CN112784494B - Training method of false positive recognition model, target recognition method and device - Google Patents
Training method of false positive recognition model, target recognition method and device Download PDFInfo
- Publication number
- CN112784494B CN112784494B CN202110113779.8A CN202110113779A CN112784494B CN 112784494 B CN112784494 B CN 112784494B CN 202110113779 A CN202110113779 A CN 202110113779A CN 112784494 B CN112784494 B CN 112784494B
- Authority
- CN
- China
- Prior art keywords
- target
- disturbance
- model
- recognition
- false positive
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000012549 training Methods 0.000 title claims abstract description 106
- 238000000034 method Methods 0.000 title claims abstract description 85
- 239000013598 vector Substances 0.000 claims abstract description 90
- 238000012545 processing Methods 0.000 claims abstract description 34
- 230000015654 memory Effects 0.000 claims description 28
- 238000010586 diagram Methods 0.000 claims description 13
- 238000007619 statistical method Methods 0.000 claims description 11
- 238000012216 screening Methods 0.000 claims description 8
- 230000011218 segmentation Effects 0.000 description 22
- 238000001514 detection method Methods 0.000 description 13
- 238000013527 convolutional neural network Methods 0.000 description 9
- 238000004891 communication Methods 0.000 description 8
- 238000000605 extraction Methods 0.000 description 5
- 238000005457 optimization Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/02—Reliability analysis or reliability optimisation; Failure analysis, e.g. worst case scenario performance, failure mode and effects analysis [FMEA]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of image processing, in particular to a training method of a false positive recognition model, a target recognition method and a device, wherein the training method comprises the steps of obtaining a target recognition model and obtaining at least one disturbance target recognition model, wherein the disturbance target recognition model is obtained by training after carrying out random disturbance processing on output characteristic information of a preset convolution layer in the target recognition model; determining a first uncertainty feature vector based on the relationship between the first recognition result of each disturbance target recognition model on the target image and the second recognition result of the target recognition model on the target image; and training the false positive identification model according to the first uncertainty feature vector and the label of the target image, and determining the target false positive identification model. The uncertainty feature vector is utilized to train the false positive identification model, and then the false positive identification model is utilized to screen the identification result, so that the accuracy of the identification result can be improved.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a training method of a false positive identification model, a target identification method and a target identification device.
Background
The Convolutional Neural Network (CNN) is the most widely applied deep learning method in the field of image processing, and can automatically extract image features by inputting a picture into the CNN and then transmitting the picture into a plurality of convolutional layers and pooling layers in the CNN, so that feature representation of a target can be automatically learned layer by layer, and object layering abstraction and description can be realized, thereby being used for target identification. Among them, object recognition is widely used in various fields, such as unmanned driving and medical image segmentation.
Regardless of the method used to perform the components of the object recognition model, they all need to rely on a convolution layer to perform feature extraction, and the result of feature extraction directly affects the result of object recognition. For some targets with correct identification, the error of feature extraction has little influence on the identification result; for some targets with wrong identification (i.e. false positive), the target identification result is unstable under the condition of feature extraction errors, and the probability value is mainly reflected to have larger fluctuation, so that the accuracy of target identification can be influenced.
Disclosure of Invention
In view of the above, the embodiments of the present invention provide a training method for a false positive recognition model, a target recognition method and a device, so as to solve the problem of low target recognition accuracy.
According to a first aspect, an embodiment of the present invention provides a training method for a false positive identification model, where the training method includes:
obtaining a target recognition model and at least one disturbance target recognition model, wherein the disturbance target recognition model is obtained by training after carrying out random disturbance processing on output characteristic information of a preset convolution layer in the target recognition model;
determining a first uncertainty feature vector based on the relation between a first recognition result of each disturbance target recognition model on a target image and a second recognition result of the target recognition model on the target image;
and training the false positive identification model according to the first uncertainty feature vector and the label of the target image to determine a target false positive identification model.
According to the training method of the false positive identification model, on the basis of not changing the structure of the target identification model, the output characteristic information generated after convolution is subjected to random disturbance processing to obtain the first identification result under the random disturbance condition, the uncertainty characteristic vector of the identification result can be quantitatively evaluated based on the first identification result, the false positive identification model is trained by using the uncertainty characteristic vector, the false positive identification model with more accurate identification can be obtained, the identification result is screened by using the false positive identification model, and the accuracy of the identification result can be improved.
With reference to the first aspect, in a first implementation manner of the first aspect, the obtaining at least one disturbance target identification model includes:
inputting a sample image into the target recognition model, and extracting output characteristic information of a preset convolution layer in the target recognition model;
randomly selecting a plurality of disturbance coordinate points;
processing pixel values corresponding to the disturbance coordinate points in the output characteristic information by utilizing the disturbance coordinate points to obtain a disturbance characteristic diagram;
training the target recognition model based on the disturbance feature map to obtain the disturbance target recognition model.
According to the training method of the false positive identification model, provided by the embodiment of the invention, the disturbance processing is carried out on the output characteristic information in a mode of randomly selecting disturbance coordinate points, so that on one hand, the randomness of disturbance can be ensured, on the other hand, the processing mode is simple, and the training efficiency of the false positive identification model is improved.
With reference to the first implementation manner of the first aspect, in a second implementation manner of the first aspect, the processing, by using the disturbance coordinate point, a pixel value corresponding to the disturbance coordinate point in the output feature information to obtain a disturbance feature map includes:
Determining a disturbance area by taking each disturbance coordinate point as a center;
and setting a pixel value corresponding to the disturbance area in the output characteristic information as a preset value to obtain the disturbance characteristic diagram.
According to the training method of the false positive identification model, provided by the embodiment of the invention, the disturbance region is determined by taking the disturbance coordinate point as the center, and the pixel value corresponding to the disturbance region in the output characteristic information is processed to adjust the disturbance size, so that the reliability of the false positive identification model is further ensured.
With reference to the first aspect, in a third implementation manner of the first aspect, the determining a first uncertainty feature vector based on a relationship between a first recognition result of the disturbance target recognition model on a target image and a second recognition result of the target recognition model on the target image includes:
corresponding to each disturbance target identification model, calculating the overlapping rate of each first target identification area in the disturbance target identification model and a corresponding second target identification area in the target identification model;
adjusting the recognition probability of the first target recognition areas by utilizing the overlapping rate to obtain target recognition probabilities of the first target recognition areas corresponding to the target categories;
And carrying out statistical analysis on the target recognition probability corresponding to each disturbance target recognition model, and determining the first uncertainty feature vector.
