CN111222409A - Vehicle brand labeling method, device and system - Google Patents
Vehicle brand labeling method, device and system Download PDFInfo
- Publication number
- CN111222409A CN111222409A CN201911179667.1A CN201911179667A CN111222409A CN 111222409 A CN111222409 A CN 111222409A CN 201911179667 A CN201911179667 A CN 201911179667A CN 111222409 A CN111222409 A CN 111222409A
- Authority
- CN
- China
- Prior art keywords
- brand
- image
- vehicle
- sub
- marked
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000002372 labelling Methods 0.000 title claims description 59
- 238000003062 neural network model Methods 0.000 claims abstract description 53
- 238000012549 training Methods 0.000 claims abstract description 49
- 238000000034 method Methods 0.000 claims abstract description 44
- 238000004590 computer program Methods 0.000 claims description 11
- 238000012937 correction Methods 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 9
- 238000001514 detection method Methods 0.000 claims description 8
- 230000004888 barrier function Effects 0.000 claims description 6
- 238000012545 processing Methods 0.000 abstract description 6
- 230000006870 function Effects 0.000 description 14
- 238000013527 convolutional neural network Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
- G06V10/464—Salient features, e.g. scale invariant feature transforms [SIFT] using a plurality of salient features, e.g. bag-of-words [BoW] representations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/08—Detecting or categorising vehicles
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a method, a device and a system for marking a vehicle brand, which relate to the technical field of image processing, and the method comprises the following steps: acquiring an image to be marked of a key area containing a vehicle; inputting the image to be marked into a pre-trained neural network model to identify and mark the main brand of the vehicle; the neural network model is obtained by training a vehicle training image based on a marked vehicle main brand; acquiring a sub-brand image set corresponding to the identified main brand of the vehicle; performing sub-brand identification on the image to be annotated based on the sub-brand image set to obtain a sub-brand of the vehicle; and marking the sub-brand of the vehicle in the image to be marked. The invention can effectively improve the marking efficiency of the vehicle brand.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a vehicle brand labeling method, device and system.
Background
Identification of the brand of the vehicle is of great importance for retrieving the target vehicle from the image or video. In order to automatically identify the vehicle brand, it is necessary to rely on a large amount of data labeled with the vehicle brand. Currently, for the labeling of vehicle brands, especially for the labeling of vehicle brands accurate to sub-brands or annual money of vehicles, professional labeling personnel mainly label the identified vehicle brands; wherein the annotating personnel need to sort and accumulate vehicle brand knowledge for a longer time. The manual labeling mode consumes a large amount of manpower, is poor in timeliness and is low in labeling efficiency.
Disclosure of Invention
In view of this, the present invention provides a method, an apparatus and a system for labeling a vehicle brand, which can effectively improve the labeling efficiency of the vehicle brand.
In order to achieve the above purpose, the embodiment of the present invention adopts the following technical solutions:
in a first aspect, an embodiment of the present invention provides a method for labeling a brand of a vehicle, where the method includes: acquiring an image to be marked of a key area containing a vehicle; inputting the image to be marked into a pre-trained neural network model to identify and mark a vehicle main brand; the neural network model is obtained by training a vehicle training image based on a marked vehicle main brand; acquiring a sub-brand image set corresponding to the identified main brand of the vehicle; performing sub-brand identification on the image to be marked based on the sub-brand image set to obtain a sub-brand of the vehicle; and marking the sub-brand of the vehicle in the image to be marked.
Further, the step of obtaining the image to be labeled including the key area of the vehicle includes: acquiring an original image containing a vehicle; detecting whether a preset key area exists in the original image; the key area comprises at least one of a car lamp area, a car logo area, a hub area and a barrier area; if yes, scoring the quality of the key area according to the definition factor and/or the distortion factor; and when the quality score is higher than a preset score threshold value, determining the original image as an image to be annotated.
Further, the sub-brand image sets are multiple, and different sub-brand image sets correspond to different vehicle parts; the step of identifying the sub-brand in the image to be marked based on the sub-brand image set to obtain the sub-brand of the vehicle comprises the following steps: acquiring the quality score of each key area in the image to be annotated; determining a first vehicle part contained in the key area with the highest quality score; selecting a first target sub-brand image set from the plurality of sub-brand image sets according to the determined first vehicle location; and determining the sub-brand of the vehicle by comparing the image to be labeled with the first target sub-brand image set.
Further, the sub-brand image sets are multiple, and different sub-brand image sets correspond to different vehicle parts; the step of identifying the sub-brand in the image to be marked based on the sub-brand image set to obtain the sub-brand of the vehicle comprises the following steps: determining a second vehicle part contained in each key area in the image to be marked; selecting a second target sub-brand image set corresponding to each key area from the plurality of sub-brand image sets according to the determined second vehicle part; calculating a first matching degree of each key area in the image to be labeled and a second target sub-brand image set corresponding to the key area, and determining an initial sub-brand of the key area according to the first matching degree; judging whether the initial sub-brands corresponding to different key areas are the same or not; if so, determining the initial sub-brand as a sub-brand of the vehicle; and if not, determining the initial sub-brand corresponding to the maximum matching degree in the first matching degrees as the sub-brand of the vehicle.
Further, the second target sub-brand image set stores reference images of the labeled sub-brands of the vehicles; the step of calculating a first matching degree of the key area and a second target sub-brand image set corresponding to the key area, and determining an initial sub-brand of the key area according to the first matching degree comprises the following steps: performing key point detection on the key area through a preset key point detection model to obtain first key point information of the key area; acquiring second key point information preset by each reference image in a second target sub-brand image set corresponding to the key area; obtaining a second matching degree of the key area and each reference image by comparing the first key point information with the second key point information of each reference image; determining the maximum matching degree in the second matching degrees as the first matching degree of the key area and a second target sub-brand image set corresponding to the key area; and acquiring the sub-brand of the vehicle marked by the reference image corresponding to the first matching degree, and determining the acquired sub-brand of the vehicle as the initial sub-brand of the key area.
Further, the step of inputting the image to be labeled into a pre-trained neural network model for identifying and labeling the main brand of the vehicle comprises: inputting the image to be marked into a pre-trained neural network model to identify a vehicle main brand to obtain an initial main brand; acquiring a discrimination operation for the initial main brand; if the judging operation is the operation for determining that the initial main brand is correct, marking the main brand of the vehicle in the image to be marked according to the initial main brand; if the judging operation is an operation for determining that the initial main brand is wrong, responding to a correction operation aiming at the initial main brand, and acquiring a corrected main brand corresponding to the correction operation; and marking the main brand of the vehicle in the image to be marked according to the corrected main brand.
Further, the training process of the neural network model comprises: inputting a plurality of vehicle training images with marked vehicle main brands to a currently trained neural network model; setting a marked vehicle main brand of the vehicle training image as a reference main brand; identifying the vehicle training image through a currently trained neural network model to obtain a predicted main brand of the vehicle training image; calculating a loss function value between a prediction main brand and a reference main brand corresponding to the vehicle training image; wherein the loss function value is used to evaluate a primary brand recognition accuracy of the vehicle training image; and adjusting the parameters of the currently trained neural network model through a back propagation algorithm according to the loss function value until the loss function value converges to a preset value, and finishing the training.
In a second aspect, an embodiment of the present invention further provides a vehicle brand labeling apparatus, including: the image acquisition module is used for acquiring an image to be marked of a key area containing a vehicle; the main brand labeling module is used for inputting the image to be labeled to a pre-trained neural network model to identify and label a vehicle main brand; the neural network model is obtained by training a vehicle training image based on a marked vehicle main brand; the sub-brand image acquisition module is used for acquiring a sub-brand image set corresponding to the identified main brand of the vehicle; the sub-brand identification module is used for carrying out sub-brand identification on the image to be marked based on the sub-brand image set to obtain a sub-brand of the vehicle; and the sub-brand marking module is used for marking the sub-brand of the vehicle in the image to be marked.
In a third aspect, an embodiment of the present invention provides a vehicle brand labeling system, including: the device comprises an image acquisition device, a processor and a storage device; the image acquisition device is used for acquiring an image to be marked; the storage means has stored thereon a computer program which, when executed by the processor, performs the method of any of the first aspects.
In a fourth aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, performs the steps of the method according to any one of the above first aspects.
The embodiment of the invention provides a method, a device and a system for marking a vehicle brand, which comprises the steps of firstly inputting an acquired image to be marked into a pre-trained neural network model to identify and mark a vehicle main brand; then acquiring a sub-brand image set corresponding to the identified main brand of the vehicle, and performing sub-brand identification on the image to be marked based on the sub-brand image set to obtain a sub-brand of the vehicle; and finally, marking the sub-brand of the vehicle in the image to be marked. Compared with the existing manual labeling mode, the mode provided by the embodiment can effectively reduce the manpower requirement and improve the labeling efficiency of the main brand by identifying the main brand of the vehicle of the image to be labeled by utilizing the neural network model; the sub-brand image set can better meet the requirement of fast recognition of vehicle brands under fine granularity; the identification process from the main brand to the sub-brand is integrated, and the labeling efficiency of the vehicle brand is greatly improved.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the above-described technology of the disclosure.
In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a schematic structural diagram of an electronic device according to an embodiment of the present invention;
FIG. 2 is a flow chart of a method for branding a vehicle according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a set of sub-brand images provided by an embodiment of the present invention;
fig. 4 shows a block diagram of a labeling apparatus for a brand of vehicle according to an embodiment of the present invention.
Detailed Description
To make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In view of the problem that the existing license plate labeling mode relying on manual work has low vehicle brand labeling efficiency, the embodiment of the invention provides a vehicle brand labeling method, device and system.
The first embodiment is as follows:
first, an example electronic device 100 for implementing the labeling method, apparatus, and system of the vehicle brand according to the embodiment of the present invention is described with reference to fig. 1.
As shown in fig. 1, an electronic device 100 includes one or more processors 102, one or more memory devices 104, an input device 106, an output device 108, and an image capture device 110, which are interconnected via a bus system 112 and/or other type of connection mechanism (not shown). It should be noted that the components and structure of the electronic device 100 shown in fig. 1 are only exemplary and not limiting, and the electronic device may have some of the components shown in fig. 1 and may also have other components and structures not shown in fig. 1, as desired.
The processor 102 may be a Central Processing Unit (CPU) or other form of processing unit having data processing capabilities and/or instruction execution capabilities, and may control other components in the electronic device 100 to perform desired functions.
The storage 104 may include one or more computer program products that may include various forms of computer-readable storage media, such as volatile memory and/or non-volatile memory. The volatile memory may include, for example, Random Access Memory (RAM), cache memory (cache), and/or the like. The non-volatile memory may include, for example, Read Only Memory (ROM), hard disk, flash memory, etc. On which one or more computer program instructions may be stored that may be executed by processor 102 to implement client-side functionality (implemented by the processor) and/or other desired functionality in embodiments of the invention described below. Various applications and various data, such as various data used and/or generated by the applications, may also be stored in the computer-readable storage medium.
The input device 106 may be a device used by a user to input instructions and may include one or more of a keyboard, a mouse, a microphone, a touch screen, and the like.
The output device 108 may output various information (e.g., images or sounds) to the outside (e.g., a user), and may include one or more of a display, a speaker, and the like.
The image capture device 110 may take images (e.g., photographs, videos, etc.) desired by the user and store the taken images in the storage device 104 for use by other components.
For example, an example electronic device for implementing a vehicle brand labeling method, apparatus and system according to an embodiment of the present invention may be implemented on a smart terminal such as a smart phone, a tablet computer, a camera, and the like.
Example two:
referring to the flowchart of a method for labeling a brand of a vehicle shown in fig. 2, the method specifically includes the following steps S202 to S210:
step S202, acquiring an image to be annotated containing a key area of the vehicle. The image to be marked is an image containing at least one key area of the vehicle; the key area is an area including a specific part capable of identifying the brand of the vehicle, such as an area including a vehicle logo, an area including a vehicle lamp, an area including a wheel hub, an area including a grille, an area for protecting vehicle lines and the like.
Step S204, inputting the image to be labeled into a pre-trained neural network model to identify and label the main brand of the vehicle; the neural network model is obtained by training a vehicle training image based on the marked vehicle main brand.
In some optional embodiments, the main brand identification and labeling may be performed on the image to be labeled based on a Neural network model such as a CNN (Convolutional Neural Networks), an R-CNN (Region-CNN) network, or a Segnet network, so as to obtain the main brand of the vehicle in the image to be labeled. The main brand is used for distinguishing vehicle categories under larger granularity, and the main brand is usually the automobile manufacturer, such as gallo, pommel, audi, honda, and the like.
Step S206, a sub-brand image set corresponding to the identified main brand of the vehicle is obtained.
It will be appreciated that after the vehicle's main brand is determined, all sub-brands under the main brand flag may be readily available. The sub-brand image set is a set in which images of sub-brands of tagged vehicles are stored. In this embodiment, in order to facilitate rapid identification of the image to be labeled, the sub-brand image sets corresponding to the main brand may be classified according to different vehicle parts, such as the sub-brand image set of the car lights, the sub-brand image set of the wheel hubs, and the sub-brand image set of the grille, so that the image to be labeled can be matched with the corresponding sub-brand image set through the included key region.
And S208, performing sub-brand identification on the image to be annotated based on the sub-brand image set to obtain a sub-brand of the vehicle.
Often the distinction between sub-brands is manifested by the characteristics of critical areas of the vehicle; for example, a key area of the inlet grill, audi A6L in 2019, relative to audi A6L in 2018, has the following characteristics: the height is unchanged and the length is increased. In this case, the sub-brand of the vehicle may be determined by comparing the image to be annotated and the set of sub-brand images.
And step S210, marking the sub-brand of the vehicle in the image to be marked.
In this embodiment, the sub-brands of the vehicle may be labeled in a variety of ways. For example, the determined sub-brand of the vehicle and the image to be annotated can be associated; or, modifying the original name field of the image to be marked into a name field containing a sub-brand of the vehicle; or adding a hyperlink carrying a sub-brand of the vehicle at the position of the vehicle of the image to be marked. Of course, the above three ways are merely exemplary illustrations of labeling sub-brands of vehicles and should not be construed as limiting.
The embodiment of the invention provides a vehicle brand labeling method, which comprises the steps of firstly inputting an acquired image to be labeled into a pre-trained neural network model to identify and label a vehicle main brand; then acquiring a sub-brand image set corresponding to the identified main brand of the vehicle, and performing sub-brand identification on the image to be marked based on the sub-brand image set to obtain a sub-brand of the vehicle; and finally, marking the sub-brand of the vehicle in the image to be marked. Compared with the existing manual labeling mode, the mode provided by the embodiment can effectively reduce the manpower requirement and improve the labeling efficiency of the main brand by identifying the main brand of the vehicle of the image to be labeled by utilizing the neural network model; the sub-brand image set can better meet the requirement of fast recognition of vehicle brands under fine granularity; the identification process from the main brand to the sub-brand is integrated, and the labeling efficiency of the vehicle brand is greatly improved.
For easy understanding, the present embodiment provides a manner of acquiring an image to be annotated in relation to the above step S202, and refers to the following steps (1) to (4):
(1) an original image containing a vehicle is acquired. The original image may be an image captured by an image capturing device such as a camera of the electronic device, or an image downloaded by the electronic device from a server in a network communication manner, or an image directly retrieved by the electronic device from a local storage area.
(2) Detecting whether a preset key area exists in an original image; the key zones include at least one of a vehicle light zone, a vehicle logo zone, a wheel hub zone, and a barrier zone.
In a possible implementation manner, a Region Of Interest (ROI) in an original image may be extracted by using a selective search method, and then whether a preset key Region exists in the extracted ROI may be determined. The Selective Search method integrates a brute force Search and a segmentation method, can provide various strategies, and can greatly reduce the Search space. Of course, other ways such as manual division or ROI extraction method can be selected to extract multiple regions of interest from the original image.
(3) If present, the key regions are quality scored according to sharpness and/or distortion factors.
In order to avoid the problem that the image cannot be brand-marked due to the fact that the key area is shielded, cut off or distorted, the quality of the key area can be scored according to the definition factor and/or the distortion factor so as to screen the original image containing the clear and complete key area. It can be understood that, besides the definition factor and the distortion factor, the noise factor, the brightness factor, the color factor, the focusing factor and the like can be further referred to screen the image to be annotated with better quality from the original image, so that the license plate annotation can be realized quickly and accurately.
(4) And when the quality score is higher than a preset score threshold value, determining the original image as the image to be annotated.
And if no key area with the quality score higher than a preset score threshold value exists in the original image, the label which is not applicable to the brand of the original image is represented, and the original image is discarded. If the quality score of a certain key area is higher than a preset score threshold value, the key area representing the original image can be applied to brand labeling, and therefore the original image is determined to be an image to be labeled.
Considering that the main brand of the vehicle determined based on the neural network model may have errors, the embodiment provides a main brand labeling method based on the neural network model and human-computer interaction, and refers to the following steps 1 to 5:
step 1, inputting an image to be marked into a pre-trained neural network model to identify a vehicle main brand, so as to obtain an initial main brand. And extracting the characteristics of each key area in the image to be marked through a neural network model, classifying the extracted characteristics, and obtaining an initial main brand of the vehicle according to a characteristic classification result.
And 2, acquiring the discrimination operation aiming at the initial main brand. In consideration of the fact that in actual life, some vehicle owners refit the vehicle, and the identified initial main brand is not accurate, the embodiment can ensure the accuracy of the main brand of the vehicle by combining with manual mode. In a specific implementation, the distinguishing operation is generally an operation of the user for the initial main brand feedback, and includes an operation of determining that the initial main brand is correct and an operation of determining that the initial main brand is wrong. If the judging operation is the operation for determining that the initial main brand is correct, executing the following step 3; if the discrimination operation is an operation for determining that the initial main brand is wrong, the following step 4 is performed.
And 3, marking the main brand of the vehicle in the image to be marked according to the initial main brand.
And 4, responding to the correction operation aiming at the initial main brand, and acquiring a corrected main brand corresponding to the correction operation. When the determined initial main brand is not correct, the user can execute a correction operation through the editing interface, and the correction operation is the operation of inputting the correct main brand, so that the corrected main brand corresponding to the correction operation is obtained.
And 5, marking the main brand of the vehicle in the image to be marked according to the corrected main brand.
In order to enable the neural network model to be directly applied to identification of a main brand, the neural network model needs to be trained in advance, parameters of the neural network model need to be obtained through training, and the purpose of training the neural network model is to finally determine the parameters which can meet requirements. With the trained parameters, the neural network model can obtain the expected main brand identification. The embodiment provides a training method of a neural network model, which comprises the following four steps:
step one, inputting a plurality of vehicle training images with marked vehicle main brands into a currently trained neural network model; and setting the marked vehicle main brand of the vehicle training image as a reference main brand. In practical applications, the vehicle main brand can be characterized in the form of feature vectors so as to facilitate the neural network model to process the vehicle main brand.
And step two, identifying the vehicle training image through the currently trained neural network model to obtain a predicted main brand of the vehicle training image.
Calculating a loss function value between a prediction main brand and a reference main brand corresponding to the vehicle training image; wherein the loss function value is used to evaluate the accuracy of the recognition of the main brand of the vehicle training image.
And step four, adjusting the parameters of the currently trained neural network model through a back propagation algorithm according to the loss function value until the loss function value converges to a preset value, and finishing the training.
Based on the trained neural network model, the main brand of the vehicle can be determined according to the image to be labeled, so that sub-brand labeling is further performed based on the main brand, and the labeling efficiency of the license plate brand is effectively improved.
It is understood that there are multiple sub-brands under each main brand; in this case, the sub-brand image sets are typically multiple, and different sub-brand image sets correspond to different vehicle parts, such as the sub-brand image set corresponding to the vehicle lights shown in fig. 3, which shows the vehicle light images of different sub-brands under the audi flag, and each vehicle light image is labeled with a specific sub-brand, for example, a3 Sportback 17/18.
Based on the plurality of sub-brand image sets, the implementation manner of sub-brand identification in the image to be annotated may include the following two examples.
For example one, the following steps a to D may be referred to:
a, acquiring quality scores of all key areas in an image to be labeled; this embodiment provides an implementation manner for acquiring the quality score of each region of interest: and for each region of interest divided from the image to be marked, performing quality detection on the region of interest according to factors such as definition factors, distortion factors and the like of image quality, and obtaining a quality score of the region of interest.
And step B, determining a first vehicle part contained in the key area with the highest quality score. Each key area obtains a corresponding quality score, and the vehicle part contained in the key area with the highest quality score is used as a first vehicle part (such as a vehicle lamp part).
And C, selecting a first target sub-brand image set from the plurality of sub-brand image sets according to the determined first vehicle part. And if the determined first vehicle part is the vehicle headlight part, selecting a sub-brand image set of the vehicle headlight from the plurality of sub-brand image sets, and taking the sub-brand image set of the vehicle headlight as a first target sub-brand image set.
And D, determining the sub-brand of the vehicle by comparing the image to be labeled with the first target sub-brand image set. The first target sub-brand image set comprises images marked with different sub-brands, and in specific implementation, the similarity between the image to be marked and each image in the first target sub-brand image set can be calculated respectively, the image corresponding to the maximum similarity is obtained, and the sub-brand marked by the image is determined as the sub-brand of the vehicle.
Example two: reference may be made to steps 1) to 6) below:
1) and determining a second vehicle part contained in each key area in the image to be marked.
2) And selecting a second target sub-brand image set corresponding to each key area from the plurality of sub-brand image sets according to the determined second vehicle part. It can be understood that the determined second target sub-brand image set corresponds to and is equal in number to the key areas contained in the image to be annotated. And assuming that key areas in the image to be marked are a vehicle sign area, a vehicle lamp area and a barrier area, determining a second vehicle part as a sub-brand image set of the vehicle sign, a sub-brand image set of the vehicle lamp and a sub-brand image set of the barrier.
3) And aiming at each key area in the image to be labeled, calculating a first matching degree of the key area and a second target sub-brand image set corresponding to the key area, and determining an initial sub-brand of the key area according to the first matching degree. And the second target sub-brand image set stores reference images for marking the sub-brands of the vehicles.
In a specific implementation manner, firstly, a preset key point detection model is used for detecting key points of the key area to obtain first key point information of the key area; then acquiring second key point information preset by each reference image in a second target sub-brand image set corresponding to the key area; then, comparing the first key point information with the second key point information of each reference image to obtain a second matching degree of the key area and each reference image; then determining the maximum matching degree in the second matching degrees as the first matching degree of the key area and a second target sub-brand image set corresponding to the key area; and finally, acquiring the sub-brand of the vehicle marked by the reference image corresponding to the first matching degree, and determining the acquired sub-brand of the vehicle as the initial sub-brand of the key area.
4) And judging whether the initial sub-brands corresponding to different key areas are the same or not. If the same, the following step 5) is executed; if not, the following step 6) is performed.
5) Determining the initial sub-brand as a sub-brand of the vehicle;
6) and determining the initial sub-brand corresponding to the maximum matching degree in the first matching degrees as the sub-brand of the vehicle.
In summary, in the manner provided by this embodiment, after the image to be labeled which can be used for labeling is obtained, the neural network model is directly utilized to identify the main brand of the vehicle of the image to be labeled, so that the manpower requirement can be effectively reduced and the labeling efficiency of the main brand can be improved; the sub-brand image set can better meet the requirement of fast recognition of vehicle brands under fine granularity; the identification process from the main brand to the sub-brand is integrated, so that the labeling efficiency of the vehicle brand is greatly improved; in addition, the method is not limited by the professional level of the user, and the expansibility and the effectiveness of vehicle labeling are greatly improved.
Example three:
based on the method for labeling the brand of the vehicle provided by the above embodiment, the embodiment provides a device for labeling the brand of the vehicle. Referring to fig. 4, a block diagram of a vehicle brand labeling apparatus is shown, which includes:
an image obtaining module 402, configured to obtain an image to be annotated, which includes a key area of a vehicle;
a main brand labeling module 404, configured to input an image to be labeled to a pre-trained neural network model to perform identification and labeling of a vehicle main brand; the neural network model is obtained by training a vehicle training image based on a marked vehicle main brand;
a sub-brand image acquisition module 406, configured to acquire a sub-brand image set corresponding to the identified main brand of the vehicle;
the sub-brand identification module 408 is configured to perform sub-brand identification on the image to be annotated based on the sub-brand image set to obtain a sub-brand of the vehicle;
and a sub-brand labeling module 410, configured to label a sub-brand of the vehicle in the image to be labeled.
According to the marking device for the vehicle brand, provided by the embodiment of the invention, firstly, the acquired image to be marked is input to a pre-trained neural network model to identify and mark the main brand of the vehicle; then acquiring a sub-brand image set corresponding to the identified main brand of the vehicle, and performing sub-brand identification on the image to be marked based on the sub-brand image set to obtain a sub-brand of the vehicle; and finally, marking the sub-brand of the vehicle in the image to be marked. Compared with the existing manual labeling mode, the mode provided by the embodiment can effectively reduce the manpower requirement and improve the labeling efficiency of the main brand by identifying the main brand of the vehicle of the image to be labeled by utilizing the neural network model; the sub-brand image set can better meet the requirement of fast recognition of vehicle brands under fine granularity; the identification process from the main brand to the sub-brand is integrated, and the labeling efficiency of the vehicle brand is greatly improved.
In some embodiments, the sub-brand image acquisition module 406 is further configured to: acquiring an original image containing a vehicle; detecting whether a preset key area exists in an original image; the key area comprises at least one of a vehicle lamp area, a vehicle sign area, a hub area and a barrier area; if yes, performing quality scoring on the key area according to the definition factor and/or the distortion factor; and when the quality score is higher than a preset score threshold value, determining the original image as the image to be annotated.
In some embodiments, the sub-brand image set is a plurality of sub-brand image sets, and different sub-brand image sets correspond to different vehicle parts; the sub-brand identification module 408 is further configured to: acquiring quality scores of all key areas in an image to be marked; determining a first vehicle part contained in a key area with the highest quality score; selecting a first target sub-brand image set from the plurality of sub-brand image sets according to the determined first vehicle part; and determining the sub-brand of the vehicle by comparing the image to be marked with the first target sub-brand image set.
In some embodiments, the sub-brand image set is a plurality of sub-brand image sets, and different sub-brand image sets correspond to different vehicle parts; the sub-brand identification module 408 is further configured to: determining a second vehicle part contained in each key area in the image to be marked; selecting a second target sub-brand image set corresponding to each key area from the plurality of sub-brand image sets according to the determined second vehicle part; aiming at each key area in the image to be marked, calculating a first matching degree of the key area and a second target sub-brand image set corresponding to the key area, and determining an initial sub-brand of the key area according to the first matching degree; judging whether the initial sub-brands corresponding to different key areas are the same or not; if so, determining the initial sub-brand as a sub-brand of the vehicle; and if not, determining the initial sub-brand corresponding to the maximum matching degree in the first matching degrees as the sub-brand of the vehicle.
In some embodiments, the second target sub-brand image set stores reference images of sub-brands of tagged vehicles; the sub-brand identification module 408 is further configured to: performing key point detection on the key area through a preset key point detection model to obtain first key point information of the key area; acquiring second key point information preset by each reference image in a second target sub-brand image set corresponding to the key area; obtaining a second matching degree of the key area and each reference image by comparing the first key point information with the second key point information of each reference image; determining the maximum matching degree in the second matching degrees as the first matching degree of the key area and a second target sub-brand image set corresponding to the key area; and acquiring the sub-brand of the vehicle marked by the reference image corresponding to the first matching degree, and determining the acquired sub-brand of the vehicle as the initial sub-brand of the key area.
In some embodiments, the primary brand labeling module 404 is further configured to: inputting the image to be marked into a pre-trained neural network model to identify a vehicle main brand to obtain an initial main brand; acquiring a distinguishing operation aiming at an initial main brand; if the judging operation is the operation for determining that the initial main brand is correct, marking the main brand of the vehicle in the image to be marked according to the initial main brand; if the judging operation is an operation for determining that the initial main brand is wrong, responding to the correction operation aiming at the initial main brand, and acquiring a corrected main brand corresponding to the correction operation; and marking the main brand of the vehicle in the image to be marked according to the corrected main brand.
In some embodiments, the vehicle brand labeling apparatus further includes a model training module; the model training module is configured to: inputting a plurality of vehicle training images with marked vehicle main brands to a currently trained neural network model; setting a marked vehicle main brand of the vehicle training image as a reference main brand; identifying the vehicle training image through the currently trained neural network model to obtain a predicted main brand of the vehicle training image; calculating a loss function value between a prediction main brand and a reference main brand corresponding to the vehicle training image; the loss function value is used for evaluating the recognition accuracy of the main brand of the vehicle training image; and adjusting the parameters of the currently trained neural network model through a back propagation algorithm according to the loss function value until the loss function value converges to a preset value, and finishing the training.
The device provided in this embodiment has the same implementation principle and technical effects as those of the foregoing embodiment, and for the sake of brief description, reference may be made to corresponding contents in the foregoing embodiment.
Example four:
based on the foregoing embodiments, the present embodiment provides a vehicle brand labeling system, including: the system comprises an image acquisition device, a processor and a storage device; the image acquisition equipment is used for acquiring an image to be annotated; the storage device has a computer program stored thereon, which when executed by the processor performs the method of labeling a brand of any of the vehicles as provided in embodiment two.
It can be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working process of the system described above may refer to the corresponding process in the foregoing method embodiment, and is not described herein again.
Further, the present embodiment also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processing device, the steps of the method for labeling a brand of a vehicle provided in any one of the second embodiment are performed.
The computer program product of the method, the apparatus, and the system for labeling a vehicle brand provided in the embodiments of the present invention includes a computer-readable storage medium storing program codes, where instructions included in the program codes may be used to execute the method described in the foregoing method embodiments, and specific implementations of the method embodiments may be referred to, and are not described herein again.
The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
Finally, it should be noted that: the above-mentioned embodiments are only specific embodiments of the present invention, which are used for illustrating the technical solutions of the present invention and not for limiting the same, and the protection scope of the present invention is not limited thereto, although the present invention is described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: any person skilled in the art can modify or easily conceive the technical solutions described in the foregoing embodiments or equivalent substitutes for some technical features within the technical scope of the present disclosure; such modifications, changes or substitutions do not depart from the spirit and scope of the embodiments of the present invention, and they should be construed as being included therein. Therefore, the protection scope of the present invention shall be subject to the protection scope of the appended claims.
Claims (10)
1. A method of labeling a brand of a vehicle, the method comprising:
acquiring an image to be marked of a key area containing a vehicle;
inputting the image to be marked into a pre-trained neural network model to identify and mark a vehicle main brand; the neural network model is obtained by training a vehicle training image based on a marked vehicle main brand;
acquiring a sub-brand image set corresponding to the identified main brand of the vehicle;
performing sub-brand identification on the image to be marked based on the sub-brand image set to obtain a sub-brand of the vehicle;
and marking the sub-brand of the vehicle in the image to be marked.
2. The method according to claim 1, wherein the step of obtaining the image to be labeled containing the key area of the vehicle comprises:
acquiring an original image containing a vehicle;
detecting whether a preset key area exists in the original image; the key area comprises at least one of a car lamp area, a car logo area, a hub area and a barrier area;
if yes, scoring the quality of the key area according to the definition factor and/or the distortion factor;
and when the quality score is higher than a preset score threshold value, determining the original image as an image to be annotated.
3. The method of claim 1, wherein the sub-brand image sets are multiple and different ones of the sub-brand image sets correspond to different vehicle parts;
the step of identifying the sub-brand in the image to be marked based on the sub-brand image set to obtain the sub-brand of the vehicle comprises the following steps:
acquiring the quality score of each key area in the image to be annotated;
determining a first vehicle part contained in the key area with the highest quality score;
selecting a first target sub-brand image set from the plurality of sub-brand image sets according to the determined first vehicle location;
and determining the sub-brand of the vehicle by comparing the image to be labeled with the first target sub-brand image set.
4. The method of claim 1, wherein the sub-brand image sets are multiple and different ones of the sub-brand image sets correspond to different vehicle parts;
the step of identifying the sub-brand in the image to be marked based on the sub-brand image set to obtain the sub-brand of the vehicle comprises the following steps:
determining a second vehicle part contained in each key area in the image to be marked;
selecting a second target sub-brand image set corresponding to each key area from the plurality of sub-brand image sets according to the determined second vehicle part;
calculating a first matching degree of each key area in the image to be labeled and a second target sub-brand image set corresponding to the key area, and determining an initial sub-brand of the key area according to the first matching degree;
judging whether the initial sub-brands corresponding to different key areas are the same or not;
if so, determining the initial sub-brand as a sub-brand of the vehicle;
and if not, determining the initial sub-brand corresponding to the maximum matching degree in the first matching degrees as the sub-brand of the vehicle.
5. The method of claim 4, wherein the second set of target sub-brand images stores reference images that label sub-brands of vehicles;
the step of calculating a first matching degree of the key area and a second target sub-brand image set corresponding to the key area, and determining an initial sub-brand of the key area according to the first matching degree comprises the following steps:
performing key point detection on the key area through a preset key point detection model to obtain first key point information of the key area;
acquiring second key point information preset by each reference image in a second target sub-brand image set corresponding to the key area;
obtaining a second matching degree of the key area and each reference image by comparing the first key point information with the second key point information of each reference image;
determining the maximum matching degree in the second matching degrees as the first matching degree of the key area and a second target sub-brand image set corresponding to the key area;
and acquiring the sub-brand of the vehicle marked by the reference image corresponding to the first matching degree, and determining the acquired sub-brand of the vehicle as the initial sub-brand of the key area.
6. The method according to claim 1, wherein the step of inputting the image to be labeled into a pre-trained neural network model for identifying and labeling the main brand of the vehicle comprises:
inputting the image to be marked into a pre-trained neural network model to identify a vehicle main brand to obtain an initial main brand;
acquiring a discrimination operation for the initial main brand;
if the judging operation is the operation for determining that the initial main brand is correct, marking the main brand of the vehicle in the image to be marked according to the initial main brand;
if the judging operation is an operation for determining that the initial main brand is wrong, responding to a correction operation aiming at the initial main brand, and acquiring a corrected main brand corresponding to the correction operation;
and marking the main brand of the vehicle in the image to be marked according to the corrected main brand.
7. The method of claim 1, wherein the training process of the neural network model comprises:
inputting a plurality of vehicle training images with marked vehicle main brands to a currently trained neural network model; setting a marked vehicle main brand of the vehicle training image as a reference main brand;
identifying the vehicle training image through a currently trained neural network model to obtain a predicted main brand of the vehicle training image;
calculating a loss function value between a prediction main brand and a reference main brand corresponding to the vehicle training image; wherein the loss function value is used to evaluate a primary brand recognition accuracy of the vehicle training image;
and adjusting the parameters of the currently trained neural network model through a back propagation algorithm according to the loss function value until the loss function value converges to a preset value, and finishing the training.
8. A vehicle brand labeling apparatus, said apparatus comprising:
the image acquisition module is used for acquiring an image to be marked of a key area containing a vehicle;
the main brand labeling module is used for inputting the image to be labeled to a pre-trained neural network model to identify and label a vehicle main brand; the neural network model is obtained by training a vehicle training image based on a marked vehicle main brand;
the sub-brand image acquisition module is used for acquiring a sub-brand image set corresponding to the identified main brand of the vehicle;
the sub-brand identification module is used for carrying out sub-brand identification on the image to be marked based on the sub-brand image set to obtain a sub-brand of the vehicle;
and the sub-brand marking module is used for marking the sub-brand of the vehicle in the image to be marked.
9. A vehicle brand labeling system, the system comprising: the device comprises an image acquisition device, a processor and a storage device;
the image acquisition device is used for acquiring an image to be marked;
the storage device has stored thereon a computer program which, when executed by the processor, performs the method of any one of claims 1 to 7.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method according to any one of the claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911179667.1A CN111222409A (en) | 2019-11-26 | 2019-11-26 | Vehicle brand labeling method, device and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911179667.1A CN111222409A (en) | 2019-11-26 | 2019-11-26 | Vehicle brand labeling method, device and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111222409A true CN111222409A (en) | 2020-06-02 |
Family
ID=70829050
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911179667.1A Pending CN111222409A (en) | 2019-11-26 | 2019-11-26 | Vehicle brand labeling method, device and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111222409A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111754784A (en) * | 2020-06-23 | 2020-10-09 | 高新兴科技集团股份有限公司 | Attention mechanism-based vehicle main and sub brand identification method of multilayer network |
| CN112364898A (en) * | 2020-10-27 | 2021-02-12 | 星火科技技术(深圳)有限责任公司 | Image identification automatic labeling method, device, equipment and storage medium |
| CN112525556A (en) * | 2020-12-24 | 2021-03-19 | 湖南联科科技有限公司 | Diagnostic method, apparatus, device and storage medium for automatically identifying vehicle brand |
| CN113065609A (en) * | 2021-04-22 | 2021-07-02 | 平安国际智慧城市科技股份有限公司 | Image classification method and device, electronic equipment and readable storage medium |
| WO2022196592A1 (en) * | 2021-03-15 | 2022-09-22 | 国立大学法人 東京大学 | Information processing device and program |
Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104361343A (en) * | 2014-10-30 | 2015-02-18 | 深圳市中盟科技股份有限公司 | Method and device for identifying vehicle types |
| CN105448103A (en) * | 2015-12-24 | 2016-03-30 | 北京旷视科技有限公司 | Vehicle fake license plate detection method and system |
| CN106295541A (en) * | 2016-08-03 | 2017-01-04 | 乐视控股（北京）有限公司 | Vehicle type recognition method and system |
| CN106650740A (en) * | 2016-12-15 | 2017-05-10 | 深圳市华尊科技股份有限公司 | License plate identification method and terminal |
| CN107301417A (en) * | 2017-06-28 | 2017-10-27 | 广东工业大学 | A kind of method and device of the vehicle brand identification of unsupervised multilayer neural network |
| CN107729818A (en) * | 2017-09-21 | 2018-02-23 | 北京航空航天大学 | A kind of multiple features fusion vehicle recognition methods again based on deep learning |
| CN108549926A (en) * | 2018-03-09 | 2018-09-18 | 中山大学 | A kind of deep neural network and training method for refining identification vehicle attribute |
| CN108875458A (en) * | 2017-05-15 | 2018-11-23 | 杭州海康威视数字技术股份有限公司 | Detection method, device, electronic equipment and the video camera that vehicular high beam lamp is opened |
| CN108984644A (en) * | 2018-06-25 | 2018-12-11 | 高新兴科技集团股份有限公司 | A kind of fake-licensed car search method and system |
| CN109033175A (en) * | 2018-06-25 | 2018-12-18 | 高新兴科技集团股份有限公司 | A kind of method and system to scheme to search vehicle |
| US20190026603A1 (en) * | 2017-07-19 | 2019-01-24 | Beijing ICETech Science & Technology Co., Ltd. | Method and system for vehicle recognition |
| CN109559355A (en) * | 2018-12-04 | 2019-04-02 | 北京航空航天大学 | A kind of polyphaser global calibration device and method without public view field based on phase unit |
| US20190114494A1 (en) * | 2016-03-29 | 2019-04-18 | Panasonic Intellectual Property Management Co., Ltd. | Vehicle model identification device, vehicle model identification system, and vehicle model identification method |
| CN109657587A (en) * | 2018-12-10 | 2019-04-19 | 南京甄视智能科技有限公司 | Side face method for evaluating quality and system for recognition of face |
| CN109948610A (en) * | 2019-03-14 | 2019-06-28 | 西南交通大学 | A kind of vehicle fine grit classification method in the video based on deep learning |
| CN110135318A (en) * | 2019-05-08 | 2019-08-16 | 佳都新太科技股份有限公司 | Cross determination method, apparatus, equipment and the storage medium of vehicle record |
-
2019
- 2019-11-26 CN CN201911179667.1A patent/CN111222409A/en active Pending
Patent Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104361343A (en) * | 2014-10-30 | 2015-02-18 | 深圳市中盟科技股份有限公司 | Method and device for identifying vehicle types |
| CN105448103A (en) * | 2015-12-24 | 2016-03-30 | 北京旷视科技有限公司 | Vehicle fake license plate detection method and system |
| US20190114494A1 (en) * | 2016-03-29 | 2019-04-18 | Panasonic Intellectual Property Management Co., Ltd. | Vehicle model identification device, vehicle model identification system, and vehicle model identification method |
| CN106295541A (en) * | 2016-08-03 | 2017-01-04 | 乐视控股（北京）有限公司 | Vehicle type recognition method and system |
| CN106650740A (en) * | 2016-12-15 | 2017-05-10 | 深圳市华尊科技股份有限公司 | License plate identification method and terminal |
| CN108875458A (en) * | 2017-05-15 | 2018-11-23 | 杭州海康威视数字技术股份有限公司 | Detection method, device, electronic equipment and the video camera that vehicular high beam lamp is opened |
| CN107301417A (en) * | 2017-06-28 | 2017-10-27 | 广东工业大学 | A kind of method and device of the vehicle brand identification of unsupervised multilayer neural network |
| US20190026603A1 (en) * | 2017-07-19 | 2019-01-24 | Beijing ICETech Science & Technology Co., Ltd. | Method and system for vehicle recognition |
| CN107729818A (en) * | 2017-09-21 | 2018-02-23 | 北京航空航天大学 | A kind of multiple features fusion vehicle recognition methods again based on deep learning |
| CN108549926A (en) * | 2018-03-09 | 2018-09-18 | 中山大学 | A kind of deep neural network and training method for refining identification vehicle attribute |
| CN108984644A (en) * | 2018-06-25 | 2018-12-11 | 高新兴科技集团股份有限公司 | A kind of fake-licensed car search method and system |
| CN109033175A (en) * | 2018-06-25 | 2018-12-18 | 高新兴科技集团股份有限公司 | A kind of method and system to scheme to search vehicle |
| CN109559355A (en) * | 2018-12-04 | 2019-04-02 | 北京航空航天大学 | A kind of polyphaser global calibration device and method without public view field based on phase unit |
| CN109657587A (en) * | 2018-12-10 | 2019-04-19 | 南京甄视智能科技有限公司 | Side face method for evaluating quality and system for recognition of face |
| CN109948610A (en) * | 2019-03-14 | 2019-06-28 | 西南交通大学 | A kind of vehicle fine grit classification method in the video based on deep learning |
| CN110135318A (en) * | 2019-05-08 | 2019-08-16 | 佳都新太科技股份有限公司 | Cross determination method, apparatus, equipment and the storage medium of vehicle record |
Non-Patent Citations (2)
| Title |
|---|
| 余烨;金强;傅云翔;路强;: "基于Fg-CarNet的车辆型号精细分类研究", 自动化学报, no. 10, pages 138 - 149 * |
| 王亚;朱明;刘成林;: "基于CNN的监控视频中人脸图像质量评估", 计算机系统应用, no. 11, 14 November 2018 (2018-11-14) * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111754784A (en) * | 2020-06-23 | 2020-10-09 | 高新兴科技集团股份有限公司 | Attention mechanism-based vehicle main and sub brand identification method of multilayer network |
| CN112364898A (en) * | 2020-10-27 | 2021-02-12 | 星火科技技术(深圳)有限责任公司 | Image identification automatic labeling method, device, equipment and storage medium |
| CN112364898B (en) * | 2020-10-27 | 2024-01-19 | 星火科技技术(深圳)有限责任公司 | Automatic labeling method, device, equipment and storage medium for image recognition |
| CN112525556A (en) * | 2020-12-24 | 2021-03-19 | 湖南联科科技有限公司 | Diagnostic method, apparatus, device and storage medium for automatically identifying vehicle brand |
| WO2022196592A1 (en) * | 2021-03-15 | 2022-09-22 | 国立大学法人 東京大学 | Information processing device and program |
| CN113065609A (en) * | 2021-04-22 | 2021-07-02 | 平安国际智慧城市科技股份有限公司 | Image classification method and device, electronic equipment and readable storage medium |
| CN113065609B (en) * | 2021-04-22 | 2024-04-09 | 深圳赛安特技术服务有限公司 | Image classification method, device, electronic equipment and readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111222409A (en) | Vehicle brand labeling method, device and system | |
| US10853638B2 (en) | System and method for extracting structured information from image documents | |
| CN107403424B (en) | Vehicle loss assessment method and device based on image and electronic equipment | |
| JP6994101B2 (en) | Target recognition methods, devices, storage media and electronic devices | |
| US9014432B2 (en) | License plate character segmentation using likelihood maximization | |
| US9171204B2 (en) | Method of perspective correction for devanagari text | |
| US7970213B1 (en) | Method and system for improving the recognition of text in an image | |
| US8358837B2 (en) | Apparatus and methods for detecting adult videos | |
| JP6163344B2 (en) | Reliable cropping of license plate images | |
| CN106547744B (en) | Image retrieval method and system | |
| US8965127B2 (en) | Method for segmenting text words in document images | |
| US20230376527A1 (en) | Generating congruous metadata for multimedia | |
| CN110705405A (en) | Target labeling method and device | |
| KR102087882B1 (en) | Device and method for media stream recognition based on visual image matching | |
| US20120224765A1 (en) | Text region detection system and method | |
| US9613299B2 (en) | Method of identifying pattern training need during verification of recognized text | |
| US20160253577A1 (en) | Image Clustering Method, Image Clustering System, And Image Clustering Server | |
| US20110156999A1 (en) | Gesture recognition methods and systems | |
| CN110291527B (en) | Information processing method, system, cloud processing device and computer program product | |
| CN112733666A (en) | Method, equipment and storage medium for collecting difficult images and training models | |
| CN108665459B (en) | Image blur detection method, computing device and readable storage medium | |
| JP2014115961A (en) | Information processor, vehicle identification method and vehicle identification program | |
| CN112241470A (en) | Video classification method and system | |
| CN111738173A (en) | Video clip detection method and device, electronic equipment and storage medium | |
| CN116665272A (en) | Airport scene face recognition fusion decision method and device, electronic equipment and medium |
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 |