer to the hidden layer. can The neural network according to another embodiment of the present disclosure may be a neural network in a combined form of the aforementioned neural networks.

A deep neural network (DNN) may refer to a neural network including a plurality of hidden layers in addition to an input layer and an output layer. Deep neural networks can be used to identify the latent structures of data. In other words, it can identify the potential structure of photos, texts, videos, voices, and music (e.g., what objects are in the photos, what the text and emotions are, what the texts and emotions are, etc.) . Deep neural networks include convolutional neural networks (CNNs), recurrent neural networks (RNNs), auto encoders, generative adversarial networks (GANs), and restricted boltzmann machines (RBMs). machine), a deep trust network (DBN), a Q network, a U network, a Siamese network, and the like. The description of the deep neural network described above is only an example, and the present disclosure is not limited thereto.

In an embodiment of the present disclosure, the network function may include an autoencoder. The auto-encoder may be a kind of artificial neural network for outputting output data similar to input data. The auto encoder may include at least one hidden layer, and an odd number of hidden layers may be disposed between the input/output layers. The number of nodes in each layer may be reduced from the number of nodes of the input layer to an intermediate layer called the bottleneck layer (encoding), and then expanded symmetrically with reduction from the bottleneck layer to the output layer (symmetrically with the input layer). The auto-encoder can perform non-linear dimensionality reduction. The number of input layers and output layers may correspond to a dimension after preprocessing the input data. In the auto-encoder structure, the number of nodes of the hidden layer included in the encoder may have a structure that decreases as the distance from the input layer increases. If the number of nodes in the bottleneck layer (the layer with the fewest nodes located between the encoder and decoder) is too small, a sufficient amount of information may not be conveyed, so a certain number or more (e.g., more than half of the input layer, etc.) ) may be maintained.

The neural network may be trained by at least one of teacher learning, unsupervised learning, semi-supervised learning, and reinforcement learning. Learning of the neural network may be a process of applying knowledge for the neural network to perform a specific operation to the neural network.

A neural network can be trained in a way that minimizes output errors. In the training of a neural network, iteratively input the training data into the neural network, calculate the output of the neural network and the target error for the training data, and calculate the error of the neural network from the output layer of the neural network to the input layer in the direction to reduce the error. It is a process of updating the weight of each node in the neural network by backpropagation in the direction. In the case of teacher learning, learning data in which the correct answer is labeled in each learning data is used (ie, labeled learning data), and in the case of comparative learning, the correct answer may not be labeled in each learning data. That is, for example, learning data in the case of teacher learning related to data classification may be data in which categories are labeled in each of the learning data. Labeled training data is input to the neural network, and an error can be calculated by comparing the output (category) of the neural network with the label of the training data. As another example, in the case of comparison learning related to data classification, an error may be calculated by comparing the input training data with the neural network output. The calculated error is back propagated in the reverse direction (ie, from the output layer to the input layer) in the neural network, and the connection weight of each node of each layer of the neural network may be updated according to the back propagation. The change amount of the connection weight of each node to be updated may be determined according to a learning rate. The computation of the neural network on the input data and the backpropagation of errors can constitute a learning cycle (epoch). The learning rate may be applied differently according to the number of repetitions of the learning cycle of the neural network. For example, in the early stage of learning of a neural network, a high learning rate can be used to enable the neural network to quickly obtain a certain level of performance, thereby increasing efficiency, and using a low learning rate at a later stage of learning can increase accuracy.

In the training of neural networks, in general, the training data may be a subset of real data (that is, data to be processed using the trained neural network), and thus, the error on the training data is reduced, but the error on the real data is reduced. There may be increasing learning cycles. Overfitting is a phenomenon in which errors on actual data increase by over-learning on training data as described above. For example, a phenomenon in which a neural network that has learned a cat by seeing a yellow cat does not recognize that it is a cat when it sees a cat other than yellow may be a type of overfitting. Overfitting can act as a cause of increasing errors in machine learning algorithms. In order to prevent such overfitting, various optimization methods can be used. In order to prevent overfitting, methods such as increasing the training data, regularization, dropout that deactivate some nodes of the network in the process of learning, and the use of a batch normalization layer can be applied.

According to an embodiment of the present disclosure, the first neural network model and the second neural network model are deep learning models for object detection, and each model may perform different functions by the processor 110 . As described above with reference to FIG. 1 , the processor 110 may perform localization through the first neural network model, and may perform classification through the second neural network model. In an embodiment, the neural network models may include deep learning models for object detection such as CNN, R-CNN, U-net, ResNet, SPP-net, YOLO, EfficientDet, and EfficientNet. The above models are merely examples, and the present disclosure is not limited thereto. In the present disclosure, the processor 110 may extract location information of the object through the first neural network model and classify the defect of the object through the second neural network model. For example, the first neural network model and the second neural network model may be configured in series.

According to an embodiment of the present disclosure, the first neural network model and the second neural network model may include a process in which model scaling is performed. Model scaling is a fine tuning of the model related to at least one of width scaling to control the number of filters of the deep learning model, depth scaling to control the number of layers of the deep learning model, or resolution scaling to control the resolution of the input image. -tuning) can be understood as In general, as the number of filters increases, the number of layers increases, and the resolution of the input image increases, the performance of the model can be improved. However, in reality, as the size of the model increases, the time required for the learning process and the inference process increases exponentially, so there is a limit to increasing the model's performance by increasing the size of the model. In an embodiment of the present disclosure, model scaling may include compound scaling that simultaneously considers width scaling, depth scaling, and resolution scaling. Compound scaling has the advantage of being able to increase the performance of the model without increasing the size of the model by considering three factors at the same time. For this purpose, a preset compound coefficient may be expressed as in the following [Equation 1]. For convenience of explanation, it has been described that all three scaling methods are considered for compound scaling, but according to an embodiment, compound scaling in the form of combining at least two of the above three scaling methods is also included within the scope of the present disclosure. can

는 compound coefficient이고,

,

,

는 scaling factor이다. compound coefficient는 depth, width, resolution을 모두 고려하기 위한 사용자 지정 값일 수 있다. 또한,

는 모델의 연산 수를

배로 증가하도록 하기 위해 설정된 것으로 모델에서 수행되는 연산의 수를 제한할 수 있다. 본 개시에 따르면, 복합 계수(compound coefficient)인

값이 설정되면, 모델의 한정된 자원 내에서 최고의 성능을 내도록

,

,

(각각 depth, width, resolution)값이 결정될 수 있다. here

is the compound coefficient,

,

,

is the scaling factor. The compound coefficient may be a user-specified value to consider all depth, width, and resolution. In addition,

is the number of operations in the model

It is set to multiply by a factor and can limit the number of operations performed on the model. According to the present disclosure, the compound coefficient

When the value is set, it is designed to give the best performance within the limited resources of the model.

,

,

(respectively, depth, width, and resolution) values can be determined.

According to an embodiment of the present disclosure, the first neural network model may include a pre-trained EfficientDet-D2 model, and the second neural network model may include a pre-trained EfficientNet-B0 model. A number included in the model name may be understood as the above-described complex coefficient, and the first and second neural network models may be the most lightweight models among EfficientDet and EfficientNet models. According to the present disclosure, the processor 110 may increase the efficiency of object defect detection through the neural network model based on a limited data set. In an embodiment, the processor 110 may determine location information of an object included in the image by performing localization through the EfficientDet-D2 model. In this case, location information about the object may be expressed based on the bounding box. As an embodiment, the processor 110 may determine whether the object included in the image is good (OK) or bad (NG) by performing classification through the EfficientNet-B0 model. In another embodiment, the pre-trained EfficientDet-D2 and the pre-trained EfficientNet-B0 may be serially connected to construct a deep learning model. In this case, EfficientDet-D2 can perform the function of classifying objects, but it may be difficult to classify objects with defects on the surface but no difference in shape. Therefore, in the present disclosure, by combining with EfficientNet-B0 optimized for a specific classification task, there is an effect of increasing the defect detection performance of an object.

According to an embodiment of the present disclosure, the second neural network model may further include a fully connected (FC) layer to increase classification performance in the pre-trained Efficient-B0 model. For example, the second neural network model may include two FC layers including 128 and 64 nodes instead of the FC layer of the existing Efficient-B0 model. According to the present disclosure, the second neural network model may be transfer-trained to improve the performance of classifying object defects based on the limited number of data sets. Details will be described later with reference to FIG. 6 .

3 is a flowchart illustrating an object defect detection process of a deep learning model according to an embodiment of the present disclosure.

Referring to FIG. 3 , in step S310 , the computing device 100 according to an embodiment of the present disclosure may input an image into the first neural network model and extract location information of an object included in the image.

In operation S320, the computing device 100 according to an embodiment of the present disclosure may perform preprocessing on the image based on the object location information extracted through the first neural network model.

In step S330 , the computing device 100 according to an embodiment of the present disclosure may input the pre-processed image to the second neural network model to determine a defect in the object.

In an embodiment of the present disclosure, the computing device 100 may determine a defect of an object by inputting an image into the deep learning model. The image may include an image obtained from any camera that captures an object in the manufacturing process. In the present disclosure, an image may include an object used for repair or maintenance in the field of aerospace, automobile, or general engineering. For example, the object may include a seal, which is a component in a manufacturing process made of at least one of synthetic rubber and heat-resistant plastic. In this case, the object may include a ring having a circular cross-section.

According to an embodiment of the present disclosure, the deep learning model may include a first neural network model and a second neural network model. A deep learning model may include any model for object detection. As an embodiment, the first neural network model may include a model performing a localization function, and the second neural network model may include a model performing a classification function. According to the present disclosure, the first neural network model may include a pre-trained EfficientDet-D2 model to extract location information of an image, and the second neural network model is a pre-trained EfficientNet-B0 model for defect determination. may include For example, the computing device 100 may distinguish whether the state of the object included in the image is defective (NG) or good (OK) through the pre-trained EfficientDet-D2 model and the pre-trained EfficientNet-B0 model.

According to an embodiment of the present disclosure, the computing device 100 may perform preprocessing on an image based on location information of an object extracted through the first neural network model. Preprocessing on an image can include loading the image, saving the image, resizing the image, cropping the image, changing the image tone, removing the background, etc. For example, the preprocessing of the image is performed by changing, enlarging, reducing, modifying, deleting, synthesizing, and adding the image for the purpose of improving the quality of the result of the deep learning model before inputting the image to the deep learning model. This may include processing. By performing preprocessing on the image, the image can be converted into features that can be used in the deep learning model. In the present disclosure, preprocessing for an image is 'image cropping' that reduces the dimension by removing the periphery of the image, or 'image cropping' that crops the image through array slicing, and 'image that visually changes the image based on the kernel color change'. Here, the kernel determines the degree of blur and softness of the image, and may mean an operation performed by neighboring pixels. In one embodiment, the image tone change may be performed by non-linearly modifying the intensity of light for the image. For example, the intensity of light for an image may be modified through gamma correction. According to an embodiment of the present disclosure, the computing device 100 performs gamma correction on an image, so that defects of an object can be clearly identified. Therefore, the performance of the deep learning model for judging object defects through image preprocessing can be improved.

4 is a diagram illustrating a process of determining an object defect according to an embodiment of the present disclosure.

Referring to FIG. 4 , the processor 110 may acquire the location information 42 of the object included in the image by inputting the image 41 into the first neural network model 410 to determine the defect of the object. The processor 110 may perform image preprocessing 420 to clearly identify a defect of an object based on the location information 42 of the object included in the acquired image. The processor 110 may input the preprocessed image 43 into the second neural network model 430 to determine whether an object included in the image is good 44 or has a defect 45 .

According to an embodiment of the present disclosure, the processor 110 inputs an image 41 including an object used in a manufacturing process into the first neural network model 410 for localization, and location information of the object included in the image (42) can be obtained. The image 41 may include images obtained from all cameras that photograph an object in the manufacturing process. In the present disclosure, an object may include a seal, which is a component in a manufacturing process made of at least one of synthetic rubber and heat-resistant plastic. In the present disclosure, the object may include a ring having a circular cross-sectional shape. For example, the object may mean an O-ring as a ring having a synthetic rubber material and having a circular cross-sectional shape. The first neural network model 410 may include a model for acquiring location information of an object to be detected. In the present disclosure, the first neural network model 410 may include a pre-trained EfficientDet-D2 model. For example, the first neural network model 410 may acquire location information of the O-ring by inputting the image 41 including the O-ring into the pre-trained EfficientDet-D2 model. In this case, the location information of the O-ring may be expressed based on a bounding box of a predetermined size.

According to an embodiment of the present disclosure, the processor 110 may perform image preprocessing 420 to clearly identify a defect of an object based on the acquired location information 42 of the object. Preprocessing on an image can include loading the image, saving the image, resizing the image, cropping the image, changing the image tone, removing the background, etc. For example, the preprocessing of the image is performed by changing, enlarging, reducing, modifying, deleting, synthesizing, and adding the image for the purpose of improving the quality of the result of the deep learning model before inputting the image to the deep learning model. This may include processing. In the present disclosure, preprocessing for an image is 'image cropping' that reduces the dimension by removing the periphery of the image, or 'image cropping' that crops the image through array slicing, and 'image that visually changes the image based on the kernel color change'. As an embodiment, the processor 110 may crop an image including the location information 42 of the object based on a bounding box of a predetermined size. Through this, the dimension of the image to be input to the second neural network 43 may be constantly adjusted. For example, when the bounding box is set based on the location information of the O-ring, the processor 110 may crop the image 41 to fit the size of the bounding box. Also, in a data augmentation algorithm to be described later in FIG. 6 , cropping of an image may contribute to constructing various types of data sets.

According to an embodiment of the present disclosure, the image tone change may be performed by non-linearly transforming the intensity of light for the image. Also, the intensity of light for the image may be modified through gamma correction. According to an embodiment of the present disclosure, the processor 110 performs gamma correction on the image, so that the defect of the object can be clearly identified. For example, the processor 110 may acquire an image in which the surface of the object can be clearly identified by performing gamma correction in the image preprocessing 420 . In the present disclosure, as the surface of an object is clearly identified by gamma correction, an effect of facilitating determination of a defect of an object through the second neural network model 430 occurs.

According to an embodiment of the present disclosure, the processor 110 inputs the pre-processed image 43 into the second neural network model 430 for classification to determine whether the object included in the image is good (44) or defective (44). 45) can be determined. The second neural network model 430 may include a model for performing classification based on object characteristics. In the present disclosure, the second neural network model 430 may include a pre-trained EfficientDet-B0 model. For example, the second neural network model 430 inputs the preprocessed image 43 for the O-ring into the pre-trained EfficientDet-B0 model to determine whether the O-ring is good (44) or has a defect (45). can be distinguished In the present disclosure, the defect of the O-ring may be determined based on a pre-learned threshold for pixel output.

5 is a diagram illustrating pre-processing of an image according to an embodiment of the present disclosure.

Referring to FIG. 5 , the processor 110 may perform pre-processing on an image. According to an embodiment of the present disclosure, pre-processing of the image may be performed on an output image in which location information is obtained through the first neural network model. Preprocessing on an image can include loading the image, saving the image, resizing the image, cropping the image, changing the image tone, removing the background, etc. For example, the preprocessing of the image is performed by changing, enlarging, reducing, modifying, deleting, synthesizing, and adding the image for the purpose of improving the quality of the result of the deep learning model before inputting the image to the deep learning model. This may include processing. In the present disclosure, preprocessing for an image is 'image cropping' that reduces the dimension by removing the periphery of the image, or 'image cropping' that crops the image through array slicing, and 'image that visually changes the image based on the kernel color change'. As an embodiment, the processor 110 may crop an image including location information of an object based on a bounding box of a predetermined size. Through this, the dimension of the image to be input to the second neural network model may be constantly adjusted. For example, when the bounding box is set based on the location information of the O-ring, the processor 110 may crop the image to fit the size of the bounding box. In the present disclosure, since the training data set is constructed based on cropping, the defect of the O-ring can be determined regardless of the type and size of the defect of the O-ring.

According to an embodiment of the present disclosure, the image tone change may be performed by non-linearly transforming the intensity of light with respect to the image through a nonlinear transfer function. Human vision can respond non-linearly to brightness according to Weber's law. Weber's law may mean that in order to feel a change in stimulus in the sensory organ, it must be stimulated at a certain rate or more with respect to the initial stimulus. In the present disclosure, when the intensity of light for an image is non-linearly modified in accordance with human visual characteristics, an effect of increasing the discrimination power of object defects occurs without incurring a large cost.

According to an embodiment of the present disclosure, the intensity of light for an image may be modified through gamma correction. Gamma correction can be expressed as [Equation 2] below.

는 입력 이미지를 나타내고,

는 밝기를 변경하기 위한 값이고,

는 출력 이미지를 나타낸다. 일 실시예로,

값이 0.5보다 크면 이미지가 원본 이미지보다 밝아지고, 0.5보다 작으면 원본 이미지보다 어두워질 수 있다. here

represents the input image,

is the value to change the brightness,

represents the output image. In one embodiment,

If the value is greater than 0.5, the image may be brighter than the original image, and if it is less than 0.5, the image may be darker than the original image.

According to an embodiment of the present disclosure, when photographing an object, the lighting may be located on the bottom surface of the object. In this case, it may be difficult to detect defects on the object surface. The above-described problem can be overcome by allowing the processor 110 to control the brightness of the image through gamma correction. In the present disclosure, the processor 110 may clearly identify defects present on the surface of the O-ring by performing gamma correction on the image. For example, when comparing the images before and after performing gamma correction in FIG. 5 , it can be seen that the discrimination power for defects is significantly different. According to the present disclosure, the processor 110 may improve the defect determination performance of the second neural network model by performing pre-processing on the image.

6 is a diagram illustrating a pre-learning process of a neural network model according to an embodiment of the present disclosure.

Referring to FIG. 6 , in operation S610 , the computing device 100 according to an embodiment of the present disclosure may acquire first result data including a bounding box through a first neural network model.

In step S620 , the first training data may be generated by cropping the result data to correspond to the bounding box of the first result data.

In step S630 , a second training data set may be generated by applying a data augmentation algorithm to the generated first training data.

In operation S640 , additional training may be performed based on the generated second training data set.

According to an embodiment of the present disclosure, the computing device 100 may obtain first result data including a bounding box indicating location information of an object through a first neural network model for localization in a pre-learning process. In the present disclosure, the first result data may correspond to the image 42 indicating the location information of the object in FIG. 4 ,

According to an embodiment of the present disclosure, the computing device 100 may generate the first training data by cropping the result data to correspond to the bounding box of the first result data obtained in the pre-learning process. In the present disclosure, the bounding box may have a predetermined size. In the present disclosure, the first training data is cropped based on a predetermined bounding box, and may be an image representing a certain dimension. For example, the first training data may be an image including an O-ring cropped to fit the size of the bounding box.

According to an embodiment of the present disclosure, the computing device 100 may generate a second training data set by applying a data augmentation algorithm to the first training data generated in the pre-learning process. The data augmentation algorithm may be a technique for securing data necessary to sufficiently train a deep learning model. In the present disclosure, a data augmentation algorithm may refer to a series of processes for securing a large amount of data based on a limited amount of data samples. In this case, the series of processes may be geometric transformations such as resizing, inverting, cropping, rotating, moving, splitting a window, and changing the shape. In an embodiment of the present disclosure, the second training data set may be a data set generated through geometric transformation. For example, when an object for which a defect is to be determined is an O-ring, the amount of samples having a defect may be very limited. When the number of data is limited, overfitting may occur, which may not respond properly to test data or new data by over-adapting only to training data. Therefore, in order to learn a deep learning model with a limited amount of data set, it can be said that securing data through a data augmentation algorithm is essential.

According to an embodiment of the present disclosure, the data augmentation algorithm may be performed by the AutoAugment technique. AutoAugment is an algorithm that automatically finds the best aggregation policy for image data through reinforcement learning. The AutoAugmnet technique is only an example, and the present disclosure is not limited thereto.

According to an embodiment of the present disclosure, the data augmentation algorithm may include zooming the first training data. In the present disclosure, the first training data may be expanded to fit a specification used in a manufacturing process. For example, when the first training data is an image about the O-ring, it may be enlarged to a size of 1.5 times or 2 times. The multiple of the magnification is merely an example, and the present disclosure is not limited thereto.

According to an embodiment of the present disclosure, a data augmentation algorithm generates first training data including defective objects by sampling first training data including good objects and distorting it according to a predetermined criterion. may include doing In the present disclosure, the first training data may be artificially changed according to a predetermined defect type. For example, when there is a sample for a normal O-ring, an arbitrary defect type may be applied to an arbitrary position by using Photoshop for the sample. Defect types will be described later in detail with reference to FIG. 7 .

According to an embodiment of the present disclosure, the computing device 100 may perform additional learning based on the second training data set generated in the pre-learning process. As an embodiment, the computing device 100 may further learn the second neural network model based on the second training data set generated through the data augmentation algorithm. As described above, in order to improve the performance of the second neural network model that performs classification, various data sets are required. Since the second neural network model may be additionally learned based on various types of data sets obtained through the data augmentation algorithm, the computing device 100 determines the size and type of objects included in the image through the second neural network model. Regardless, the defect can be detected.

According to an embodiment of the present disclosure, the neural network model may be pre-trained and transfer-trained. Transfer learning can mean taking and reusing a lower layer of a model when there is a similar model that has been trained previously. This not only speeds up learning, but also reduces the number of training sets required for learning. In an embodiment, the first neural network model may be transfer-trained based on all layers of EfficientDet-D2 that have been pre-trained in order to identify object location information. In an embodiment, the second neural network model may be transfer-trained based on an average pooling layer located at the end of EfficientNet-B0 that has been pre-trained for classification.

7 is a flowchart illustrating a defect type of an object according to an embodiment of the present disclosure.

Referring to FIG. 7 , the processor 110 may secure a training data set for pre-training a neural network model through a data augmentation algorithm.

According to an embodiment of the present disclosure, a data augmentation algorithm generates first training data including defective objects by sampling first training data including good objects and distorting it according to a predetermined criterion. may include doing In the present disclosure, the first training data may be artificially changed according to a predetermined defect type. For example, when there is a sample for a normal O-ring, an arbitrary defect type may be applied to an arbitrary position by using Photoshop for the sample.

According to an embodiment of the present disclosure, the predetermined defect type of the object includes at least one of a protrusion 71 , a warpage 72 , a deficiency 73 , a filling 74 , a scratch 75 , or a defect 76 . can do. In the present disclosure, the defect type may include a defect type related to the O-ring. For example, the protrusion 71 may refer to a portion protruding from the shape of the object. The bending 72 may mean a portion in which the shape of the object is distorted. The deficiency 73 may mean that the shape of the object is partially lacking and thus does not have the shape of the completed object. Filling 74 may mean that the inner circle portion is filled in the original ring shape. The scratch 75 may mean that the shape surface of the object is scratched. The defect 76 may mean that a part of the shape of the object is broken. The types of defects of the above-described objects are only examples, and the present disclosure is not limited thereto.

According to an embodiment of the present disclosure, the processor 110 may generate the first training data by combining the above-described defect types through a data augmentation algorithm. For example, processor 110 may generate first training data including scratches 75 and deficits 76 .

8 is a simplified, general schematic diagram of an exemplary computing environment in which embodiments of the present disclosure may be implemented.

Although the present disclosure has been described above as being generally capable of being implemented by a computing device, those skilled in the art will appreciate that the present disclosure may be implemented in hardware and software in combination with computer-executable instructions and/or other program modules and/or in combination with computer-executable instructions and/or other program modules that may be executed on one or more computers. It will be appreciated that it can be implemented as a combination.

Generally, program modules include routines, programs, components, data structures, etc. that perform particular tasks or implement particular abstract data types. In addition, those skilled in the art will appreciate that the methods of the present disclosure can be applied to single-processor or multiprocessor computer systems, minicomputers, mainframe computers as well as personal computers, handheld computing devices, microprocessor-based or programmable consumer electronics, and the like. It will be appreciated that each of these may be implemented in other computer system configurations, including those capable of operating in connection with one or more associated devices.

The described embodiments of the present disclosure may also be practiced in distributed computing environments where certain tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote memory storage devices.

Computers typically include a variety of computer-readable media. Any medium accessible by a computer can be a computer-readable medium, and such computer-readable media includes volatile and nonvolatile media, transitory and non-transitory media, removable and non-transitory media. including removable media. By way of example, and not limitation, computer-readable media may include computer-readable storage media and computer-readable transmission media. Computer readable storage media includes volatile and nonvolatile media, temporary and non-transitory media, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. includes media. A computer-readable storage medium may be RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital video disk (DVD) or other optical disk storage device, magnetic cassette, magnetic tape, magnetic disk storage device, or other magnetic storage device. device, or any other medium that can be accessed by a computer and used to store the desired information.

Computer readable transmission media typically embodies computer readable instructions, data structures, program modules or other data, etc. in a modulated data signal such as a carrier wave or other transport mechanism, and Includes any information delivery medium. The term modulated data signal means a signal in which one or more of the characteristics of the signal is set or changed so as to encode information in the signal. By way of example, and not limitation, computer-readable transmission media includes wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, RF, infrared, and other wireless media. Combinations of any of the above are also intended to be included within the scope of computer-readable transmission media.

An example environment 1100 implementing various aspects of the disclosure is shown including a computer 1102 , the computer 1102 including a processing unit 1104 , a system memory 1106 , and a system bus 1108 . do. A system bus 1108 couples system components, including but not limited to system memory 1106 , to the processing device 1104 . The processing device 1104 may be any of a variety of commercially available processors. Dual processor and other multiprocessor architectures may also be used as processing unit 1104 .

The system bus 1108 may be any of several types of bus structures that may further interconnect a memory bus, a peripheral bus, and a local bus using any of a variety of commercial bus architectures. System memory 1106 includes read only memory (ROM) 1110 and random access memory (RAM) 1112 . A basic input/output system (BIOS) is stored in non-volatile memory 1110, such as ROM, EPROM, EEPROM, etc., the BIOS is the basic input/output system (BIOS) that helps transfer information between components within computer 1102, such as during startup. contains routines. RAM 1112 may also include high-speed RAM, such as static RAM, for caching data.

The computer 1102 may also include an internal hard disk drive (HDD) 1114 (eg, EIDE, SATA) - this internal hard disk drive 1114 may also be configured for external use within a suitable chassis (not shown). Yes—a magnetic floppy disk drive (FDD) 1116 (eg, for reading from or writing to removable diskette 1118), and an optical disk drive 1120 (eg, a CD-ROM) for reading from, or writing to, disk 1122, or other high capacity optical media, such as DVD. The hard disk drive 1114 , the magnetic disk drive 1116 , and the optical disk drive 1120 are connected to the system bus 1108 by the hard disk drive interface 1124 , the magnetic disk drive interface 1126 , and the optical drive interface 1128 , respectively. ) can be connected to The interface 1124 for implementing an external drive includes at least one or both of Universal Serial Bus (USB) and IEEE 1394 interface technologies.

These drives and their associated computer-readable media provide non-volatile storage of data, data structures, computer-executable instructions, and the like. In the case of computer 1102, drives and media correspond to storing any data in a suitable digital format. Although the description of computer readable media above refers to HDDs, removable magnetic disks, and removable optical media such as CDs or DVDs, those skilled in the art will use zip drives, magnetic cassettes, flash memory cards, cartridges, etc. It will be appreciated that other tangible computer-readable media such as etc. may also be used in the exemplary operating environment and any such media may include computer-executable instructions for performing the methods of the present disclosure.

A number of program modules may be stored in the drive and RAM 1112 , including an operating system 1130 , one or more application programs 1132 , other program modules 1134 , and program data 1136 . All or portions of the operating system, applications, modules, and/or data may also be cached in RAM 1112 . It will be appreciated that the present disclosure may be implemented in various commercially available operating systems or combinations of operating systems.

A user may enter commands and information into the computer 1102 via one or more wired/wireless input devices, for example, a pointing device such as a keyboard 1138 and a mouse 1140 . Other input devices (not shown) may include a microphone, IR remote control, joystick, game pad, stylus pen, touch screen, and the like. Although these and other input devices are connected to the processing unit 1104 through an input device interface 1142 that is often connected to the system bus 1108, parallel ports, IEEE 1394 serial ports, game ports, USB ports, IR interfaces, It may be connected by other interfaces, etc.

A monitor 1144 or other type of display device is also coupled to the system bus 1108 via an interface, such as a video adapter 1146 . In addition to the monitor 1144, the computer typically includes other peripheral output devices (not shown), such as speakers, printers, and the like.

Computer 1102 may operate in a networked environment using logical connections to one or more remote computers, such as remote computer(s) 1148 via wired and/or wireless communications. Remote computer(s) 1148 may be workstations, computing device computers, routers, personal computers, portable computers, microprocessor-based entertainment devices, peer devices, or other common network nodes, and are typically connected to computer 1102 . Although it includes many or all of the components described for it, only memory storage device 1150 is shown for simplicity. The logical connections shown include wired/wireless connections to a local area network (LAN) 1152 and/or a larger network, eg, a wide area network (WAN) 1154 . Such LAN and WAN networking environments are common in offices and companies, and facilitate enterprise-wide computer networks, such as intranets, all of which can be connected to a worldwide computer network, for example, the Internet.

When used in a LAN networking environment, the computer 1102 is coupled to the local network 1152 through a wired and/or wireless communication network interface or adapter 1156 . Adapter 1156 may facilitate wired or wireless communication to LAN 1152 , which LAN 1152 also includes a wireless access point installed therein for communicating with wireless adapter 1156 . When used in a WAN networking environment, the computer 1102 may include a modem 1158, be connected to a communication computing device on the WAN 1154, or establish communications over the WAN 1154, such as over the Internet. have other means. A modem 1158 , which may be internal or external and a wired or wireless device, is coupled to the system bus 1108 via a serial port interface 1142 . In a networked environment, program modules described for computer 1102 , or portions thereof, may be stored in remote memory/storage device 1150 . It will be appreciated that the network connections shown are exemplary and other means of establishing a communication link between the computers may be used.

The computer 1102 may be associated with any wireless device or object that is deployed and operates in wireless communication, for example, a printer, scanner, desktop and/or portable computer, portable data assistant (PDA), communication satellite, wireless detectable tag. It operates to communicate with any device or place, and phone. This includes at least Wi-Fi and Bluetooth wireless technologies. Accordingly, the communication may be a predefined structure as in a conventional network or may simply be an ad hoc communication between at least two devices.

Wi-Fi (Wireless Fidelity) makes it possible to connect to the Internet, etc. without a wired connection. Wi-Fi is a wireless technology such as cell phones that allows these devices, eg, computers, to transmit and receive data indoors and outdoors, ie anywhere within range of a base station. Wi-Fi networks use a radio technology called IEEE 802.11 (a, b, g, etc.) to provide secure, reliable, and high-speed wireless connections. Wi-Fi can be used to connect computers to each other, to the Internet, and to wired networks (using IEEE 802.3 or Ethernet). Wi-Fi networks may operate in unlicensed 2.4 and 5 GHz radio bands, for example at 11 Mbps (802.11a) or 54 Mbps (802.11b) data rates, or in products that include both bands (dual band). .

On the other hand, a computer-readable medium storing a data structure according to an embodiment of the present disclosure is disclosed.

The data structure may refer to the organization, management, and storage of data that enables efficient access and modification of data. A data structure may refer to an organization of data to solve a specific problem (eg, data retrieval, data storage, and data modification in the shortest time). A data structure may be defined as a physical or logical relationship between data elements designed to support a particular data processing function. The logical relationship between data elements may include a connection relationship between user-defined data elements. Physical relationships between data elements may include actual relationships between data elements physically stored on a computer-readable storage medium (eg, persistent storage). A data structure may specifically include a set of data, relationships between data, and functions or instructions applicable to data. Through an effectively designed data structure, the computing device can perform an operation while using the resource of the computing device to a minimum. Specifically, the computing device may increase the efficiency of operations, reads, insertions, deletions, comparisons, exchanges, and retrievals through effectively designed data structures.

A data structure may be classified into a linear data structure and a non-linear data structure according to the type of the data structure. The linear data structure may be a structure in which only one piece of data is connected after one piece of data. The linear data structure may include a list, a stack, a queue, and a deck. A list may mean a set of data in which an order exists internally. The list may include a linked list. The linked list may be a data structure in which data is linked in such a way that each data is linked in a line with a pointer. In a linked list, a pointer may contain information about a link with the next or previous data. A linked list may be expressed as a single linked list, a doubly linked list, or a circularly linked list according to a shape. A stack can be a data enumeration structure with limited access to data. A stack can be a linear data structure in which data can be manipulated (eg, inserted or deleted) at only one end of the data structure. The data stored in the stack may be a data structure LIFO-Last in First Out. A queue is a data listing structure that allows limited access to data. Unlike a stack, the queue may be a data structure that comes out later (FIFO-First in First Out) as data stored later. A deck can be a data structure that can process data at either end of the data structure.

The non-linear data structure may be a structure in which a plurality of data is connected after one data. The nonlinear data structure may include a graph data structure. A graph data structure may be defined as a vertex and an edge, and the edge may include a line connecting two different vertices. A graph data structure may include a tree data structure. The tree data structure may be a data structure in which one path connects two different vertices among a plurality of vertices included in the tree. That is, it may be a data structure that does not form a loop in the graph data structure.

Throughout this specification, computational model, neural network, network function, and neural network may be used interchangeably. Hereinafter, the neural network is unified and described. The data structure may include a neural network. And the data structure including the neural network may be stored in a computer-readable medium. Data structures, including neural networks, also include preprocessed data for processing by the neural network, data input to the neural network, weights of the neural network, hyperparameters of the neural network, data obtained from the neural network, activation functions associated with each node or layer of the neural network, and neural networks. It may include a loss function for learning of . A data structure comprising a neural network may include any of the components disclosed above. That is, the data structure including the neural network includes preprocessed data for processing by the neural network, data input to the neural network, weights of the neural network, hyperparameters of the neural network, data obtained from the neural network, activation function associated with each node or layer of the neural network, and neural network It may be configured to include all or any combination thereof, such as a loss function for learning of. In addition to the above-described configurations, a data structure including a neural network may include any other information that determines a characteristic of a neural network. In addition, the data structure may include all types of data used or generated in the operation process of the neural network, and is not limited to the above. Computer-readable media may include computer-readable recording media and/or computer-readable transmission media. A neural network may be composed of a set of interconnected computational units, which may generally be referred to as nodes. These nodes may also be referred to as neurons. A neural network is configured by including at least one or more nodes.

The data structure may include data input to the neural network. A data structure including data input to the neural network may be stored in a computer-readable medium. The data input to the neural network may include learning data input in a neural network learning process and/or input data input to the neural network in which learning is completed. Data input to the neural network may include pre-processing data and/or pre-processing target data. The preprocessing may include a data processing process for inputting data into the neural network. Accordingly, the data structure may include data to be pre-processed and data generated by pre-processing. The above-described data structure is merely an example, and the present disclosure is not limited thereto.

The data structure may include the weights of the neural network. In this specification, a weight, a parameter, and a parameter may be used with the same meaning. And the data structure including the weight of the neural network may be stored in a computer-readable medium. The neural network may include a plurality of weights. The weight may be variable, and may be changed by a user or an algorithm in order for the neural network to perform a desired function. For example, when one or more input nodes are interconnected to one output node by respective links, the output node sets values input to input nodes connected to the output node and links corresponding to the respective input nodes. A data value output from the output node may be determined based on the weight. The above-described data structure is merely an example, and the present disclosure is not limited thereto.

By way of example and not limitation, the weight may include a weight variable in a neural network learning process and/or a weight in which neural network learning is completed. The variable weight in the neural network learning process may include a weight at a time point at which a learning cycle starts and/or a weight variable during the learning cycle. The weight for which neural network learning is completed may include a weight for which a learning cycle has been completed. Accordingly, the data structure including the weights of the neural network may include a data structure including the weights that vary in the process of learning the neural network and/or the weights on which the learning of the neural network is completed. Therefore, it is assumed that the above-described weights and/or combinations of weights are included in the data structure including the weights of the neural network. The above-described data structure is merely an example, and the present disclosure is not limited thereto.

The data structure including the weights of the neural network may be stored in a computer-readable storage medium (eg, memory, hard disk) after being serialized. Serialization can be the process of converting a data structure into a form that can be reconstructed and used later by storing it on the same or a different computing device. The computing device may serialize the data structure to send and receive data over the network. A data structure including weights of the serialized neural network may be reconstructed in the same computing device or in another computing device through deserialization. The data structure including the weight of the neural network is not limited to serialization. Furthermore, the data structure including the weights of the neural network is a data structure to increase the efficiency of computation while using the resources of the computing device to a minimum (e.g., B-Tree, Trie, m-way search tree, AVL tree, Red-Black Tree). The foregoing is merely an example, and the present disclosure is not limited thereto.

The data structure may include hyper-parameters of the neural network. In addition, the data structure including the hyperparameters of the neural network may be stored in a computer-readable medium. The hyper parameter may be a variable variable by a user. Hyperparameters are, for example, learning rate, cost function, number of iterations of the learning cycle, weight initialization (e.g., setting the range of weight values to be initialized for weights), Hidden Unit It may include the number (eg, the number of hidden layers, the number of nodes of the hidden layer). The above-described data structure is merely an example, and the present disclosure is not limited thereto.

One of ordinary skill in the art of this disclosure will understand that information and signals may be represented using any of a variety of different technologies and techniques. For example, data, instructions, instructions, information, signals, bits, symbols and chips that may be referenced in the above description are voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields particles or particles, or any combination thereof.

Those of ordinary skill in the art of the present disclosure will recognize that the various illustrative logical blocks, modules, processors, means, circuits, and algorithm steps described in connection with the embodiments disclosed herein include electronic hardware, (convenience For this purpose, it will be understood that it may be implemented by various forms of program or design code (referred to herein as software) or a combination of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. A person skilled in the art of the present disclosure may implement the described functionality in various ways for each specific application, but such implementation decisions should not be interpreted as a departure from the scope of the present disclosure.

The various embodiments presented herein may be implemented as methods, apparatus, or articles of manufacture using standard programming and/or engineering techniques. The term article of manufacture includes a computer program, carrier, or media accessible from any computer-readable storage device. For example, computer-readable storage media include magnetic storage devices (eg, hard disks, floppy disks, magnetic strips, etc.), optical disks (eg, CDs, DVDs, etc.), smart cards, and flash drives. memory devices (eg, EEPROMs, cards, sticks, key drives, etc.). Also, various storage media presented herein include one or more devices and/or other machine-readable media for storing information.

It is understood that the specific order or hierarchy of steps in the presented processes is an example of exemplary approaches. Based on design priorities, it is understood that the specific order or hierarchy of steps in the processes may be rearranged within the scope of the present disclosure. The appended method claims present elements of the various steps in a sample order, but are not meant to be limited to the specific order or hierarchy presented.

The description of the presented embodiments is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the scope of the present disclosure. Thus, the present disclosure is not intended to be limited to the embodiments presented herein, but is to be construed in the widest scope consistent with the principles and novel features presented herein.

Claims (15)

A deep learning-based end-to-end O-ring fault detection method performed by a computing device comprising at least one processor, the method comprising:
inputting an image into a first neural network model and extracting location information of an object included in the image;
performing pre-processing on the image based on the location information of the object extracted through the first neural network model; and
determining a defect of the object by inputting the pre-processed image into a second neural network model;
includes,
The first neural network model and the second neural network model are different neural network models,
The first neural network model is a transfer-trained model to extract the position information of the object based on the layer of the pre-trained EfficientDet-D2 model, and the second neural network model is the last of the pre-trained EfficientNet-B0 model. It is a transfer-trained model to determine the defect of the object based on the average pooling layer located in
Way.
The method of claim 1,
The object is
Including a seal, which is a part in the manufacturing process made of at least one of synthetic rubber or heat-resistant plastic,
Way.
The method of claim 1,
The object is
Including a ring having a circular cross-section made of synthetic rubber or synthetic resin,
Way.
4. The method of claim 2 or 3,
The pre-processing of the image includes:
non-linearly transforming an intensity of light for the image;
containing,
Way.
4. The method of claim 2 or 3,
The pre-processing of the image includes:
performing gamma correction on the image;
containing,
Way.
4. The method of claim 2 or 3,
The first neural network model is
a model performing model scaling related to at least one of width scaling to control the number of filters, depth scaling to control the number of layers, or resolution scaling to control the resolution of an input image;
containing,
Way.
4. The method of claim 2 or 3,
The step of determining the defect of the object by inputting the pre-processed image into the second neural network model,
inputting the pre-processed image into the second neural network model and extracting a feature capable of classifying the defect of the object; and
classifying the object as good or defective through the second neural network model based on the extracted features;
containing,
Way.
8. The method of claim 7,
The second neural network model is
a model performing model scaling related to at least one of width scaling to control the number of filters, depth scaling to control the number of layers, or resolution scaling to control the resolution of an input image;
containing,
Way.
4. The method of claim 2 or 3,
The second neural network model is
obtaining first result data including a bounding box through the first neural network model;
generating first training data by cropping the result data to correspond to a bounding box of the first result data;
generating a second training data set by applying a data augmentation algorithm to the generated first training data; and
performing additional training based on the generated second training data set;
Pre-learned through a learning process that includes
Way.
10. The method of claim 9,
The data augmentation algorithm,
Enlarging the first training data including the object (zooming);
sampling the first training data including good objects; and
sampling the first training data including the defective object by distorting the first training data according to a predetermined criterion;
containing,
Way.
11. The method of claim 10,
The predetermined criterion is determined according to the defect type of the object,
The defect type of the object includes at least one of protruding, dent, shortage, filled, scratch, or breakage,
Way.
delete A computer program stored in a computer readable storage medium, wherein, when the computer program is executed on one or more processors, it performs operations of detecting an O-ring defect with an end-to-end deep learning-based operation, the operations comprising:
inputting an image into a first neural network model and extracting location information of an object included in the image;
performing pre-processing on the image based on the location information of the object extracted through the first neural network model; and
inputting the pre-processed image into a second neural network model to determine a defect in the object;
includes,
The first neural network model and the second neural network model are different neural network models,
The first neural network model is a transfer-trained model to extract the position information of the object based on the layer of the pre-trained EfficientDet-D2 model, and the second neural network model is the last of the pre-trained EfficientNet-B0 model. It is a transfer-trained model to determine the defect of the object based on the average pooling layer located in
A computer program stored on a computer-readable storage medium.
As a computing device that performs O-RING defect detection with end-to-end deep learning,
a processor including at least one core;
a memory containing program codes executable by the processor; and
a network unit for receiving an image;
including,
The processor is
Input the image to the first neural network model to extract the location information of the object included in the image,
performing pre-processing on the image based on the location information of the object extracted through the first neural network model;
Input the pre-processed image to a second neural network model to determine a defect in the object, and
The first neural network model and the second neural network model are different neural network models,
The first neural network model is a transfer-trained model to extract the position information of the object based on the layer of the pre-trained EfficientDet-D2 model, and the second neural network model is the last of the pre-trained EfficientNet-B0 model. It is a transfer-trained model to determine the defect of the object based on the average pooling layer located in
Device.
A computer-readable recording medium storing a data structure storing data related to operation of a neural network for O-ring defect detection, wherein the data is obtained through the following operations, the operations comprising:
inputting an image into a first neural network model and extracting location information of an object included in the image;
performing pre-processing on the image based on the location information of the object extracted through the first neural network model; and
determining a defect in the object by inputting the pre-processed image into a second neural network model;
includes,
The first neural network model and the second neural network model are different neural network models,
The first neural network model is a transfer-trained model to extract the position information of the object based on the layer of the pre-trained EfficientDet-D2 model, and the second neural network model is the last of the pre-trained EfficientNet-B0 model. It is a transfer-trained model to determine the defect of the object based on the average pooling layer located in
A computer-readable recording medium having a data structure stored thereon.

Images