According to the training method of the false positive identification model, provided by the embodiment of the invention, the identification probability of the identification result corresponding to the disturbance identification model is adjusted by utilizing the disturbance target identification model and the overlapping rate corresponding to the identification result of the target identification model, so that the identification probability can be changed along with the change of the overlapping rate, and the reliability of the false positive identification model is further ensured.
With reference to the first aspect, in a fourth implementation manner of the first aspect, the determining a first uncertainty feature vector based on a relationship between a first recognition result of the disturbance target recognition model on a target image and a second recognition result of the target recognition model on the target image includes:
calculating the average value of the probabilities that all pixels in the second target recognition area belong to each target category so as to obtain the target recognition probability of the first target recognition area corresponding to each target category;
and carrying out statistical analysis on the target recognition probability corresponding to each disturbance target recognition model, and determining the first uncertainty feature vector.
With reference to the third implementation manner or the fourth implementation manner of the first aspect, in a fifth implementation manner of the first aspect, the performing statistical analysis on the target recognition probabilities corresponding to the disturbance target recognition models to determine the first uncertainty feature vector includes:
calculating the statistic value of the target recognition probability corresponding to each target category by utilizing the target recognition probability corresponding to each disturbance target recognition model, wherein the statistic value of the target recognition probability comprises at least one of entropy, standard deviation, root mean square error, range and average absolute deviation;
and determining a first uncertainty characteristic vector corresponding to each target category based on the statistic value of the target recognition probability corresponding to each target category.
According to the training method for the false positive recognition model, the statistical value of the target recognition probability can accurately measure the heterogeneity or the heterogeneity of the data, so that the statistical value of the target recognition probability calculates the first uncertainty feature vector corresponding to each target category, and the reliability of a calculation result can be ensured.
With reference to the first aspect, in a sixth implementation manner of the first aspect, the training the false positive identification model according to the first uncertainty feature vector and the label of the target image, and determining a target false positive identification model includes:
Inputting the first uncertainty feature vector into the false positive identification model to obtain a prediction classification result;
and adjusting parameters of the false positive identification model by using the label of the target image and the prediction classification result to determine the target false positive identification model.
According to the training method for the false positive identification model, provided by the embodiment of the invention, the false positive rate can be reduced by training the false positive identification model by using the uncertain characteristic vector, and the accuracy rate of target identification is finally improved.
According to a second aspect, an embodiment of the present invention further provides a target identification method, where the identification method includes:
acquiring an image to be identified;
respectively inputting the image to be identified into a target identification model and at least one disturbance target identification model to respectively obtain a third identification result and a fourth identification result of the image to be identified, wherein the disturbance target identification model is obtained by training after carrying out random disturbance treatment on output characteristic information of a preset convolution layer in the target identification model;
determining a second uncertainty feature vector based on a relationship of the third recognition result and the fourth recognition result;
And inputting the second uncertainty feature vector into the false positive identification model, screening the third identification result, and determining a target identification result, wherein the false positive identification model is obtained by training according to the first aspect of the invention or the training method of the false positive identification model in any embodiment of the first aspect.
According to the target recognition method provided by the embodiment of the invention, the false positive recognition model is utilized to screen the target recognition result, so that the false positive rate can be reduced, the recognition accuracy is improved, and further optimization of the recognition result of the target recognition model is realized; the method has strong universality and can be applied to all target identification based on the convolutional neural network.
According to a third aspect, an embodiment of the present invention further provides a training device for a false positive identification model, where the training device includes:
the first acquisition module is used for acquiring a target recognition model and at least one disturbance target recognition model, wherein the disturbance target recognition model is obtained by training after carrying out random disturbance processing on output characteristic information of a preset convolution layer in the target recognition model;
the first determining module is used for determining a first uncertainty feature vector based on the relation between a first recognition result of each disturbance target recognition model on a target image and a second recognition result of the target recognition model on the target image;
And the training module is used for training the false positive identification model according to the first uncertainty feature vector and the label of the target image to determine a target false positive identification model.
According to the training device for the false positive identification model, provided by the embodiment of the invention, on the basis of not changing the structure of the target identification model, the output characteristic information generated after convolution is subjected to random disturbance processing to obtain the first identification result under the random disturbance condition, the uncertainty characteristic vector of the identification result can be quantitatively evaluated based on the first identification result, the false positive identification model is trained by using the uncertainty characteristic vector, the false positive identification model with more accurate identification can be obtained, the identification result is screened by using the false positive identification model, and the accuracy of the identification result can be improved.
According to a fourth aspect, an embodiment of the present invention further provides an object recognition apparatus, including:
the second acquisition module is used for acquiring the image to be identified;
the recognition module is used for respectively inputting the image to be recognized into a target recognition model and at least one disturbance target recognition model to respectively obtain a third recognition result and a fourth recognition result of the image to be recognized, wherein the disturbance target recognition model is obtained by training after carrying out random disturbance treatment on output characteristic information of a preset convolution layer in the target recognition model;
A second determining module, configured to determine a second uncertainty feature vector based on a relationship between the third recognition result and the fourth recognition result;
the screening module is configured to input the second uncertainty feature vector into the false positive identification model, screen the third identification result, and determine a target identification result, where the false positive identification model is obtained by training according to the first aspect of the present invention or the training method of the false positive identification model in any embodiment of the first aspect.
According to the target recognition device provided by the embodiment of the invention, the false positive recognition model is utilized to screen the target recognition result, so that the false positive rate can be reduced, the recognition accuracy is improved, and further optimization of the recognition result of the target recognition model is realized; the method has strong universality and can be applied to all target identification based on the convolutional neural network.
According to a fourth aspect, an embodiment of the present invention provides an electronic device, including: the processor executes the computer instructions, thereby executing the training method of the false positive identification model in the first aspect or any implementation manner of the first aspect, or executing the target identification method in the second aspect.
According to a fifth aspect, an embodiment of the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to perform the training method of the first aspect or any implementation manner of the first aspect, or the target recognition method of the second aspect.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are needed in the description of the embodiments or the prior art will be briefly described, and it is obvious that the drawings in the description below are some embodiments of the present invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a training method of a false positive identification model according to an embodiment of the present invention;
FIG. 2 is a flow chart of a training method of a false positive identification model according to an embodiment of the present invention;
FIG. 3 is a flow chart of a training method of a false positive identification model according to an embodiment of the present invention;
FIG. 4 is a flow chart of a target recognition method according to an embodiment of the invention;
FIG. 5 is a block diagram of a training apparatus for a false positive identification model according to an embodiment of the present invention;
FIG. 6 is a block diagram of a target recognition apparatus according to an embodiment of the present invention;
fig. 7 is a schematic diagram of a hardware structure of an electronic device according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The false positive identification model provided by the embodiment of the invention is a classification model and is used for judging whether the identification result is false positive or not. Specifically, in the target recognition process, an uncertainty feature vector can be formed by using a recognition result, and the uncertainty feature vector is input into a false positive recognition model so as to screen false positives in the recognition result, and a more accurate recognition result is obtained.
The target recognition may be target detection, target segmentation, etc., and is not limited in this regard. According to the method provided by the embodiment of the invention, the uncertainty characteristic vector corresponding to the identification result of the target identification model is calculated on the basis of not changing the network structure of the original target identification model; and then, the uncertainty feature vector is input into a false positive identification model to identify a false positive target detection or segmentation result, so that the false positive rate is reduced, further optimization of the target detection or segmentation result of the target identification model is realized, and finally, the accuracy is improved.
According to an embodiment of the present invention, there is provided a training method embodiment of a false positive identification model, it being noted that the steps shown in the flowcharts of the figures may be performed in a computer system such as a set of computer executable instructions, and that although a logical order is shown in the flowcharts, in some cases the steps shown or described may be performed in an order different from that shown or described herein.
In this embodiment, a training method of a false positive identification model is provided, which may be used in an electronic device, such as a computer, a mobile phone, a tablet computer, etc., fig. 1 is a flowchart of a training method of a false positive identification model according to an embodiment of the present invention, and as shown in fig. 1, the flowchart includes the following steps:
S11, acquiring a target recognition model and at least one disturbance target recognition model.
The disturbance target recognition model is obtained by training after carrying out random disturbance processing on output characteristic information of a preset convolution layer in the target recognition model.
The target recognition model is a model which can be used for target detection or target segmentation, the specific structure of the target recognition model is not limited, and the target recognition model can be set correspondingly according to actual conditions. For example, the object recognition model is an object detection or segmentation model based on a convolutional neural network. The target recognition model can be trained and stored in the electronic device in advance, can be obtained by training the electronic device in real time when training the false positive recognition model, can be obtained by the electronic device from the outside, and the like.
The object recognition model is used for recognizing objects in the image and probabilities that the objects correspond to various categories. For example, for target detection, the output of the target recognition model may be A bounding boxes B i (i=1, 2, …, a), each bounding box B i The corresponding probability value is PB i (c) Where C is the class of the target (c=1, 2, …, C); for object segmentation, the result of the output of the object recognition model may be C segmentation graphs S c (c＝1，2，…，C)，S c Each pixel represents the probability that the pixel belongs to the c-th object class, and the pixel which is larger than the preset threshold T is judged to belong to the object, so as to generate the segmentation result of the object, namely an area R consisting of similar pixels c 。
The number of the disturbance recognition models obtained by the electronic device may be one, two or more, and the specific number of the disturbance recognition models is not limited, and the disturbance recognition models can be set correspondingly according to actual situations. The disturbance recognition model is obtained by training after adding random disturbance to the output of a convolution layer of the disturbance recognition model on the basis of the target recognition model. Specifically, the output information of the convolution layer is the extracted characteristic information, and the accuracy of the characteristic information directly influences the accuracy of the identification result, so that random disturbance is added to the output information of the convolution layer, and the degree of disturbance is dequantized by using an uncertainty characteristic vector, so that the false positive identification model obtained through training has higher identification accuracy. From this, it can be seen that the inputs of the disturbance recognition model and the target recognition model are images, and the outputs are probabilities that each target in the images belongs to each category.
The predetermined convolution layer may be any one or several convolution layers in the target recognition model (for example, a convolution layer in a middle portion of the target recognition model), or may be all convolution layers, which is not limited in any way. And outputting the characteristic information from the preset convolution layer to the next module of the preset convolution layer for processing after random disturbance processing. For example, the preset convolution layer is the 2 nd convolution layer in the target recognition model, and in the target recognition model, the output of the 2 nd convolution layer is connected with the input of the 3 rd convolution layer; in the disturbance target identification model, the output of the 2 nd convolution layer is connected with the input of the 3 rd convolution layer after random disturbance processing.
The random perturbation may be a random modification of the output characteristic information, e.g. a random modification of a characteristic value in the output characteristic information, etc., as will be described in more detail below.
The random disturbance target recognition model can be obtained by training the electronic equipment in the process of training the false positive recognition model, can be stored in the electronic equipment after training in advance, or can be obtained by the electronic equipment from the outside, and the like.
S12, determining a first uncertainty feature vector based on the relation between the first recognition result of each disturbance target recognition model on the target image and the second recognition result of the target recognition model on the target image.
When the electronic equipment acquires the target recognition model and the disturbance target recognition model, the target images are respectively input into the models to obtain second recognition results corresponding to the disturbance target recognition models and second recognition results corresponding to the target recognition models.
As described above, the recognition result may be represented by a bounding box, or a segmentation map, and probabilities corresponding to respective categories. If the same object is identified, then the recognition result obtained using the perturbed object recognition model and the object recognition model is that there is a portion of intersection, e.g., intersection of bounding boxes. If the pixel values in the segmentation map belong to each other.
Using this relationship, the electronic device can determine a first uncertainty feature vector. For example, the relationship between the first recognition result and the second recognition result may measure a representation of random disturbances by the disturbance target recognition model, which represents uncertainty. Taking target detection as an example, if the intersection area is large, the recognition result of the disturbance target recognition model is reliable; if the intersection area is smaller, the recognition result of the disturbance target recognition model is inaccurate, and the recognition result of the part can be ignored in the calculation process of the uncertainty characteristic vector.
This step will be described in detail hereinafter, and is not described in detail here.
And S13, training the false positive identification model according to the first uncertainty feature vector and the label of the target image, and determining the target false positive identification model.
The label of the target image is used to indicate whether each target in the target image is of each category, and the specific labeling mode is not limited in this description.
The electronic equipment inputs the first uncertainty feature vector into a false positive identification model to obtain whether each identification result belongs to each category; and calculating a loss function by using the label of the target image, further training the false positive identification model, and determining the target false positive identification model. From this, it is known whether the input of the target false positive recognition model is an uncertainty feature vector and the output is of each category. Therefore, the false positive recognition model may be a classifier model, for example, constructed based on a neural network, a support vector machine, a random forest, or the like, and the specific network structure is not limited at all, and the corresponding setting may be performed according to actual situations.
According to the training method for the false positive identification model, on the basis of not changing the structure of the target identification model, the output characteristic information generated after convolution is subjected to random disturbance processing to obtain the first identification result under the random disturbance condition, the uncertainty characteristic vector of the identification result can be quantitatively evaluated based on the first identification result, the false positive identification model is trained by using the uncertainty characteristic vector, the false positive identification model with more accurate identification can be obtained, the identification result is screened by using the false positive identification model, and the accuracy of the identification result can be improved.
In this embodiment, a training method of a false positive identification model is provided, which may be used in an electronic device, such as a computer, a mobile phone, a tablet computer, etc., fig. 2 is a flowchart of a training method of a false positive identification model according to an embodiment of the present invention, and as shown in fig. 2, the flowchart includes the following steps:
s21, acquiring a target recognition model and at least one disturbance target recognition model.
The disturbance target recognition model is obtained by training after carrying out random disturbance processing on output characteristic information of a preset convolution layer in the target recognition model.
Specifically, the step S21 may include the following steps:
s211, acquiring a target recognition model.
Please refer to the related description of the object recognition model in S11 of the embodiment shown in fig. 1 in detail, which is not repeated here.
S212, inputting the sample image into a target recognition model, and extracting output characteristic information of a preset convolution layer in the target recognition model.
The electronic device inputs the sample image into a target recognition model, and the convolution layers in the target recognition model perform feature extraction on the sample image, and corresponding output feature information is provided for each convolution layer. The electronic equipment performs random disturbance processing on the output characteristic information of the preset convolution layer by extracting the output characteristic information.
For example, the object recognition model has Z convolutional layers, where the electronic device isN (N is not less than 1 and not more than Z) convolution layers are selected, wherein the selected convolution layers are marked as Conv s (s=1, 2, …, N). The output characteristic information of each convolution layer is expressed in the form of a characteristic diagram, and the Conv of the convolution layer s The generated feature map is F s (s＝1，2，…，N)。
S213, randomly selecting a plurality of disturbance coordinate points.
The electronic device may generate K (K>0) Random numbers, which are used as a plurality of disturbance coordinate points P k (k=1, 2, …, K). The specific value of K may be set correspondingly according to the actual situation, and the specific value of K is not limited herein, and may be set correspondingly according to the actual situation.
S214, processing pixel values corresponding to the disturbance coordinate points in the output characteristic information by using the disturbance coordinate points to obtain a disturbance characteristic diagram.
After obtaining the disturbance coordinate point, the electronic device may correspond the disturbance coordinate point to the output feature information, determine a pixel value of a pixel point of the disturbance coordinate point in the output feature information, and after determining the pixel value, the electronic device may set the pixel value to a preset value, for example, to 0, or other values. The disturbance map can be obtained after the setting.
In some optional implementations of this embodiment, the step S214 may include the following steps:
(1) And determining a disturbance area by taking each disturbance coordinate point as a center.
The electronic equipment uses the disturbance coordinate point P k A disturbance zone is defined for the center. The disturbance area may be a template area m with a size of l×l, or may be a template area with R as a radius, etc., and the shape of the disturbance area is not limited in any way.
(2) And setting a pixel value corresponding to the disturbance area in the output characteristic information as a preset value to obtain a disturbance characteristic diagram.
After the disturbance area is determined, the electronic equipment maps the disturbance area to the output characteristic information, determines the position of the disturbance area in the output characteristic information, and sets the pixel value corresponding to the disturbance area in the output characteristic information as a preset value, so that a disturbance characteristic diagram can be obtained. For example, the pixel value in the output characteristic information corresponding to the disturbance region is set to 0.
And determining a disturbance region by taking the disturbance coordinate point as a center, and processing pixel values corresponding to the disturbance region in the output characteristic information to adjust the disturbance size, so that the reliability of the false positive identification model is further ensured.
S215, training the target recognition model based on the disturbance feature map to obtain a disturbance target recognition model.
The electronic equipment trains the target recognition model on the basis of the disturbance characteristic diagram, and adjusts parameters of the target recognition model to obtain a disturbance target recognition model.
It should be noted that the number of obtained disturbance target recognition models may be set correspondingly according to actual situations. The training process of the plurality of disturbance target identification models may be repeated in steps S212 to S215. For example, the electronic device may set q random numbers to control the random process, and retrain the target recognition model using S212-S215 described above to obtain q disturbance target recognition models.
S22, determining a first uncertainty feature vector based on the relation between the first recognition result of each disturbance target recognition model on the target image and the second recognition result of the target recognition model on the target image.
Please refer to the embodiment S12 shown in fig. 1 in detail, which is not described herein.
S23, training the false positive identification model according to the first uncertainty feature vector and the label of the target image, and determining the target false positive identification model.
Please refer to the embodiment S13 shown in fig. 1 in detail, which is not described herein.
According to the training method of the false positive identification model, the output characteristic information is subjected to disturbance processing in a mode of randomly selecting disturbance coordinate points, so that on one hand, the randomness of disturbance can be guaranteed, on the other hand, the processing mode is simple, and the training efficiency of the false positive identification model is improved.
In this embodiment, a training method of a false positive identification model is provided, which may be used in an electronic device, such as a computer, a mobile phone, a tablet computer, etc., fig. 3 is a flowchart of a training method of a false positive identification model according to an embodiment of the present invention, and as shown in fig. 3, the flowchart includes the following steps:
s31, acquiring a target recognition model and at least one disturbance target recognition model.
The disturbance target recognition model is obtained by training after carrying out random disturbance processing on output characteristic information of a preset convolution layer in the target recognition model.
Please refer to the embodiment S21 shown in fig. 2 in detail, which is not described herein.
S32, determining a first uncertainty feature vector based on the relation between the first recognition result of each disturbance target recognition model on the target image and the second recognition result of the target recognition model on the target image.
Specifically, when the target recognition model is used for target detection, the above S32 may include the steps of:
s321, corresponding to each disturbance target identification model, calculating the overlapping rate of each first target identification area in the disturbance target identification model and the corresponding second target identification area in the target identification model.
After the disturbance target recognition model is obtained, the electronic equipment respectively utilizes each disturbance target recognition model to carry out target recognition on the target image, and a first target recognition area and the probability of the area corresponding to each category are obtained; and carrying out target recognition on the target image by using the target recognition model to obtain a second target recognition area and the probability of the area corresponding to each category.
After the first target recognition area and the second target recognition area are obtained, the electronic device can determine the overlapping rate by calculating the intersection ratio of the two areas.
Using the object recognition model M_D for object detection, and obtaining in S31For q disturbance target identification models, the disturbance target identification model is denoted as M_D o (o=1, 2, …, q). The second target recognition area output by the target recognition model M_D adopts a bounding box B i Representing the disturbance recognition model M_D o The output first target recognition area adopts a bounding box B o Representing, electronic device statistics M_D o And the output bounding box and Bi overlap ratio IOU.
S322, the recognition probability of the first target recognition areas is adjusted by utilizing the overlapping rate, and the target recognition probabilities of the first target recognition areas corresponding to the target categories are obtained.
The electronic device calculates the overlapping rate and the threshold IOU in the step S321 threshold And comparing, and adjusting the recognition probability of the first target recognition area based on the comparison result. Specifically, when the calculated overlap ratio is greater than the threshold IOU threshold At the time M_D o Output bounding box B o The corresponding recognition probability value is recorded as PB o (c) Wherein PB o (c) The recognition probability corresponding to the category c is output for the disturbance target recognition model. When with B i The overlap ratio IOU is less than the threshold IOU threshold Then M_D is taken o Output bounding box B o The corresponding recognition probability value is recorded as PBo (c) =0, thereby obtaining PB o (c)(o＝1，2，…，q)。
It should be noted that, the target recognition probabilities herein are that each first target recognition area corresponds to each target category.
S323, carrying out statistical analysis on the target recognition probability corresponding to each disturbance target recognition model, and determining a first uncertainty feature vector.
The statistical analysis can be to calculate the mean value, variance, etc. of the target recognition probabilities corresponding to the disturbance target recognition models under all target categories. And determining a first uncertainty characteristic vector by using the calculated statistical information, wherein each element in the vector is a specific value of each statistical information.
As an alternative implementation manner of this embodiment, the step S323 may include the following steps:
(1) Calculating the statistic value of the target recognition probability corresponding to each target category by utilizing the target recognition probability corresponding to each disturbance target recognition model, wherein the statistic value of the target recognition probability comprises at least one of entropy, standard deviation, root mean square error, range and average absolute deviation;
for example, the statistic value of the target recognition probability may be any combination of the above 5 statistic values, which is not limited in any way. In the following description, a detailed description will be given taking an example in which the first uncertain feature vector includes the statistical value of 5 above.
Taking object detection as an example, the first uncertain feature vector is vd= [ Vd1, vd2, vd3, vd4, vd5]:
vd4＝max(PB o (c))-min(PB o (c))
wherein,is PB o (c) (o=1, 2, …, q).
(2) And determining a first uncertainty characteristic vector corresponding to each target category based on the statistic value of the target recognition probability corresponding to each target category.
And the electronic equipment obtains the first uncertainty characteristic vector corresponding to each target category as Vd= [ Vd1, vd2, vd3, vd4, vd5] according to the calculated statistical value.
Because the statistic value of the target recognition probability can accurately measure the heterogeneity or the heterogeneity degree of the data, the statistic value of the target recognition probability calculates the first uncertainty feature vector corresponding to each target category, and the reliability of a calculation result can be ensured.
In some optional implementations of this embodiment, when the object recognition model is used for object segmentation, the step S32 may include the following steps:
(1) And calculating the average value of the probabilities that all pixels in the second target recognition area belong to each target category to obtain the target recognition probability of the first target recognition area corresponding to each target category.
Taking the target segmentation as an example, q retrained disturbance target recognition models M_S can be obtained o (o=1, 2, …, q). Obtaining M_S o The output segmentation graph, and the segmentation result R output by the target segmentation model M_S is found on the segmentation graph c Region, R in the calculation segmentation map c Average PS of probabilities that all pixels in the region belong to the c-th target class o (c)(o＝1，2，…，q)。
(2) And carrying out statistical analysis on the target recognition probability corresponding to each disturbance target recognition model, and determining a first uncertainty characteristic vector.
The specific statistical analysis method is similar to the statistical analysis method of the above-mentioned target detection, specifically, the first uncertain feature vector may be expressed as: vs= [ Vs1, vs2, vs3, vs4, vs5]:
vs4＝max(PS o (c))-min(PS o (c))
wherein,is the mean of PSo (c) (o=1, 2, …, q).
And S33, training the false positive identification model according to the first uncertainty feature vector and the label of the target image, and determining the target false positive identification model.
The reason for uncertainty is employed in the inventive embodiments: for easy identification of correct targets, even if some disturbance exists, the probability of the network prediction does not change greatly, but for identification of false (i.e. false positive) targets, the target identification result of the network is unstable and is reflected by large fluctuation of probability value. The degree of fluctuation is dequantized by using a statistical index that characterizes the degree of data non-uniformity, the set of statistical indices is defined as uncertainty feature vectors, and then a model is trained based on the uncertainty feature vectors, thereby identifying false positive results.
Specifically, the step S33 may include the following steps:
s331, inputting the first uncertainty feature vector into a false positive identification model to obtain a prediction classification result.
The electronic device takes the first uncertainty feature vector calculated in the step S32 as the input of a false positive identification model, and inputs the first uncertainty feature vector into the false positive identification model to obtain a prediction classification result. As described above, the false positive recognition model is a classifier model, and then the predicted classification result is whether the recognition result of the target recognition model is a false positive.
S332, adjusting parameters of the false positive identification model by using the label of the target image and the prediction classification result to determine the target false positive identification model.
For target detection, the tag is bounding box B i Whether it overlaps with object c; for target segmentation, the label is the segmentation result R c Whether the region belongs to object c. After training is completed, a recognition bounding box B can be constructed i Model Md of false positive or not FP Or to identify the segmentation result R c Model Ms of whether a region is false positive FP 。
It should be noted that, the training may be performed by using the first uncertainty feature vector corresponding to each target class, where the first uncertainty feature vector corresponds to each target class; or all the first uncertainty feature vectors corresponding to the target categories are combined into an uncertainty feature matrix, the false positive identification model is trained by using the uncertainty feature matrix, and the like, and the corresponding setting can be specifically performed according to actual conditions.
According to the training method for the false positive recognition model, the recognition probability of the recognition result corresponding to the disturbance recognition model is adjusted by using the disturbance target recognition model and the overlapping rate corresponding to the recognition result of the target recognition model, so that the recognition probability can be changed along with the change of the overlapping rate, and the reliability of the false positive recognition model is further ensured. Further, the false positive recognition model is trained by utilizing the uncertain feature vector, so that the false positive rate can be reduced, and the accuracy rate of target recognition is finally improved.
According to an embodiment of the present invention, there is provided an object recognition method embodiment, it being noted that the steps shown in the flowcharts of the drawings may be performed in a computer system such as a set of computer executable instructions, and that although a logical order is shown in the flowcharts, in some cases the steps shown or described may be performed in an order different from that herein.
In this embodiment, a target recognition method is provided, which may be used in an electronic device, such as a computer, a mobile phone, a tablet computer, etc., fig. 4 is a flowchart of a training method of a false positive recognition model according to an embodiment of the present invention, and as shown in fig. 4, the flowchart includes the following steps:
S41, acquiring an image to be identified.
The image to be identified can be obtained by the electronic device from the outside, or can be stored in the electronic device, and the manner of obtaining the image to be identified by the electronic device is not limited.
S42, respectively inputting the images to be identified into the target identification model and at least one disturbance target identification model to respectively obtain a third identification result and a fourth identification result of the images to be identified.
The disturbance target recognition model is obtained by training after carrying out random disturbance processing on output characteristic information of a preset convolution layer in the target recognition model.
The step is similar to the method of obtaining the first recognition result and the second recognition result in the above embodiment, and the detailed description is omitted herein.
S43, determining a second uncertainty characteristic vector based on the relation between the third identification result and the fourth identification result.
This step is similar to the determination of the first uncertainty feature vector in the above embodiment, and is described in detail above, and will not be repeated here.
S44, inputting the second uncertainty feature vector into the false positive identification model, screening the third identification result, and determining the target identification result.
The false positive identification model is trained according to the training method of the false positive identification model in the embodiment of the invention.
And the electronic equipment inputs the second uncertainty feature vector corresponding to the obtained image to be identified into a false positive identification model, and screens the third identification result to determine the target identification result.
Specifically, the second uncertainty feature vector corresponds to the recognition result corresponding to each target category in the image to be recognized, so that each recognition result of different target categories can be characterized by using the second uncertainty feature vector, and the false positive recognition model is used for screening the false positive recognition result, so that screening of the recognition result can be realized, and the target recognition result can be obtained.
For example, the electronic device, after obtaining the second uncertainty feature vector of each bounding box or segmented result region, inputs it to the established false positive identification model Md FP Or Ms FP Judgment of bounding Box B i Or the segmentation result R c If the region belongs to false positive, the result is removed, and after the false positive result is removed, the accuracy of detection or segmentation is improved.
According to the target recognition method provided by the embodiment, the false positive recognition model is utilized to screen the target recognition result, so that the false positive rate can be reduced, the recognition accuracy is improved, and further optimization of the recognition result of the target recognition model is realized; the method has strong universality and can be applied to all target identification based on the convolutional neural network.
In this embodiment, a training device for a false positive recognition model, or a target recognition device, which is used to implement the foregoing embodiments and preferred embodiments, is provided, and details thereof are omitted. As used below, the term "module" may be a combination of software and/or hardware that implements a predetermined function. While the means described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and contemplated.
The present embodiment provides a training device for a false positive identification model, as shown in fig. 5, including:
the first obtaining module 51 is configured to obtain a target recognition model and obtain at least one disturbance target recognition model, where the disturbance target recognition model is obtained by performing random disturbance processing on output feature information of a preset convolution layer in the target recognition model, and then training the disturbance target recognition model;
A first determining module 52, configured to determine a first uncertainty feature vector based on a relationship between a first recognition result of each disturbance target recognition model on a target image and a second recognition result of the target recognition model on the target image;
the training module 53 is configured to train the false positive identification model according to the first uncertainty feature vector and the label of the target image, and determine a target false positive identification model.
According to the training device for the false positive identification model, on the basis of not changing the structure of the target identification model, the output characteristic information generated after convolution is subjected to random disturbance processing to obtain the first identification result under the random disturbance condition, the uncertainty characteristic vector of the identification result can be quantitatively evaluated based on the first identification result, the false positive identification model is trained by using the uncertainty characteristic vector, the false positive identification model with more accurate identification can be obtained, the identification result is screened by using the false positive identification model, and the accuracy of the identification result can be improved.
The present embodiment also provides an object recognition apparatus, as shown in fig. 6, including:
A second acquiring module 61, configured to acquire an image to be identified;
the recognition module 62 is configured to input the image to be recognized into a target recognition model and at least one disturbance target recognition model, to obtain a third recognition result and a fourth recognition result of the image to be recognized, where the disturbance target recognition model is obtained by performing random disturbance processing on output feature information of a preset convolution layer in the target recognition model, and then training the disturbance target recognition model;
a second determining module 63, configured to determine a second uncertainty feature vector based on a relationship between the third recognition result and the fourth recognition result;
the screening module 64 is configured to input the second uncertainty feature vector into the false positive identification model, screen the third identification result, and determine a target identification result, where the false positive identification model is trained according to the training method of the false positive identification model according to any one of the foregoing embodiments of the present invention.
According to the target recognition device provided by the embodiment, the false positive recognition model is utilized to screen the target recognition result, so that the false positive rate can be reduced, the recognition accuracy is improved, and further optimization of the recognition result of the target recognition model is realized; the method has strong universality and can be applied to all target identification based on the convolutional neural network.
The training means of the false positive identification model, or the object identification means in this embodiment is presented in the form of functional units, here referred to as ASIC circuits, processors and memories executing one or more software or firmware programs, and/or other devices that can provide the above described functionality.
Further functional descriptions of the above respective modules are the same as those of the above corresponding embodiments, and are not repeated here.
The embodiment of the invention also provides electronic equipment, which is provided with the training device of the false positive identification model shown in the figure 5 or the target identification device shown in the figure 6.
Referring to fig. 7, fig. 7 is a schematic structural diagram of an electronic device according to an alternative embodiment of the present invention, and as shown in fig. 7, the electronic device may include: at least one processor 71, such as a CPU (Central Processing Unit ), at least one communication interface 73, a memory 74, at least one communication bus 72. Wherein the communication bus 72 is used to enable connected communication between these components. The communication interface 73 may include a Display screen (Display) and a Keyboard (Keyboard), and the optional communication interface 73 may further include a standard wired interface and a wireless interface. The memory 74 may be a high-speed RAM memory (Random Access Memory, volatile random access memory) or a non-volatile memory (non-volatile memory), such as at least one disk memory. The memory 74 may alternatively be at least one memory device located remotely from the processor 71. Where the processor 71 may be in conjunction with the apparatus described in fig. 5 or 6, the memory 74 stores an application program, and the processor 71 invokes the program code stored in the memory 74 for performing any of the method steps described above.
The communication bus 72 may be a peripheral component interconnect standard (peripheral component interconnect, PCI) bus, an extended industry standard architecture (extended industry standard architecture, EISA) bus, or the like. The communication bus 72 may be divided into an address bus, a data bus, a control bus, and the like. For ease of illustration, only one thick line is shown in fig. 7, but not only one bus or one type of bus.
Wherein the memory 74 may include volatile memory (English) such as random-access memory (RAM); the memory may also include a nonvolatile memory (english: non-volatile memory), such as a flash memory (english: flash memory), a hard disk (english: hard disk drive, abbreviated as HDD) or a solid state disk (english: solid-state drive, abbreviated as SSD); memory 74 may also include a combination of the above types of memory.
The processor 71 may be a central processor (English: central processing unit, abbreviated: CPU), a network processor (English: network processor, abbreviated: NP) or a combination of CPU and NP.
The processor 71 may further include a hardware chip, among others. The hardware chip may be an application-specific integrated circuit (ASIC), a Programmable Logic Device (PLD), or a combination thereof (English: programmable logic device). The PLD may be a complex programmable logic device (English: complex programmable logic device, abbreviated: CPLD), a field programmable gate array (English: field-programmable gate array, abbreviated: FPGA), a general-purpose array logic (English: generic array logic, abbreviated: GAL), or any combination thereof.
Optionally, the memory 74 is also used for storing program instructions. Processor 71 may invoke program instructions to implement the training method of the false positive identification model as shown in the embodiments of fig. 1-3 of the present application, or the target identification method shown in the embodiment of fig. 4.
The embodiment of the invention also provides a non-transitory computer storage medium, which stores computer executable instructions, and the computer executable instructions can execute the training method of the false positive identification model or the target identification method in any of the method embodiments. Wherein the storage medium may be a magnetic Disk, an optical Disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a Flash Memory (Flash Memory), a Hard Disk (HDD), or a Solid State Drive (SSD); the storage medium may also comprise a combination of memories of the kind described above.
Although embodiments of the present invention have been described in connection with the accompanying drawings, various modifications and variations may be made by those skilled in the art without departing from the spirit and scope of the invention, and such modifications and variations fall within the scope of the invention as defined by the appended claims.
Claims (12)
1. A training method for a false positive recognition model, the training method comprising:
obtaining a target recognition model and at least one disturbance target recognition model, wherein the disturbance target recognition model is obtained by training after carrying out random disturbance processing on output characteristic information of a preset convolution layer in the target recognition model;
determining a first uncertainty feature vector based on the relation between a first recognition result of each disturbance target recognition model on a target image and a second recognition result of the target recognition model on the target image;
and training the false positive identification model according to the first uncertainty feature vector and the label of the target image to determine a target false positive identification model.
2. The training method of claim 1, wherein the obtaining at least one disturbance target identification model comprises:
inputting a sample image into the target recognition model, and extracting output characteristic information of a preset convolution layer in the target recognition model;
randomly selecting a plurality of disturbance coordinate points;
processing pixel values corresponding to the disturbance coordinate points in the output characteristic information by utilizing the disturbance coordinate points to obtain a disturbance characteristic diagram;
Training the target recognition model based on the disturbance feature map to obtain the disturbance target recognition model.
3. The training method according to claim 2, wherein the processing, with the disturbance coordinate point, the pixel value corresponding to the disturbance coordinate point in the output feature information to obtain a disturbance feature map includes:
determining a disturbance area by taking each disturbance coordinate point as a center;
and setting a pixel value corresponding to the disturbance area in the output characteristic information as a preset value to obtain the disturbance characteristic diagram.
4. The training method of claim 1, wherein the determining a first uncertainty feature vector based on a relationship between a first recognition result of the disturbance target recognition model on a target image and a second recognition result of the target recognition model on the target image comprises:
corresponding to each disturbance target identification model, calculating the overlapping rate of each first target identification area in the disturbance target identification model and a corresponding second target identification area in the target identification model;
adjusting the recognition probability of the first target recognition areas by utilizing the overlapping rate to obtain target recognition probabilities of the first target recognition areas corresponding to the target categories;
And carrying out statistical analysis on the target recognition probability corresponding to each disturbance target recognition model, and determining the first uncertainty feature vector.
5. The training method of claim 1, wherein the determining a first uncertainty feature vector based on a relationship between a first recognition result of the disturbance target recognition model on a target image and a second recognition result of the target recognition model on the target image comprises:
calculating the average value of probabilities that all pixels in a first target recognition area corresponding to a second target recognition area belong to each target category to obtain target recognition probabilities that the first target recognition area corresponds to each target category, wherein the first target recognition area corresponds to the first recognition result, and the second target recognition area corresponds to the second recognition result;
and carrying out statistical analysis on the target recognition probability corresponding to each disturbance target recognition model, and determining the first uncertainty feature vector.
6. The training method of claim 4 or 5, wherein the performing a statistical analysis on the target recognition probabilities corresponding to the disturbance target recognition models to determine the first uncertainty feature vector includes:
Calculating the statistic value of the target recognition probability corresponding to each target category by utilizing the target recognition probability corresponding to each disturbance target recognition model, wherein the statistic value of the target recognition probability comprises at least one of entropy, standard deviation, root mean square error, range and average absolute deviation;
and determining a first uncertainty characteristic vector corresponding to each target category based on the statistic value of the target recognition probability corresponding to each target category.
7. The training method of claim 1, wherein the training the false positive recognition model based on the first uncertainty feature vector and the label of the target image to determine a target false positive recognition model comprises:
inputting the first uncertainty feature vector into the false positive identification model to obtain a prediction classification result;
and adjusting parameters of the false positive identification model by using the label of the target image and the prediction classification result to determine the target false positive identification model.
8. A method of identifying a target, the method comprising:
acquiring an image to be identified;
Respectively inputting the image to be identified into a target identification model and at least one disturbance target identification model to respectively obtain a third identification result and a fourth identification result of the image to be identified, wherein the disturbance target identification model is obtained by training after carrying out random disturbance treatment on output characteristic information of a preset convolution layer in the target identification model;
determining a second uncertainty feature vector based on a relationship of the third recognition result and the fourth recognition result;
inputting the second uncertainty feature vector into the false positive identification model, screening the third identification result, and determining a target identification result, wherein the false positive identification model is trained according to the training method of the false positive identification model of any one of claims 1-7.
9. A training device for a false positive identification model, the training device comprising:
the first acquisition module is used for acquiring a target recognition model and at least one disturbance target recognition model, wherein the disturbance target recognition model is obtained by training after carrying out random disturbance processing on output characteristic information of a preset convolution layer in the target recognition model;
The first determining module is used for determining a first uncertainty feature vector based on the relation between a first recognition result of each disturbance target recognition model on a target image and a second recognition result of the target recognition model on the target image;
and the training module is used for training the false positive identification model according to the first uncertainty feature vector and the label of the target image to determine a target false positive identification model.
10. An object recognition apparatus, characterized in that the recognition apparatus comprises:
the second acquisition module is used for acquiring the image to be identified;
the recognition module is used for respectively inputting the image to be recognized into a target recognition model and at least one disturbance target recognition model to respectively obtain a third recognition result and a fourth recognition result of the image to be recognized, wherein the disturbance target recognition model is obtained by training after carrying out random disturbance treatment on output characteristic information of a preset convolution layer in the target recognition model;
a second determining module, configured to determine a second uncertainty feature vector based on a relationship between the third recognition result and the fourth recognition result;
The screening module is configured to input the second uncertainty feature vector into the false positive identification model, screen the third identification result, and determine a target identification result, where the false positive identification model is trained according to the training method of the false positive identification model according to any one of claims 1-7.
11. An electronic device, comprising:
a memory and a processor, the memory and the processor being communicatively connected to each other, the memory having stored therein computer instructions, the processor executing the computer instructions to perform the training method of the false positive identification model of any one of claims 1-7, or the target identification method of claim 8.
12. A computer-readable storage medium storing computer instructions for causing a computer to perform the training method of the false positive identification model according to any one of claims 1 to 7 or the target identification method according to claim 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110113779.8A CN112784494B (en) | 2021-01-27 | 2021-01-27 | Training method of false positive recognition model, target recognition method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110113779.8A CN112784494B (en) | 2021-01-27 | 2021-01-27 | Training method of false positive recognition model, target recognition method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112784494A CN112784494A (en) | 2021-05-11 |
| CN112784494B true CN112784494B (en) | 2024-02-06 |
Family
ID=75759170
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110113779.8A Active CN112784494B (en) | 2021-01-27 | 2021-01-27 | Training method of false positive recognition model, target recognition method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112784494B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113408650B (en) * | 2021-07-12 | 2023-07-18 | 厦门大学 | Semi-supervised three-dimensional shape recognition method based on consistency training |
| CN114638829A (en) * | 2022-05-18 | 2022-06-17 | 安徽数智建造研究院有限公司 | Anti-interference training method of tunnel lining detection model and tunnel lining detection method |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110070034A (en) * | 2019-04-19 | 2019-07-30 | 北京朗镜科技有限责任公司 | Model training method, section recognition methods, device, equipment and medium |
| CN111062871A (en) * | 2019-12-17 | 2020-04-24 | 腾讯科技（深圳）有限公司 | Image processing method and device, computer equipment and readable storage medium |
| CN111401396A (en) * | 2019-01-03 | 2020-07-10 | 阿里巴巴集团控股有限公司 | Image recognition method and device |
| CN111523597A (en) * | 2020-04-23 | 2020-08-11 | 北京百度网讯科技有限公司 | Target recognition model training method, device, equipment and storage medium |
| WO2020215557A1 (en) * | 2019-04-24 | 2020-10-29 | 平安科技（深圳）有限公司 | Medical image interpretation method and apparatus, computer device and storage medium |
| CN111914939A (en) * | 2020-08-06 | 2020-11-10 | 平安科技（深圳）有限公司 | Method, device and equipment for identifying blurred image and computer readable storage medium |
| CN112116613A (en) * | 2020-09-25 | 2020-12-22 | 贝壳技术有限公司 | Model training method, image segmentation method, image vectorization method and system thereof |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106485192B (en) * | 2015-09-02 | 2019-12-06 | 富士通株式会社 | Training method and device of neural network for image recognition |
-
2021
- 2021-01-27 CN CN202110113779.8A patent/CN112784494B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111401396A (en) * | 2019-01-03 | 2020-07-10 | 阿里巴巴集团控股有限公司 | Image recognition method and device |
| CN110070034A (en) * | 2019-04-19 | 2019-07-30 | 北京朗镜科技有限责任公司 | Model training method, section recognition methods, device, equipment and medium |
| WO2020215557A1 (en) * | 2019-04-24 | 2020-10-29 | 平安科技（深圳）有限公司 | Medical image interpretation method and apparatus, computer device and storage medium |
| CN111062871A (en) * | 2019-12-17 | 2020-04-24 | 腾讯科技（深圳）有限公司 | Image processing method and device, computer equipment and readable storage medium |
| CN111523597A (en) * | 2020-04-23 | 2020-08-11 | 北京百度网讯科技有限公司 | Target recognition model training method, device, equipment and storage medium |
| CN111914939A (en) * | 2020-08-06 | 2020-11-10 | 平安科技（深圳）有限公司 | Method, device and equipment for identifying blurred image and computer readable storage medium |
| CN112116613A (en) * | 2020-09-25 | 2020-12-22 | 贝壳技术有限公司 | Model training method, image segmentation method, image vectorization method and system thereof |
Non-Patent Citations (1)
| Title |
|---|
| 基于相空间重构和卷积神经网络的电能质量扰动分类;陈伟;何家欢;裴喜平;;电力系统保护与控制(14);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112784494A (en) | 2021-05-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110020592B (en) | Object detection model training method, device, computer equipment and storage medium | |
| WO2021129105A1 (en) | Mask rcnn network model-based target identification method and apparatus | |
| CN110378837B (en) | Target detection method and device based on fish-eye camera and storage medium | |
| CN112784494B (en) | Training method of false positive recognition model, target recognition method and device | |
| EP3702957A1 (en) | Target detection method and apparatus, and computer device | |
| CN111242026B (en) | Remote sensing image target detection method based on spatial hierarchy perception module and metric learning | |
| CN109858327B (en) | Character segmentation method based on deep learning | |
| CN109598298B (en) | Image object recognition method and system | |
| CN112419202B (en) | Automatic wild animal image recognition system based on big data and deep learning | |
| CN112861785B (en) | Instance segmentation and image restoration-based pedestrian re-identification method with shielding function | |
| CN111368682A (en) | Method and system for detecting and identifying station caption based on faster RCNN | |
| CN111310826A (en) | Method and device for detecting labeling abnormity of sample set and electronic equipment | |
| WO2023160666A1 (en) | Target detection method and apparatus, and target detection model training method and apparatus | |
| CN116740758A (en) | Bird image recognition method and system for preventing misjudgment | |
| CN115880260A (en) | Method, device and equipment for detecting base station construction and computer readable storage medium | |
| CN111539456A (en) | Target identification method and device | |
| CN111310837A (en) | Vehicle refitting recognition method, device, system, medium and equipment | |
| CN112308061B (en) | License plate character recognition method and device | |
| CN116052189A (en) | Text recognition method, system and storage medium | |
| CN116363583A (en) | Human body identification method, device, equipment and medium for top view angle | |
| CN114927236A (en) | Detection method and system for multiple target images | |
| CN115620083A (en) | Model training method, face image quality evaluation method, device and medium | |
| CN112699809B (en) | Vaccinia category identification method, device, computer equipment and storage medium | |
| CN115424250A (en) | License plate recognition method and device | |
| CN108830166B (en) | Real-time bus passenger flow volume statistical method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |