CN109325945B - Image processing method, image processing device, electronic equipment and storage medium - Google Patents
Image processing method, image processing device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN109325945B CN109325945B CN201811071355.4A CN201811071355A CN109325945B CN 109325945 B CN109325945 B CN 109325945B CN 201811071355 A CN201811071355 A CN 201811071355A CN 109325945 B CN109325945 B CN 109325945B
- Authority
- CN
- China
- Prior art keywords
- convolution kernel
- abscissa
- ordinate
- sampling
- point
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Quality & Reliability (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Processing (AREA)
Abstract
The embodiment of the application provides an image processing method and device, electronic equipment and a storage medium, and relates to the technical field of image processing. The method comprises the following steps: obtaining shape parameters related to the shape of the image sampling frame; adjusting each sampling point of at least two sampling points of the convolution kernel from a current first position to a second position corresponding to the shape parameter according to the shape parameter and the convolution kernel parameter of the convolution kernel to obtain an adjusted convolution kernel, wherein the current positions of any two sampling points of the at least two sampling points are different, and the adjusted positions of any two sampling points are also different; and carrying out convolution processing on the image in the image sampling frame according to the adjusted convolution kernel. The convolution processing is carried out on the image in the image sampling frame through the convolution kernel of which the sampling point position distribution is matched with the shape of the image sampling frame, so that the deviation of the detection result of the image can be avoided, and the detection precision is improved.
Description
Technical Field
The present disclosure relates to the field of image processing technologies, and in particular, to an image processing method and apparatus, an electronic device, and a storage medium.
Background
Currently, in the target detection technology, an image sampling frame for selecting an object in an image may be generated to further detect and identify a target object image in the image sampling frame.
However, since the randomness of the size and the position of the target image in the image is relatively large, the sizes and the shapes of the generated image sampling frames are also greatly different. Therefore, the shape and size of the image sampling frame are not fixed, so that the detection result of the target object is deviated in the process of detecting the target object image in the image sampling frame, and the detection precision is influenced.
Disclosure of Invention
The present application provides an image processing method, an image processing apparatus, an electronic device, and a storage medium, so as to effectively prevent a detection result of a target object from being biased, thereby improving detection accuracy.
In order to achieve the above object, embodiments of the present application are implemented as follows:
in a first aspect, an embodiment of the present application provides an image processing method, where the method includes:
obtaining shape parameters related to the shape of the image sampling frame;
adjusting each sampling point of at least two sampling points of the convolution kernel from a current first position to a second position corresponding to the shape parameter according to the shape parameter and the convolution kernel parameter of the convolution kernel to obtain an adjusted convolution kernel, wherein the current positions of any two sampling points of the at least two sampling points are different, and the adjusted positions of any two sampling points are different;
and performing convolution processing on the image according to the adjusted convolution kernel.
With reference to the first aspect, an embodiment of the present application provides a first possible implementation manner of the first aspect, where the shape parameter includes: the length and width of the image sampling frame, and the reduction multiple of the image sampling frame in the down-sampling; adjusting each of at least two sampling points of the convolution kernel from a current first position to a second position corresponding to the shape parameter according to the shape parameter and a convolution kernel parameter of the convolution kernel to obtain an adjusted convolution kernel, including:
according to the length, the reduction multiple and the convolution kernel parameter, adjusting the abscissa of each of at least two sampling points of the convolution kernel from the current first abscissa to a second abscissa corresponding to the length;
according to the width, the reduction multiple and the convolution kernel parameter, adjusting the ordinate of each sampling point from the current first ordinate to the second ordinate corresponding to the width, wherein the coordinate of each sampling point before adjustment is the first abscissa and the first ordinate, which indicates that the position of each sampling point is at the current first position, and the coordinate of each sampling point after adjustment is the second abscissa and the second ordinate, which indicates that the position of each sampling point is at the second position after adjustment;
and obtaining an adjusted convolution kernel, wherein the position of each sampling point in the adjusted convolution kernel is located at the second position.
With reference to the first aspect, the present embodiment provides a second possible implementation manner of the first aspect, where the adjusting, according to the length, the reduction multiple, and the convolution kernel parameter, an abscissa of each of at least two sampling points of the convolution kernel from a current first abscissa to a second abscissa corresponding to the length includes:
determining a first deviation value of the abscissa of each of at least two sampling points of the convolution kernel according to the length, the reduction multiple and the convolution kernel parameter;
and adjusting the abscissa of each sampling point from the current first abscissa to a second abscissa corresponding to the length according to the first deviation value of the abscissa of each sampling point.
With reference to the first aspect, an embodiment of the present application provides a third possible implementation manner of the first aspect, where the convolution kernel parameter includes: the total column number of the convolution kernel and the interval column number between the column of each sampling point and the central point of the convolution kernel; the determining a first deviation value of the abscissa of each sampling point in the convolution kernel according to the length, the reduction multiple and the convolution kernel parameter includes:
and determining that the first deviation value of the abscissa of each sampling point in the convolution kernel is W × M/M × S according to the length W, the reduction multiple S, the total column number M and the interval column number M.
With reference to the first aspect, an embodiment of the present application provides a fourth possible implementation manner of the first aspect, where, according to the width, the reduction multiple, and the convolution kernel parameter, the adjusting a vertical coordinate of each sample point from a current first vertical coordinate to a second vertical coordinate corresponding to the width includes:
determining a second deviation value of the ordinate of each sampling point according to the width, the reduction multiple and the convolution kernel parameter;
and adjusting the ordinate of each sampling point from the current first ordinate to a second ordinate corresponding to the width according to the second deviation value of the ordinate of each sampling point.
With reference to the first aspect, an embodiment of the present application provides a fifth possible implementation manner of the first aspect, where the convolution kernel parameter further includes: the total number of rows of the convolution kernel and the number of spaced rows between the row of each sampling point and the central point of the convolution kernel; determining a second deviation value of the ordinate of each sampling point according to the width, the reduction multiple and the convolution kernel parameter, wherein the second deviation value comprises;
and determining a second deviation value H x N/N x S of the vertical coordinate of each sampling point according to the width H, the reduction multiple S, the total line number N and the interval line number N.
With reference to the first aspect and any one of the first to fifth possible implementation manners of the first aspect, an embodiment of the present application provides a sixth possible implementation manner of the first aspect, and before the obtaining the shape of the image sample frame, the method further includes:
obtaining a first position parameter of a central point of the image sampling frame;
and adjusting each sampling point from the current third position to the first position according to the first position parameter and a second position parameter of the central point of the convolution kernel, wherein the position of each sampling point at the first position indicates that the central point of the convolution kernel coincides with the central point of the image sampling frame.
With reference to the first aspect, an embodiment of the present application provides a seventh possible implementation manner of the first aspect, where the first location parameter includes: a first center abscissa and a first center ordinate, the second position parameter comprising: a second central abscissa and a second central ordinate, wherein the adjusting each sample point from a current third position to a first position according to the first position parameter and a second position parameter of a central point of the convolution kernel comprises:
adjusting the abscissa of each sampling point from the current third abscissa to the first abscissa according to the first central abscissa and the second central abscissa;
and adjusting the ordinate of each sampling point from the current third ordinate to the first ordinate according to the first central ordinate and the second central ordinate, wherein the coordinate of each sampling point is the third abscissa and the third ordinate indicates that the position of each sampling point is located at the current third position.
In combination with the first aspect, this embodiment provides an eighth possible implementation manner of the first aspect, where the adjusting the abscissa of each sample point from the current third abscissa to the first abscissa according to the first central abscissa and the second central abscissa includes:
determining a third deviation value between the first central abscissa and the second central abscissa according to the first central abscissa and the second central abscissa;
and adjusting the abscissa of each sampling point from the current third abscissa to the first abscissa according to the third deviation value.
With reference to the first aspect, this embodiment provides a ninth possible implementation manner of the first aspect, where the adjusting the ordinate of each sampling point from the current third ordinate to the first ordinate according to the first center ordinate and the second center ordinate includes:
determining a fourth deviation value between the first central ordinate and the second central ordinate according to the first central ordinate and the second central ordinate;
and adjusting the ordinate of each sampling point from the current third ordinate to the first ordinate according to the fourth deviation value.
In combination with the second aspect, an embodiment of the present application provides an image processing apparatus, including:
the shape obtaining module is used for obtaining shape parameters related to the shape of the image sampling frame;
the shape adjusting module is used for adjusting each sampling point of at least two sampling points of the convolution kernel from a current first position to a second position corresponding to the shape parameter according to the shape parameter and the convolution kernel parameter of the convolution kernel to obtain an adjusted convolution kernel, wherein the current positions of any two sampling points of the at least two sampling points are different, and the adjusted positions of any two sampling points are different;
and the convolution module is used for performing convolution processing on the image according to the adjusted convolution kernel.
In combination with the second aspect, the present embodiments provide a first possible implementation manner of the second aspect, where the shape parameter includes: the length and width of the image sampling frame, and the reduction multiple of the image sampling frame in the down-sampling;
the shape adjusting module is further configured to adjust an abscissa of each of at least two sampling points of the convolution kernel from a current first abscissa to a second abscissa corresponding to the length according to the length, the reduction multiple, and the convolution kernel parameter; according to the width, the reduction multiple and the convolution kernel parameter, adjusting the ordinate of each sampling point from the current first ordinate to the second ordinate corresponding to the width, wherein the coordinate of each sampling point before adjustment is the first abscissa and the first ordinate, which indicates that the position of each sampling point is at the current first position, and the coordinate of each sampling point after adjustment is the second abscissa and the second ordinate, which indicates that the position of each sampling point is at the second position after adjustment; and obtaining an adjusted convolution kernel, wherein the position of each sampling point in the adjusted convolution kernel is located at the second position.
In combination with the second aspect, the present application provides a second possible implementation manner of the second aspect, where according to the length, the reduction multiple and the convolution kernel parameter,
the shape adjusting module is further configured to determine a first deviation value of an abscissa of each of at least two sampling points of the convolution kernel according to the length, the reduction multiple and the convolution kernel parameter; and adjusting the abscissa of each sampling point from the current first abscissa to a second abscissa corresponding to the length according to the first deviation value of the abscissa of each sampling point.
With reference to the second aspect, an embodiment of the present application provides a third possible implementation manner of the second aspect, where the convolution kernel parameter includes: the total column number of the convolution kernel and the interval column number between the column of each sampling point and the central point of the convolution kernel;
the shape adjusting module is further configured to determine that a first deviation value of an abscissa of each sampling point in the convolution kernel is W × M/M × S according to the length W, the reduction multiple S, the total column number M, and the interval column number M.
In combination with the second aspect, the present application provides a fourth possible implementation manner of the second aspect, where according to the width, the reduction multiple and the convolution kernel parameter,
the shape adjusting module is further configured to determine a second deviation value of the ordinate of each sampling point according to the width, the reduction multiple and the convolution kernel parameter; and adjusting the ordinate of each sampling point from the current first ordinate to a second ordinate corresponding to the width according to the second deviation value of the ordinate of each sampling point.
With reference to the second aspect, an embodiment of the present application provides a fifth possible implementation manner of the second aspect, where the convolution kernel parameter further includes: the total number of rows of the convolution kernel and the number of spaced rows between the row of each sampling point and the central point of the convolution kernel;
and the shape adjusting module is further used for determining a second deviation value H x N/N x S of the vertical coordinate of each sampling point according to the width H, the reduction multiple S, the total line number N and the interval line number N.
With reference to the second aspect and any one of the first to fifth possible implementation manners of the second aspect, in this application, an embodiment provides a sixth possible implementation manner of the second aspect, and the apparatus further includes:
the position obtaining module is used for obtaining a first position parameter of the central point of the image sampling frame;
and the position adjusting module is used for adjusting each sampling point from a current third position to a first position according to the first position parameter and a second position parameter of the center point of the convolution kernel, wherein the first position of each sampling point represents that the center point of the convolution kernel coincides with the center point of the image sampling frame.
With reference to the second aspect, embodiments of the present application provide a seventh possible implementation manner of the second aspect, where the first location parameter includes: a first center abscissa and a first center ordinate, the second position parameter comprising: a second central abscissa and a second central ordinate,
the position adjusting module is further used for adjusting the abscissa of each sampling point from the current third abscissa to the first abscissa according to the first central abscissa and the second central abscissa; and adjusting the ordinate of each sampling point from the current third ordinate to the first ordinate according to the first central ordinate and the second central ordinate, wherein the coordinate of each sampling point is the third abscissa and the third ordinate indicates that the position of each sampling point is located at the current third position.
In combination with the second aspect, the present application provides an eighth possible implementation manner of the second aspect,
the position adjusting module is further configured to determine a third deviation value between the first central abscissa and the second central abscissa according to the first central abscissa and the second central abscissa; and adjusting the abscissa of each sampling point from the current third abscissa to the first abscissa according to the third deviation value.
In combination with the second aspect, the present application provides a ninth possible implementation manner of the second aspect,
the position adjusting module is further configured to determine a fourth deviation value between the first center ordinate and the second center ordinate according to the first center ordinate and the second center ordinate; and adjusting the ordinate of each sampling point from the current third ordinate to the first ordinate according to the fourth deviation value.
In a third aspect, an embodiment of the present application provides an electronic device, where the electronic device includes: a processor, a memory, a bus and a communication interface; the processor, the communication interface and the memory are connected through the bus;
the memory is used for storing programs;
the processor is configured to execute the image processing method according to the first aspect and any one of the embodiments of the first aspect by calling a program stored in the memory.
In a fourth aspect, the present application provides a computer-readable storage medium having a non-volatile program code executable by a computer, where the program code causes the computer to execute the first aspect and the image processing method described in any one of the embodiments of the first aspect.
The beneficial effects of the embodiment of the application are that:
according to the shape parameter relevant to the shape of the image sampling frame, each of at least two sampling points of the convolution kernel is adjusted from a current first position to a second position corresponding to the shape parameter, so that the position distribution of the at least two sampling points in the adjusted convolution kernel is matched with the shape of the image sampling frame. Therefore, the convolution processing is carried out on the image in the image sampling frame through the convolution kernel of which the sampling point position distribution is matched with the shape of the image sampling frame, so that the deviation of the detection result of the image can be avoided, and the detection precision is improved.
In order to make the aforementioned objects, features and advantages of the present application more comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained from the drawings without inventive effort.
Fig. 1 shows a block diagram of an electronic device according to a first embodiment of the present application;
FIG. 2 is a first flowchart of an image processing method provided in a second embodiment of the present application;
fig. 3 shows a sub-flowchart of step S120 in an image processing method according to a second embodiment of the present application;
fig. 4 shows a sub-flowchart of step S140 in an image processing method according to a second embodiment of the present application;
fig. 5 is a schematic diagram illustrating a first scene of an image processing method according to a second embodiment of the present application;
fig. 6 is a diagram illustrating a second scene of an image processing method according to a second embodiment of the present application;
fig. 7 is a block diagram showing a configuration of an image processing apparatus according to a third embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. The components of the embodiments of the present application, generally described and illustrated in the figures herein, can be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the present application, presented in the accompanying drawings, is not intended to limit the scope of the claimed application, but is merely representative of selected embodiments of the application. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present application without inventive step, are within the scope of the present application.
It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures. The terms "first," "second," and the like are used solely to distinguish one from another and are not to be construed as indicating or implying relative importance.
First embodiment
Referring to fig. 1, an electronic device 10 is provided in the embodiment of the present application, and the electronic device 10 may be a terminal device or a server. The terminal device may be a Personal Computer (PC), a tablet PC, a smart phone, a Personal Digital Assistant (PDA), or the like; the server may be a web server, a database server, a cloud server, or a server assembly composed of a plurality of sub servers, etc.
In this embodiment, the electronic device 10 may include: memory 11, communication module 12, bus 13, and processor 14. Wherein the processor 14, the communication module 12 and the memory 11 are connected by a bus 13. The processor 14 is arranged to execute executable modules, such as computer programs, stored in the memory 11. The components and configurations of electronic device 10 shown in FIG. 1 are for example, and not for limitation, and electronic device 10 may have other components and configurations as desired.
The Memory 11 may include a high-speed Random Access Memory (Random Access Memory RAM) and may also include a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. In the present embodiment, the memory 11 stores a program necessary for executing the image processing method.
The bus 13 may be an ISA bus, a PCI bus, an EISA bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one double-headed arrow is shown in FIG. 1, but this does not indicate only one bus or one type of bus.
The processor 14 may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by instructions in the form of hardware, integrated logic circuits, or software in the processor 14. The Processor 14 may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; but may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components. The various methods, steps and logic blocks disclosed in the embodiments of the present invention may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present invention may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module may be located in ram, flash memory, rom, prom, or eprom, registers, etc. storage media as is well known in the art.
The method performed by the flow process or the defined device disclosed in any of the embodiments of the present invention may be applied to the processor 14, or may be implemented by the processor 14. After the processor 14 receives the execution instruction and calls the program stored in the memory 11 through the bus 13, the processor 14 controls the communication module 12 through the bus 13 to execute the flow of the image processing method.
Second embodiment
The present embodiment provides an image processing method, it should be noted that the steps shown in the flowchart of the figure may be performed in a computer system such as a set of computer executable instructions, and that while a logical order is shown in the flowchart, in some cases the steps shown or described may be performed in an order different than here. The present embodiment will be described in detail below.
Referring to fig. 2, in the image processing method provided in the present embodiment, the image processing method includes: step S110, step S120, step S130, step S140, and step S150.
Step S110: and obtaining a first position parameter of the central point of the image sampling frame.
Step S120: and adjusting each sampling point from the current third position to the first position according to the first position parameter and a second position parameter of the central point of the convolution kernel, wherein the position of each sampling point at the first position indicates that the central point of the convolution kernel coincides with the central point of the image sampling frame.
Step S130: shape parameters related to the shape of the image sampling frame are obtained.
Step S140: and adjusting each sampling point of at least two sampling points of the convolution kernel from a current first position to a second position corresponding to the shape parameter according to the shape parameter and the convolution kernel parameter of the convolution kernel to obtain an adjusted convolution kernel, wherein the current positions of any two sampling points of the at least two sampling points are different, and the adjusted positions of any two sampling points are different.
Step S150: and performing convolution processing on the image according to the adjusted convolution kernel.
The steps of the present application will be described in detail below with reference to fig. 2 to 4.
Step S110: and obtaining a first position parameter of the central point of the image sampling frame.
The electronic device may perform detection and identification on the object in the original image, for example, the electronic device may perform detection and identification on the object in the original image by using FasterRCNN, RFCN, SSD, RetinaNet, RefineDet, YOLOv2, and the like. The electronic device can generate an image sampling frame for framing the target object in the original image according to a preset control program and the position and the size of the target object in the original image; or the electronic equipment can respond to the image sampling frame generation operation of the user to generate the image sampling frame expected by the user.
It will be appreciated that the image sample box may exist in a virtual manner, i.e. the actual original image may not have an actual box that can be seen by the user.
In this embodiment, the electronic device may pre-establish a reference coordinate system according to the relative position of each pixel in the original image. Since the image sample frame is located in the original image, the image sample frame is correspondingly located in the reference coordinate system. Therefore, in the process of generating the image sampling frame by the electronic device, the electronic device may also correspondingly obtain the position of the image sampling frame in the reference coordinate system, that is, the electronic device may also correspondingly obtain the first position parameter of the central point of the image sampling frame in the reference coordinate system. The first position parameter may be a first center abscissa and a first center ordinate of the center point of the image sample frame in the reference coordinate system.
Step S120: and adjusting each sampling point from the current third position to the first position according to the first position parameter and a second position parameter of the central point of the convolution kernel, wherein the position of each sampling point at the first position indicates that the central point of the convolution kernel coincides with the central point of the image sampling frame.
In this embodiment, step S120 may include: step S121 and step S122.
Step S121: and adjusting the abscissa of each sampling point from the current third abscissa to the first abscissa according to the first central abscissa and the second central abscissa.
Step S122: and adjusting the ordinate of each sampling point from the current third ordinate to the first ordinate according to the first central ordinate and the second central ordinate, wherein the coordinate of each sampling point is the third abscissa and the third ordinate indicates that the position of each sampling point is located at the current third position.
The electronic device generates an image sampling frame, and simultaneously, the electronic device can also generate a convolution kernel according to a preset control program to perform convolution processing on the image so as to realize sampling of the image, wherein the image can be any image which contains the original image in a feature pyramid of the image generated by the electronic device based on the downsampling of the original image.
In this embodiment, in order to facilitate accurate convolution of an image, a central point of a convolution kernel may coincide with a central point of an image sampling frame when the image is stepped to a certain position in the convolution process. Then, in order to achieve the effect, after the convolution kernel is generated, the central point of the convolution kernel and the central point of the image sampling frame are adjusted to coincide, so that the convolution kernel can be ensured to step to coincide with the central point of the image sampling frame in the subsequent convolution process.
In addition, when the image sample frame is generated in response to the user's image sample frame generation operation, since the image sample frame is generated based on the user operation, the user can ensure that the center point of the image sample frame coincides with the center point of the generated convolution kernel at the time of generation. However, if the image sampling frame is automatically generated by the electronic device, the automatically generated image sampling frame may be a sampling frame that is more accurate with respect to the image and is regressed after the electronic device performs sub-sampling on the image sampling frame generated by the user operation, so that the center point of the image sampling frame may not coincide with the center point of the convolution kernel.
For convenience of understanding, the description of adjusting the center point of the convolution kernel to coincide with the center point of the image sampling frame in this embodiment does not refer to whether the image sampling frame is generated by a user operation or automatically by an electronic device, but is not limited to this. In practical cases, if the image sampling frame is generated based on a user operation, the electronic device may not adjust the center point of the convolution kernel.
In this embodiment, the electronic device may adjust the center point of the convolution kernel to coincide with the center point of the image sampling frame by executing steps S121 and S122, and the execution of steps S121 and S122 will be described below:
in detail, the electronic device may obtain a second position parameter of the center point of the convolution kernel in the reference coordinate system based on the position of the convolution kernel in the reference coordinate system. The second location parameter may be a second center abscissa and a second center ordinate of the center point of the convolution kernel in the reference coordinate system.
Since the convolution kernel may have at least two samples, for example, a convolution kernel with a convolution kernel a of 5 × 5, the convolution kernel a may have 25 samples. Therefore, for the convolution kernel, each of the at least two samples of the convolution kernel may be located at a current third location with respect to the center point of the convolution kernel, and any two samples of the at least two samples are located at different current locations, that is, any two third locations where any two samples are located are not the same.
When the center point of the convolution kernel is adjusted to coincide with the center point of the graphic sample frame, the position of each sample point in the convolution kernel needs to be changed correspondingly because the position of the center point of the convolution kernel is changed. Therefore, the difference value between the center point of the convolution kernel and the center point of the graph sampling frame can be used as a basis for adjusting each sampling point, so that each sampling point is adjusted from the current third position to the first position, wherein the position of each sampling point at the current third position can represent that the coordinates of each sampling point are the third abscissa and the third ordinate.
And the electronic equipment determines a third deviation value between the first central abscissa and the second central abscissa according to the first central abscissa and the second central abscissa, and the electronic equipment can adjust the abscissa of each sampling point from the current third abscissa to the first abscissa according to the third deviation value. And the electronic equipment determines a fourth deviation value between the first center ordinate and the second center ordinate according to the first center ordinate and the second center ordinate, and the electronic equipment can adjust the ordinate of each sampling point from the current third ordinate to the first ordinate according to the fourth deviation value. Thus, before the adjustment is subsequently continued, the coordinate of each sampling point being the first abscissa and the first ordinate may indicate that each sampling point has been currently adjusted to the first position, where any two first positions where any two sampling points of the at least two sampling points are currently located are also different.
Assuming that the generated convolution kernel a is a convolution kernel of 5 × 5, the second position parameter of the coordinate of the center point of the convolution kernel a in the reference coordinate system includes: (Xa, Ya). And the coordinates of the center point of the convolution kernel a are used as a reference for at least two sampling points in the convolution kernel a, the coordinates of the at least two sampling points in the convolution kernel a may be:
wherein, formula 1 is a distribution of at least two second center abscissas of at least two sampling points of the convolution kernel a, formula 2 is a distribution of at least two second center ordinates of at least two sampling points of the convolution kernel a, and formula 1 and formula 2 show that each sampling point is located at the current third position.
The first position parameter of the coordinates of the central point of the image sampling frame B in the reference coordinate system comprises: (Xb, Yb), when the midpoint of the convolution kernel a is adjusted to coincide with the central point of the image sampling frame B according to the first position parameter (Xb, Yb) and the second position parameter (Xa, Ya), the coordinates of at least two sampling points in the convolution kernel a may be correspondingly adjusted as:
wherein, formula 3 is a distribution when the midpoints of the convolution kernel a and the image sampling frame B are adjusted to coincide with the central point of the image sampling frame B, formula 4 is a distribution when the midpoints of the convolution kernel a and the image sampling frame B are adjusted to coincide with the central point of the image sampling frame B, and formulas 3 and 4 show that each sampling point is located at the current first position.
In this embodiment, the shape of the generated convolution kernel is generally a standard shape, that is, each sampling point in the convolution kernel is equidistant from adjacent sampling points, so that the shape of the convolution kernel is not matched with the shape of an image sampling frame with strong shape randomness, and sampling deviation occurs. The electronic device may not only adjust the center point of the convolution kernel to coincide with the center point of the image sampling frame, but the electronic device may also perform steps S130 to S140 to adjust the shape of the convolution kernel to match the shape of the image sampling frame.
Step S130: shape parameters related to the shape of the image sampling frame are obtained.
Since the size of the image can be determined by the number of the pixels, the size of the image can be correspondingly represented in the reference coordinate system, and the size of the image sampling frame in the image can be represented in the reference coordinate system. Therefore, while generating the image sampling frame, the electronic device may also obtain shape parameters related to the shape of the image sampling frame, where the shape parameters related to the shape may include: the length and width of the image sample box in the reference coordinate system. In the detection process, since the image is any image in which the feature pyramid of the image includes the original image, and the image may be reduced compared to the original image, and the reduced image sampling frame in the image may also be reduced in proportion to an image sampling frame that is generated automatically or based on a user operation at the beginning, the shape-related shape parameters obtained by the electronic device may further include: the reduction multiple of the image sampling frame in the down-sampling, namely the reduction multiple of the image sampling frame, can obtain the reduced image sampling frame in the image.
Step S140: and adjusting each sampling point of at least two sampling points of the convolution kernel from a current first position to a second position corresponding to the shape parameter according to the shape parameter and the convolution kernel parameter of the convolution kernel to obtain an adjusted convolution kernel, wherein the current positions of any two sampling points of the at least two sampling points are different, and the adjusted positions of any two sampling points are different.
In this embodiment, step S140 may include: step S141, step S142, and step S143.
Step S141: and adjusting the abscissa of each of at least two sampling points of the convolution kernel from the current first abscissa to a second abscissa corresponding to the length according to the length, the reduction multiple and the convolution kernel parameter.
Step S142: and adjusting the vertical coordinate of each sampling point from the current first vertical coordinate to a second vertical coordinate corresponding to the width according to the width, the reduction multiple and the convolution kernel parameter.
Step S143: and obtaining an adjusted convolution kernel, wherein the position of each sampling point in the adjusted convolution kernel is located at the second position.
In this embodiment, while the electronic device generates the convolution kernel, the electronic device may also obtain a convolution kernel parameter of the convolution kernel. The convolution kernel parameters may include: the total number of rows and the total number of columns of the convolution kernel, and, in the case of determining the center point of the convolution kernel, the parameters of the convolution kernel may further include: the number of spaced columns between the column of each sample point and the center point of the convolution kernel, and the number of spaced rows between the row of each sample point and the center point of the convolution kernel.
For example, if the convolution kernel a is a 5 × 5 convolution kernel, then the total number of rows and the total number of columns of the convolution kernel a may be: 5 rows and 5 columns. The number of spaced columns of sample points located on the edge-most column and the center point of the convolution kernel a may be 3. And, the number of spaced rows of the sampling points located on a row adjacent to the edge-most row from the center point of the convolution kernel a may be 2.
In this embodiment, by executing step S141 and step S142, the electronic device may adjust the current first abscissa of each sampling point to the second abscissa corresponding to the length, and may further adjust the current first ordinate of each sampling point to the second ordinate corresponding to the length. Therefore, the coordinates of each sampling point correspond to the length and the width, the shape of the convolution kernel is matched with the shape of the image sampling frame, and the center point of the convolution kernel can be adjusted to coincide with the center point of the image sampling frame. The execution of step S141, step S142, and step S143 will be described below:
according to the length, the reduction multiple, and the total column number and the interval column number in the parameter of the convolution kernel, the electronic equipment can determine a first deviation value generated by the abscissa of each of at least two sampling points of the convolution kernel relative to the length reduced by the reduction multiple. Then, the electronic device determines a first deviation value according to the abscissa of each sampling point, and then may adjust the abscissa of each sampling point from the current first abscissa to a second abscissa corresponding to the length.
Meanwhile, according to the width, the reduction multiple, and the total line number and the interval line number in the parameter of the convolution kernel, the electronic device can determine a second deviation value generated by the ordinate of each of the at least two sampling points of the convolution kernel relative to the width reduced by the reduction multiple. Then, the electronic device determines a second deviation value according to the ordinate of each sampling point, and may adjust the ordinate of each sampling point from the current first ordinate to a second ordinate corresponding to the width. In this way, the coordinates of each adjusted sampling point as the second abscissa and the second ordinate may indicate that each sampling point has been adjusted to the second position at present, where any two first positions where any two sampling points of the at least two sampling points are located are also different.
Optionally, the relationship between the determined first deviation value and the length, the reduction multiple, the total number of columns, and the number of interval columns may be: the length W, the reduction factor S, the total number of columns M, and the number of interval columns M, and the first deviation value of the abscissa of each sampling point in the determined convolution kernel may be W × M/M × S. And the determined relationship between the second deviation value and the width, the reduction multiple, the total line number and the spacing line number can be as follows: the width H, the reduction multiple S, the total number of rows N, and the number of spacing rows N, and the determined second deviation value of the ordinate of each sampling point in the convolution kernel may be H × N/N × S.
It will be appreciated that the more samples in the convolution kernel are in a row or column near the edge of the convolution kernel, the greater the magnitude of the adjustment of these samples.
Continuing with the foregoing assumptions, equations 3 and 4 then indicate that each sample point in convolution kernel a is located at the current first position. If the length of the image sampling frame B is W, the reduction multiple is S, the total column number is M and the interval column number is M, and the width of the image sampling frame B is H, the total row number is N and the interval row number is N. The coordinate distribution of each sample point of the adjusted convolution kernel a at the current second position can be as follows:
where equation 5 is an abscissa distribution when each of the at least two sample points of the convolution kernel a is located at the second position, and equation 6 is an ordinate distribution when each of the at least two sample points of the convolution kernel a is located at the second position. Then equations 5 and 6 represent the modulated convolution kernel.
Step S150: and performing convolution processing on the image according to the adjusted convolution kernel.
After the electronic device obtains the adjusted convolution kernel, the features in the image sampling frame of the image can be convolved according to the adjusted convolution kernel.
It is to be understood that the convolution kernel adjusted to convolve the feature objects within the image sample frame of the image may be: and the adjusted convolution kernel carries out convolution processing on the characteristics of the position where the adjusted convolution kernel is located at present. If the preset step length of the convolution kernel is stepped to the next sampling position in the image sampling frame, the position distribution of at least two sampling points of the convolution kernel can be readjusted to be matched with the shape of the image sampling frame according to the method, so that the adjusted convolution kernel is obtained again to sample the next sampling position.
It should be noted that, in the convolution process of the adjusted convolution kernel, the sampling characteristic of each sampling point in the adjusted convolution kernel may be obtained by performing bilinear interpolation on the values of the four nearest discrete characteristic points to each sampling point.
As shown in fig. 5 and 6, the convolution kernel a1 before adjustment in fig. 5 is A3 × 3 convolution kernel, each sample point (circle shown by dotted line) in the convolution kernel a1 is at the first position before adjustment, and the image sample box is B. Then, each sample point (circle shown by solid line) in the adjusted convolution kernel a2 with the image sample frame B is obtained by the adjustment at the second position before the adjustment.
Third embodiment
Referring to fig. 7, an embodiment of the present application provides an image processing apparatus 100, where the image processing apparatus 100 may be applied to an electronic device, and the image processing apparatus 100 includes:
a shape obtaining module 110, configured to obtain shape parameters related to the shape of the image sample frame.
And a shape adjusting module 120, configured to adjust, according to the shape parameter and a convolution kernel parameter of a convolution kernel, each of at least two sampling points of the convolution kernel from a current first position to a second position corresponding to the shape parameter, so as to obtain an adjusted convolution kernel, where current positions of any two sampling points of the at least two sampling points are different, and adjusted positions of any two sampling points are also different.
And a convolution module 130, configured to perform convolution processing on the image according to the adjusted convolution kernel.
It should be noted that, as those skilled in the art can clearly understand, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, embodiments of the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, embodiments of the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
In summary, the embodiments of the present application provide an image processing method, an image processing apparatus, an electronic device, and a storage medium. The method comprises the following steps: obtaining shape parameters related to the shape of the image sampling frame; adjusting each sampling point of at least two sampling points of the convolution kernel from a current first position to a second position corresponding to the shape parameter according to the shape parameter and the convolution kernel parameter of the convolution kernel to obtain an adjusted convolution kernel, wherein the current positions of any two sampling points of the at least two sampling points are different, and the adjusted positions of any two sampling points are also different; and carrying out convolution processing on the image in the image sampling frame according to the adjusted convolution kernel.
According to the shape parameter relevant to the shape of the image sampling frame, each of at least two sampling points of the convolution kernel is adjusted from a current first position to a second position corresponding to the shape parameter, so that the position distribution of the at least two sampling points in the adjusted convolution kernel is matched with the shape of the image sampling frame. Therefore, the convolution processing is carried out on the image in the image sampling frame through the convolution kernel of which the sampling point position distribution is matched with the shape of the image sampling frame, so that the deviation of the detection result of the image can be avoided, and the detection precision is improved.
The above description is only a preferred embodiment of the present application and is not intended to limit the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application. It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures.
The above description is only for the specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present application, and shall be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (12)
1. An image processing method, characterized in that the method comprises:
obtaining shape parameters related to the shape of the image sampling frame;
adjusting each sampling point of at least two sampling points of the convolution kernel from a current first position to a second position corresponding to the shape parameter according to the shape parameter and the convolution kernel parameter of the convolution kernel to obtain an adjusted convolution kernel, wherein the current positions of any two sampling points of the at least two sampling points are different, and the adjusted positions of any two sampling points are different;
performing convolution processing on the image according to the adjusted convolution kernel;
before the obtaining the shape of the image sample box, the method further comprises:
obtaining a first position parameter of a central point of the image sampling frame;
and adjusting each sampling point from the current third position to the first position according to the first position parameter and a second position parameter of the central point of the convolution kernel, wherein the position of each sampling point at the first position indicates that the central point of the convolution kernel coincides with the central point of the image sampling frame.
2. The image processing method according to claim 1, wherein the shape parameter includes: the length and width of the image sampling frame, and the reduction multiple of the image sampling frame in the down-sampling; adjusting each of at least two sampling points of the convolution kernel from a current first position to a second position corresponding to the shape parameter according to the shape parameter and a convolution kernel parameter of the convolution kernel to obtain an adjusted convolution kernel, including:
according to the length, the reduction multiple and the convolution kernel parameter, adjusting the abscissa of each of at least two sampling points of the convolution kernel from the current first abscissa to a second abscissa corresponding to the length;
according to the width, the reduction multiple and the convolution kernel parameter, adjusting the ordinate of each sampling point from the current first ordinate to the second ordinate corresponding to the width, wherein the coordinate of each sampling point before adjustment is the first abscissa and the first ordinate, which indicates that the position of each sampling point is at the current first position, and the coordinate of each sampling point after adjustment is the second abscissa and the second ordinate, which indicates that the position of each sampling point is at the second position after adjustment;
and obtaining an adjusted convolution kernel, wherein the position of each sampling point in the adjusted convolution kernel is located at the second position.
3. The image processing method according to claim 2, wherein said adjusting an abscissa of each of at least two sample points of the convolution kernel from a current first abscissa to a second abscissa corresponding to the length according to the length, the reduction factor, and the convolution kernel parameter comprises:
determining a first deviation value of the abscissa of each of at least two sampling points of the convolution kernel according to the length, the reduction multiple and the convolution kernel parameter;
and adjusting the abscissa of each sampling point from the current first abscissa to a second abscissa corresponding to the length according to the first deviation value of the abscissa of each sampling point.
4. The image processing method according to claim 3, wherein the convolution kernel parameters include: the total column number of the convolution kernel and the interval column number between the column of each sampling point and the central point of the convolution kernel; the determining a first deviation value of the abscissa of each sampling point in the convolution kernel according to the length, the reduction multiple and the convolution kernel parameter includes:
and determining that the first deviation value of the abscissa of each sampling point in the convolution kernel is W × M/M × S according to the length W, the reduction multiple S, the total column number M and the interval column number M.
5. The method according to claim 2, wherein said adjusting the ordinate of each sample point from the current first ordinate to the second ordinate corresponding to the width according to the width, the reduction factor and the convolution kernel parameter comprises:
determining a second deviation value of the ordinate of each sampling point according to the width, the reduction multiple and the convolution kernel parameter;
and adjusting the ordinate of each sampling point from the current first ordinate to a second ordinate corresponding to the width according to the second deviation value of the ordinate of each sampling point.
6. The image processing method of claim 5, wherein the convolution kernel parameters further comprise: the total number of rows of the convolution kernel and the number of spaced rows between the row of each sampling point and the central point of the convolution kernel; determining a second deviation value of the ordinate of each sampling point according to the width, the reduction multiple and the convolution kernel parameter, wherein the second deviation value comprises;
and determining a second deviation value H x N/N x S of the vertical coordinate of each sampling point according to the width H, the reduction multiple S, the total line number N and the interval line number N.
7. The image processing method according to claim 1, wherein the first position parameter comprises: a first center abscissa and a first center ordinate, the second position parameter comprising: a second central abscissa and a second central ordinate, wherein the adjusting each sample point from a current third position to a first position according to the first position parameter and a second position parameter of a central point of the convolution kernel comprises:
adjusting the abscissa of each sampling point from the current third abscissa to the first abscissa according to the first central abscissa and the second central abscissa;
and adjusting the ordinate of each sampling point from the current third ordinate to the first ordinate according to the first central ordinate and the second central ordinate, wherein the coordinate of each sampling point is the third abscissa and the third ordinate indicates that the position of each sampling point is located at the current third position.
8. The image processing method according to claim 7, wherein the adjusting the abscissa of each sample point from the current third abscissa to the first abscissa according to the first central abscissa and the second central abscissa comprises:
determining a third deviation value between the first central abscissa and the second central abscissa according to the first central abscissa and the second central abscissa;
and adjusting the abscissa of each sampling point from the current third abscissa to the first abscissa according to the third deviation value.
9. The image processing method according to claim 7, wherein the adjusting the ordinate of each sample point from the current third ordinate to the first ordinate according to the first center ordinate and the second center ordinate comprises:
determining a fourth deviation value between the first central ordinate and the second central ordinate according to the first central ordinate and the second central ordinate;
and adjusting the ordinate of each sampling point from the current third ordinate to the first ordinate according to the fourth deviation value.
10. An image processing apparatus, characterized in that the apparatus comprises:
the shape obtaining module is used for obtaining shape parameters related to the shape of the image sampling frame;
the shape adjusting module is used for adjusting each sampling point of at least two sampling points of the convolution kernel from a current first position to a second position corresponding to the shape parameter according to the shape parameter and the convolution kernel parameter of the convolution kernel to obtain an adjusted convolution kernel, wherein the current positions of any two sampling points of the at least two sampling points are different, and the adjusted positions of any two sampling points are different;
the convolution module is used for carrying out convolution processing on the image according to the adjusted convolution kernel;
the device further comprises: the position obtaining module is used for obtaining a first position parameter of the central point of the image sampling frame;
and the position adjusting module is used for adjusting each sampling point from a current third position to a first position according to the first position parameter and a second position parameter of the center point of the convolution kernel, wherein the first position of each sampling point represents that the center point of the convolution kernel coincides with the center point of the image sampling frame.
11. An electronic device, characterized in that the electronic device comprises: the number of operations performed by the processor, the memory,
a bus and a communication interface; the processor, the communication interface and the memory are connected through the bus;
the memory is used for storing programs;
the processor for executing the image processing method according to any one of claims 1 to 9 by calling a program stored in the memory.
12. A computer-readable storage medium having computer-executable non-volatile program code, wherein the program code causes the computer to perform the image processing method according to any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811071355.4A CN109325945B (en) | 2018-09-13 | 2018-09-13 | Image processing method, image processing device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811071355.4A CN109325945B (en) | 2018-09-13 | 2018-09-13 | Image processing method, image processing device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109325945A CN109325945A (en) | 2019-02-12 |
| CN109325945B true CN109325945B (en) | 2021-01-08 |
Family
ID=65265372
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811071355.4A Active CN109325945B (en) | 2018-09-13 | 2018-09-13 | Image processing method, image processing device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109325945B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112308105B (en) * | 2019-08-02 | 2024-04-12 | 北京图森智途科技有限公司 | Target detection method, target detector and related equipment |
| CN114511712A (en) * | 2022-01-29 | 2022-05-17 | 上海商汤智能科技有限公司 | Target object determination method and device, electronic equipment and storage medium |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9665802B2 (en) * | 2014-11-13 | 2017-05-30 | Nec Corporation | Object-centric fine-grained image classification |
| CN105550750B (en) * | 2015-12-21 | 2017-10-17 | 长沙网动网络科技有限公司 | The method for improving convolutional neural networks accuracy of identification |
| US10572777B2 (en) * | 2016-03-11 | 2020-02-25 | Nec Corporation | Deep deformation network for object landmark localization |
| CN107292319A (en) * | 2017-08-04 | 2017-10-24 | 广东工业大学 | The method and device that a kind of characteristic image based on deformable convolutional layer is extracted |
| CN107609541B (en) * | 2017-10-17 | 2020-11-10 | 哈尔滨理工大学 | Human body posture estimation method based on deformable convolution neural network |
| CN108510467B (en) * | 2018-03-28 | 2022-04-08 | 西安电子科技大学 | SAR image target identification method based on depth deformable convolution neural network |
-
2018
- 2018-09-13 CN CN201811071355.4A patent/CN109325945B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN109325945A (en) | 2019-02-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11854173B2 (en) | System and method for finding lines in an image with a vision system | |
| CN109086734B (en) | Method and device for positioning pupil image in human eye image | |
| CN106560840B (en) | A kind of image information identifying processing method and device | |
| CN107886026B (en) | graphic code processing method and device | |
| CN109784250B (en) | Positioning method and device of automatic guide trolley | |
| CN109102026B (en) | Vehicle image detection method, device and system | |
| CN109325945B (en) | Image processing method, image processing device, electronic equipment and storage medium | |
| CN108961260B (en) | Image binarization method and device and computer storage medium | |
| CN112418243A (en) | Feature extraction method and device and electronic equipment | |
| CN111754429A (en) | Motion vector post-processing method and device, electronic device and storage medium | |
| CN106663317B (en) | Morphological processing method and digital image processing device for digital image | |
| CN112766256B (en) | Grating phase diagram processing method and device, electronic equipment and storage medium | |
| TW201911230A (en) | Surveillance method, computing device, and non-transitory storage medium | |
| CN111768345B (en) | Correction method, device, equipment and storage medium for identity card back image | |
| CN111340788B (en) | Hardware Trojan horse layout detection method and device, electronic equipment and readable storage medium | |
| CN113129298A (en) | Definition recognition method of text image | |
| CN113051778A (en) | Clothing design method and device, electronic equipment and storage medium | |
| CN117057376A (en) | Method, device, equipment and medium for obtaining target positioning frame | |
| CN112580638B (en) | Text detection method and device, storage medium and electronic equipment | |
| CN109766953B (en) | Object identification method and device | |
| CN111428707B (en) | Method and device for identifying pattern identification code, storage medium and electronic equipment | |
| CN110782425A (en) | Image processing method, image processing device and electronic equipment | |
| CN115100225A (en) | Method and device for determining error field of camera view field, electronic equipment and medium | |
| CN113033542A (en) | Method and device for generating text recognition model | |
| CN109214230B (en) | Data matrix code identification method and device and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |