CN113408632A - Method and device for improving image classification accuracy, electronic equipment and storage medium - Google Patents
Method and device for improving image classification accuracy, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN113408632A CN113408632A CN202110719033.1A CN202110719033A CN113408632A CN 113408632 A CN113408632 A CN 113408632A CN 202110719033 A CN202110719033 A CN 202110719033A CN 113408632 A CN113408632 A CN 113408632A
- Authority
- CN
- China
- Prior art keywords
- category
- threshold parameter
- classification
- value
- feedback
- 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
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000012549 training Methods 0.000 claims abstract description 20
- 238000013145 classification model Methods 0.000 claims abstract description 19
- 238000012545 processing Methods 0.000 claims abstract description 18
- 238000012795 verification Methods 0.000 claims description 36
- 230000002452 interceptive effect Effects 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 10
- 238000005070 sampling Methods 0.000 claims 2
- 238000013473 artificial intelligence Methods 0.000 abstract description 8
- 238000013135 deep learning Methods 0.000 abstract description 4
- 238000005516 engineering process Methods 0.000 description 10
- 238000004891 communication Methods 0.000 description 8
- 241000282326 Felis catus Species 0.000 description 7
- 238000010586 diagram Methods 0.000 description 5
- 230000003993 interaction Effects 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Image Analysis (AREA)
Abstract
The disclosure provides a method, a device, electronic equipment and a storage medium for improving image classification accuracy, and relates to the field of artificial intelligence such as deep learning and computer vision, wherein the method comprises the following steps: acquiring an image input by a first user, and classifying the image by using an image classification model obtained by pre-training and a predetermined threshold parameter corresponding to each category to obtain a classification result comprising at least one category; obtaining a feedback result of the first user for any category in the classification results, wherein the feedback result is a positive feedback result or a negative feedback result; when the number of the feedback results of the first user acquired for any category reaches M, the M is a positive integer larger than one, and the threshold parameter corresponding to the category is adjusted according to the M feedback results. By applying the scheme disclosed by the invention, the accuracy of the classification result can be improved, and personalized processing aiming at different users can be realized.
Description
Technical Field
The present disclosure relates to the field of artificial intelligence technology, and more particularly to the field of deep learning and computer vision.
Background
When the image classification model is used for image classification, aiming at the images input by a user, the probabilities that the images belong to different categories can be output through the model, then the different probabilities can be compared with the threshold parameters of the corresponding categories respectively, and if the probabilities are larger than the threshold parameters, the images can be determined to belong to the categories.
After the image classification model is obtained by training with the training set, namely the model parameters are determined, the threshold parameter of each category can be determined through the verification set, and the threshold parameters are suitable for all users.
However, the data distribution of the verification set is difficult to keep consistent with that of the real user, for example, for the category of "cat", the image in the verification set is difficult to cover all cats in the real scene, so that the accuracy of the determined threshold parameter is reduced, and further, the accuracy of the classification result is reduced.
Disclosure of Invention
The present disclosure provides a method, an apparatus, an electronic device, and a storage medium for improving image classification accuracy.
A method of improving image classification accuracy, comprising:
acquiring an image input by a first user, and classifying the image by using an image classification model obtained by pre-training and a predetermined threshold parameter corresponding to each category to obtain a classification result comprising at least one category;
obtaining a feedback result of the first user for any category in the classification results, wherein the feedback result is a positive feedback result or a negative feedback result;
when the number of the feedback results of the first user acquired for any category reaches M, the M is a positive integer larger than one, and the threshold parameter corresponding to the category is adjusted according to the M feedback results.
An apparatus for improving image classification accuracy, comprising: the device comprises a classification module, an acquisition module and an adjustment module;
the classification module is used for acquiring an image input by a first user, classifying the image by utilizing an image classification model obtained by pre-training and a predetermined threshold parameter corresponding to each category to obtain a classification result comprising at least one category;
the obtaining module is configured to obtain a feedback result of the first user for any category in the classification results, where the feedback result is a positive or negative feedback result;
the adjusting module is configured to, when the number of the feedback results of the first user obtained for any category reaches M, where M is a positive integer greater than one, adjust the threshold parameter corresponding to the category according to the M feedback results.
An electronic device, comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform a method as described above.
A non-transitory computer readable storage medium storing computer instructions for causing a computer to perform the method as described above.
A computer program product comprising a computer program which, when executed by a processor, implements a method as described above.
One embodiment in the above disclosure has the following advantages or benefits: the threshold parameters corresponding to any category can be adjusted, namely optimized, through interaction with the user according to the feedback result of the user, so that the accuracy of the classification result is improved, and the threshold parameters corresponding to the user are adjusted by taking the user as the adjustment granularity, so that personalized processing and the like aiming at different users are realized.
It should be understood that the statements in this section do not necessarily identify key or critical features of the embodiments of the present disclosure, nor do they limit the scope of the present disclosure. Other features of the present disclosure will become apparent from the following description.
Drawings
The drawings are included to provide a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:
FIG. 1 is a flow chart of a first embodiment of a method for improving image classification accuracy according to the present disclosure;
FIG. 2 is a flowchart illustrating a second embodiment of a method for improving image classification accuracy according to the present disclosure;
FIG. 3 is a schematic diagram illustrating a structure of an embodiment 300 of an apparatus for improving image classification accuracy according to the present disclosure;
FIG. 4 shows a schematic block diagram of an example electronic device 400 that may be used to implement embodiments of the present disclosure.
Detailed Description
Exemplary embodiments of the present disclosure are described below with reference to the accompanying drawings, in which various details of the embodiments of the disclosure are included to assist understanding, and which are to be considered as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.
In addition, it should be understood that the term "and/or" herein is merely one type of association relationship that describes an associated object, meaning that three relationships may exist, e.g., a and/or B may mean: a exists alone, A and B exist simultaneously, and B exists alone. In addition, the character "/" herein generally indicates that the former and latter related objects are in an "or" relationship.
Fig. 1 is a flowchart illustrating a first embodiment of a method for improving image classification accuracy according to the present disclosure. As shown in fig. 1, the following detailed implementation is included.
In step 101, an image input by a first user is obtained, and the image is classified by using an image classification model obtained through pre-training and a predetermined threshold parameter corresponding to each category, so as to obtain a classification result including at least one category.
In step 102, a feedback result of the first user for any category in the classification results is obtained, where the feedback result is a positive or negative feedback result.
In step 103, when the number of the feedback results of the first user obtained for any category reaches M, where M is a positive integer greater than one, the threshold parameter corresponding to the category is adjusted according to the M feedback results.
It can be seen that, in the scheme of the embodiment of the method, the threshold parameter corresponding to any category can be adjusted, i.e., optimized, according to the feedback result of the user through interaction with the user, so that the accuracy of the classification result is improved, and the threshold parameter corresponding to the user is adjusted by using the user as the adjustment granularity, so that personalized processing and the like for different users are realized.
The image classification model may be obtained by pre-training, that is, the image classification model may be obtained by pre-training using a training set (i.e., training data), and the model parameters may be determined, and then the threshold parameters corresponding to each category may be determined by using a verification set, where the threshold parameters are applicable to all users.
On the basis, any user can be respectively processed according to the mode disclosed by the disclosure. For ease of description, "any user" will be referred to as a first user.
For any image input by the first user, the image classification model obtained by pre-training and the threshold parameter corresponding to each predetermined category can be used for classifying the image, so that a classification result comprising at least one category is obtained.
For example, assuming that 10 classes, i.e., class a to class j, are included, using the image classification model, the probability that the image belongs to the category a to the category j can be respectively obtained, then the probability belonging to the category a can be compared with the threshold parameter corresponding to the category a, the probability belonging to the category b can be compared with the threshold parameter corresponding to the category b, the probability belonging to the category c can be compared with the threshold parameter corresponding to the category c, and so on, and if the probability belonging to the category b is greater than the threshold parameter corresponding to the category b, the image can be determined to belong to the category b, i.e., category b is included in the classification result, assuming that the probability of belonging to category b is greater than the threshold parameter corresponding to category b, and the probability of belonging to the class c is greater than the threshold parameter corresponding to the class c, then it can be determined that the image belongs to the class b and the class c, i.e. the classification result includes the class b and the class c.
And then, a feedback result of the first user for any category in the classification results can be obtained, wherein the feedback result is a positive feedback result or a negative feedback result, the positive feedback result indicates that the classification is correct, and the negative feedback result indicates that the classification is wrong.
In one embodiment of the disclosure, an interactive interface may be generated and displayed for the classification result, and a feedback result of the user is obtained through the interactive interface.
For example, a classification result, such as a category b, may be displayed in the interactive interface, and for the category b, a button may be displayed, after the user clicks the button, two options of "yes" and "no" may be displayed in a pull-down manner, and the user may select one of the options, if the option selected by the user is "yes", the feedback result of the category b is correct, that is, the image belongs to the category b, and if the option selected by the user is "no", the feedback result of the category b is incorrect, that is, the image does not belong to the category b.
For another example, classification results such as a category b and a category c may be displayed in the interactive interface, and for the category b, a button may be displayed, when the user clicks the button, two options of "yes" and "no" may be displayed in a pull-down manner, the user may select one of the options, if the option selected by the user is "yes", the feedback result of the category b is correct, that is, the image belongs to the category b, whereas, if the option selected by the user is "no", the feedback result of the category b is incorrect, that is, the image does not belong to the category b, similarly, for the category c, a button may be displayed, when the user clicks the button, two options of "yes" and "no" may be displayed in a pull-down manner, the user may select one of the options, if the option selected by the user is "yes", the feedback result of the category c is correct, that is, the image belongs to the category c, whereas if the option selected by the user is "no", the feedback result indicating the category c is an error, that is, the image does not belong to the category c.
In practical applications, the classification result may include only one category, or may include a plurality of categories, and when the classification result includes a plurality of categories, one of the categories may be wrong, the other category may be correct, both categories may be wrong, and the like.
Through the interactive interface, the feedback result of the user can be conveniently and efficiently obtained, the user can feed back the feedback result by taking the category as the granularity, and the categories do not influence each other, so that the respective optimization aiming at different categories can be realized.
Further, when the number of the feedback results of the first user acquired for any category reaches M, where M is a positive integer greater than one, the threshold parameter corresponding to the category may be adjusted according to the M feedback results.
In an embodiment of the present disclosure, a verification set may be first constructed using images corresponding to M feedback results, and then threshold parameters corresponding to the category may be adjusted according to the constructed verification set. The specific value of M can be determined according to actual needs.
Taking the category a as an example, when the number of the feedback results of the first user acquired for the category a reaches M, a verification set may be constructed using images corresponding to M feedback results, and if the value of M is 100 (the numerical value is merely an example), a verification set may be constructed using 100 images corresponding to the 100 feedback results, and then, the threshold parameter corresponding to the category a may be adjusted according to the constructed verification set.
Compared with the prior art, the method has the advantages that the function of user feedback is added, so that the closed loop of the customized dynamic classification algorithm is realized.
In an embodiment of the present disclosure, the manner of adjusting the threshold parameter corresponding to the category a by using the constructed verification set may be: traversing all possible values of the threshold parameter, and aiming at each traversed value, respectively performing the following processing: determining the precision (precision) and recall (recall) corresponding to the category a when images in the verification set are classified according to the value, and determining a balanced average (F1-score) according to the obtained precision and recall; and taking the value corresponding to the maximum F1-score as the value of the adjusted threshold parameter. Among them, F1-score can also be called F1 score.
The value of the threshold parameter is usually between 0 and 1, and in practical application, all possible values of the threshold parameter, such as 0.01, 0.02, 0.03 and the like, can be traversed according to a preset step length (such as 0.01) and in a sequence from small to large.
For each traversed value, taking a value x as an example, the value x can be used to replace the original value of the threshold parameter corresponding to the category a, and the image classification model and the threshold parameter corresponding to each category can be used to classify each image in the verification set, so as to obtain the classification result of each image. Then, for the category a, an accuracy rate and a recall rate can be respectively calculated, wherein the accuracy rate indicates how many of the samples predicted to be positive are real positive samples, and the recall rate indicates how many of the positive samples in the samples are predicted to be correct. For example, taking the category "a" as "cat", assuming that the verification set includes 100 images, the accuracy rate indicates how many images in the images classified as "cat" actually belong to the category "cat", and the recall rate indicates how many images in the images classified as "cat" in the verification set are classified as "cat". Since the correct classification result for each image in the validation set is known, the above accuracy and recall can be conveniently calculated. According to the calculated accuracy and recall, F1-score can be further calculated, namely F1-score corresponding to the value x is taken.
In general terms,
where precision represents precision and recall represents recall.
According to the mode, F1-score corresponding to different traversed values can be obtained respectively, and further the value corresponding to the maximum value of F1-score can be used as the value of the adjusted threshold parameter. For example, if F1-score corresponding to the value x is the maximum, the value x may be used as the value of the threshold parameter corresponding to the adjusted category a.
In another embodiment of the present disclosure, the manner of adjusting the threshold parameter corresponding to the category a by using the constructed verification set may also be: and determining the value of the threshold parameter corresponding to the category a for classifying the images in the verification set when the accuracy rate or the recall rate corresponding to the category a is a preset value, and taking the value as the value of the adjusted threshold parameter.
The specific value of the predetermined value may be determined according to actual needs, and taking the accuracy rate as an example, the corresponding predetermined value may be 0.9. Accordingly, it can be determined that the value of the threshold parameter corresponding to the category a is obtained when the accuracy rate is 0.9, and the value can be used as the value of the adjusted threshold parameter.
In different scenarios, the accuracy rate may be more biased, or the recall rate may be more biased, i.e., the predetermined value may be set based on the accuracy rate, or the predetermined value may be set based on the recall rate.
It can be seen that two ways of adjusting the threshold parameter are provided in the scheme of the disclosure, one of the ways can be arbitrarily selected for use according to actual needs, which is very flexible and convenient, and each way has a better effect.
With the above introduction, fig. 2 is a flowchart of a second embodiment of the method for improving image classification accuracy according to the present disclosure. As shown in fig. 2, the following detailed implementation is included.
In step 201, an image classification model is obtained by training with a training set.
In step 202, a threshold parameter corresponding to each category is determined through the verification set.
The specific implementation of step 201-step 202 is prior art.
In step 203, an image input by a first user is obtained, and the image is classified by using an image classification model and a threshold parameter corresponding to each category, so as to obtain a classification result including at least one category.
In step 204, a feedback result of the first user for any category in the classification results is obtained, where the feedback result is a positive or negative feedback result.
For example, an interactive interface may be generated and displayed for the classification result, and a feedback result of the user may be obtained through the interactive interface.
In step 205, when the number of the feedback results of the first user acquired for any category reaches M, where M is a positive integer greater than one, the processing is performed in the manner shown in steps 206 to 207.
In step 206, a verification set is constructed using the images corresponding to the M feedback results.
In step 207, the threshold parameter corresponding to the category is adjusted according to the constructed verification set.
The adjustment method can be as follows: traversing all possible values of the threshold parameter, and aiming at each traversed value, respectively performing the following processing: determining the accuracy and recall rate corresponding to the type when images in the verification set are classified according to the value, and determining F1-score according to the obtained accuracy and recall rate; and taking the value corresponding to the maximum F1-score as the value of the adjusted threshold parameter.
Alternatively, the adjustment method may be: and determining the value of a threshold parameter used for classifying the images in the verification set when the accuracy rate or the recall rate corresponding to the category is a preset value, and taking the value as the value of the adjusted threshold parameter.
In addition, in practical applications, if necessary, the process shown in steps 205 to 207 may be repeatedly executed subsequently, and the threshold parameter corresponding to a certain category may be continuously optimized according to practical situations.
It is noted that while for simplicity of explanation, the foregoing method embodiments are described as a series of acts, those skilled in the art will appreciate that the present disclosure is not limited by the order of acts, as some steps may, in accordance with the present disclosure, occur in other orders and concurrently. Further, those skilled in the art should also appreciate that the embodiments described in the specification are preferred embodiments and that the acts and modules referred to are not necessarily required for the disclosure. In addition, for parts which are not described in detail in a certain embodiment, reference may be made to relevant descriptions in other embodiments.
The above is a description of embodiments of the method, and the embodiments of the apparatus are further described below.
Fig. 3 is a schematic structural diagram illustrating a composition of an embodiment 300 of the apparatus for improving image classification accuracy according to the present disclosure. As shown in fig. 3, includes: a classification module 301, an acquisition module 302, and an adjustment module 303.
The classification module 301 is configured to obtain an image input by a first user, and classify the image by using an image classification model obtained through pre-training and a predetermined threshold parameter corresponding to each category to obtain a classification result including at least one category.
An obtaining module 302, configured to obtain a feedback result of the first user for any category in the classification results, where the feedback result is a positive or negative feedback result.
The adjusting module 303 is configured to, when the number of the feedback results of the first user obtained for any category reaches M, where M is a positive integer greater than one, adjust the threshold parameter corresponding to the category according to the M feedback results.
The image classification model can be obtained by pre-training, namely, the image classification model can be obtained by pre-training with a training set, model parameters are determined, and then, the threshold parameters corresponding to each category can be determined through a verification set, wherein the threshold parameters are suitable for all users.
On this basis, for any image input by the first user, the classification module 301 may first classify the image by using the image classification model and the threshold parameter corresponding to each category, so as to obtain a classification result including at least one category.
Thereafter, the obtaining module 302 may obtain a feedback result of the first user for any category in the classification results, where the feedback result is a positive or negative feedback result.
For example, for the classification result, the obtaining module 302 may generate and display an interactive interface, and obtain the feedback result of the user through the interactive interface.
Further, the adjusting module 303 may adjust the threshold parameter corresponding to any one of the categories according to the M feedback results when the number of the feedback results of the first user obtained for the category reaches M, where M is a positive integer greater than one.
For example, the adjusting module 303 may first construct a verification set by using the images corresponding to the M feedback results, and then may adjust the threshold parameter corresponding to the category according to the constructed verification set. The specific value of M can be determined according to actual needs.
Specifically, the adjusting module 303 may traverse all possible values of the threshold parameter, and perform the following processing for each traversed value: determining the accuracy and recall rate corresponding to the type when images in the verification set are classified according to the value, and determining F1-score according to the obtained accuracy and recall rate; and taking the value corresponding to the maximum F1-score as the value of the adjusted threshold parameter.
Alternatively, the adjusting module 303 may determine a value of a threshold parameter used for classifying the images in the verification set when the accuracy rate or the recall rate corresponding to the category is a predetermined value, and use the value as the value of the adjusted threshold parameter.
For a specific work flow of the apparatus embodiment shown in fig. 3, reference is made to the related description in the foregoing method embodiment, and details are not repeated.
In a word, by adopting the scheme of the embodiment of the device disclosed by the disclosure, the threshold parameter corresponding to any category can be adjusted according to the feedback result of the user through interaction with the user, so that the accuracy of the classification result is improved, and the threshold parameter corresponding to the user is adjusted by taking the user as the adjustment granularity, so that personalized processing and the like aiming at different users are realized.
The scheme disclosed by the invention can be applied to the field of artificial intelligence, in particular to the fields of deep learning, computer vision and the like. Artificial intelligence is a subject for studying a computer to simulate some thinking processes and intelligent behaviors (such as learning, reasoning, thinking, planning and the like) of a human, and has a hardware technology and a software technology, the artificial intelligence hardware technology generally comprises technologies such as a sensor, a special artificial intelligence chip, cloud computing, distributed storage, big data processing and the like, and the artificial intelligence software technology mainly comprises a computer vision technology, a voice recognition technology, a natural language processing technology, machine learning/deep learning, a big data processing technology, a knowledge graph technology and the like.
The present disclosure also provides an electronic device, a readable storage medium, and a computer program product according to embodiments of the present disclosure.
FIG. 4 shows a schematic block diagram of an example electronic device 400 that may be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital assistants, cellular telephones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the disclosure described and/or claimed herein.
As shown in fig. 4, the apparatus 400 includes a computing unit 401 that can perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM)402 or a computer program loaded from a storage unit 408 into a Random Access Memory (RAM) 403. In the RAM 403, various programs and data required for the operation of the device 400 can also be stored. The computing unit 401, ROM 402, and RAM 403 are connected to each other via a bus 404. An input/output (I/O) interface 405 is also connected to bus 404.
A number of components in device 400 are connected to I/O interface 405, including: an input unit 406 such as a keyboard, a mouse, or the like; an output unit 407 such as various types of displays, speakers, and the like; a storage unit 408 such as a magnetic disk, optical disk, or the like; and a communication unit 409 such as a network card, modem, wireless communication transceiver, etc. The communication unit 409 allows the device 400 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunication networks.
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuitry, Field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (ASSPs), system on a chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.
Program code for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the program codes, when executed by the processor or controller, cause the functions/operations specified in the flowchart and/or block diagram to be performed. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.
In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), Wide Area Networks (WANs), and the Internet.
The computer system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so as to solve the defects of high management difficulty and weak service expansibility in the traditional physical host and Virtual Private Server (VPS). The server may also be a server of a distributed system, or a server incorporating a blockchain. Cloud computing refers to accessing an elastically extensible shared physical or virtual resource pool through a network, resources can include servers, operating systems, networks, software, applications, storage devices and the like, a technical system for deploying and managing the resources in a self-service mode as required can be achieved, and efficient and powerful data processing capacity can be provided for technical applications and model training of artificial intelligence, block chains and the like through a cloud computing technology.
It should be understood that various forms of the flows shown above may be used, with steps reordered, added, or deleted. For example, the steps described in the present disclosure may be executed in parallel, sequentially, or in different orders, as long as the desired results of the technical solutions disclosed in the present disclosure can be achieved, and the present disclosure is not limited herein.
The above detailed description should not be construed as limiting the scope of the disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made in accordance with design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present disclosure should be included in the scope of protection of the present disclosure.
Claims (13)
1. A method of improving image classification accuracy, comprising:
acquiring an image input by a first user, and classifying the image by using an image classification model obtained by pre-training and a predetermined threshold parameter corresponding to each category to obtain a classification result comprising at least one category;
obtaining a feedback result of the first user for any category in the classification results, wherein the feedback result is a positive feedback result or a negative feedback result;
when the number of the feedback results of the first user acquired for any category reaches M, the M is a positive integer larger than one, and the threshold parameter corresponding to the category is adjusted according to the M feedback results.
2. The method of claim 1, wherein the obtaining feedback results of the first user for any of the categories of the classification results comprises:
and generating and displaying an interactive interface aiming at the classification result, and acquiring the feedback result through the interactive interface.
3. The method according to claim 1 or 2, wherein the adjusting the threshold parameter corresponding to the category according to the M feedback results comprises:
constructing a verification set by using images corresponding to the M feedback results;
and adjusting the threshold parameter corresponding to the category according to the verification set.
4. The method of claim 3, wherein said adjusting threshold parameters corresponding to said categories according to said verification set comprises:
traversing all possible values of the threshold parameter, and aiming at each traversed value, respectively performing the following processing: determining the accuracy and the recall rate corresponding to the categories when the images in the verification set are classified according to the sampling values, and determining a balanced average F1-score according to the accuracy and the recall rate;
and taking the value corresponding to the maximum F1-score as the value of the adjusted threshold parameter.
5. The method of claim 3, wherein said adjusting threshold parameters corresponding to said categories according to said verification set comprises:
and determining the value of the threshold parameter used for image classification of the images in the verification set when the accuracy rate or the recall rate corresponding to the category is a preset value, and taking the value as the value of the adjusted threshold parameter.
6. An apparatus for improving image classification accuracy, comprising: the device comprises a classification module, an acquisition module and an adjustment module;
the classification module is used for acquiring an image input by a first user, classifying the image by utilizing an image classification model obtained by pre-training and a predetermined threshold parameter corresponding to each category to obtain a classification result comprising at least one category;
the obtaining module is configured to obtain a feedback result of the first user for any category in the classification results, where the feedback result is a positive or negative feedback result;
the adjusting module is configured to, when the number of the feedback results of the first user obtained for any category reaches M, where M is a positive integer greater than one, adjust the threshold parameter corresponding to the category according to the M feedback results.
7. The apparatus of claim 6, wherein,
and the acquisition module generates and displays an interactive interface aiming at the classification result, and acquires the feedback result through the interactive interface.
8. The apparatus of claim 6 or 7,
and the adjusting module constructs a verification set by using the images corresponding to the M feedback results, and adjusts the threshold parameters corresponding to the categories according to the verification set.
9. The apparatus of claim 8, wherein,
the adjusting module traverses all possible values of the threshold parameter, and respectively performs the following processing for each traversed value: determining the accuracy and the recall rate corresponding to the categories when the images in the verification set are classified according to the sampling values, and determining a balanced average F1-score according to the accuracy and the recall rate; and taking the value corresponding to the maximum F1-score as the value of the adjusted threshold parameter.
10. The apparatus of claim 8, wherein,
and the adjusting module determines the value of the threshold parameter used for image classification of the images in the verification set when the accuracy rate or the recall rate corresponding to the category is a preset value, and takes the value as the adjusted value of the threshold parameter.
11. An electronic device, comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-5.
12. A non-transitory computer readable storage medium having stored thereon computer instructions for causing a computer to perform the method of any one of claims 1-5.
13. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110719033.1A CN113408632A (en) | 2021-06-28 | 2021-06-28 | Method and device for improving image classification accuracy, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110719033.1A CN113408632A (en) | 2021-06-28 | 2021-06-28 | Method and device for improving image classification accuracy, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113408632A true CN113408632A (en) | 2021-09-17 |
Family
ID=77679834
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110719033.1A Pending CN113408632A (en) | 2021-06-28 | 2021-06-28 | Method and device for improving image classification accuracy, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113408632A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114330221A (en) * | 2021-11-22 | 2022-04-12 | 北京百度网讯科技有限公司 | Score board implementation method, score board, electronic device and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107729946A (en) * | 2017-10-26 | 2018-02-23 | 广东欧珀移动通信有限公司 | Picture classification method, device, terminal and storage medium |
| CN109543713A (en) * | 2018-10-16 | 2019-03-29 | 北京奇艺世纪科技有限公司 | The modification method and device of training set |
| CN109767001A (en) * | 2019-01-07 | 2019-05-17 | 深圳增强现实技术有限公司 | Construction method, device and the mobile terminal of neural network model |
| CN111507467A (en) * | 2019-01-31 | 2020-08-07 | 北京奇虎科技有限公司 | Neural network model training method and device, computer equipment and storage medium |
| CN112529083A (en) * | 2020-12-15 | 2021-03-19 | 深圳开立生物医疗科技股份有限公司 | Ultrasonic scanning method, device, equipment and storage medium |
| CN112580400A (en) * | 2019-09-29 | 2021-03-30 | 华为技术有限公司 | Image optimization method and electronic equipment |
| CN112669321A (en) * | 2021-03-22 | 2021-04-16 | 常州微亿智造科技有限公司 | Sand blasting unevenness detection method based on feature extraction and algorithm classification |
-
2021
- 2021-06-28 CN CN202110719033.1A patent/CN113408632A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107729946A (en) * | 2017-10-26 | 2018-02-23 | 广东欧珀移动通信有限公司 | Picture classification method, device, terminal and storage medium |
| CN109543713A (en) * | 2018-10-16 | 2019-03-29 | 北京奇艺世纪科技有限公司 | The modification method and device of training set |
| CN109767001A (en) * | 2019-01-07 | 2019-05-17 | 深圳增强现实技术有限公司 | Construction method, device and the mobile terminal of neural network model |
| CN111507467A (en) * | 2019-01-31 | 2020-08-07 | 北京奇虎科技有限公司 | Neural network model training method and device, computer equipment and storage medium |
| CN112580400A (en) * | 2019-09-29 | 2021-03-30 | 华为技术有限公司 | Image optimization method and electronic equipment |
| CN112529083A (en) * | 2020-12-15 | 2021-03-19 | 深圳开立生物医疗科技股份有限公司 | Ultrasonic scanning method, device, equipment and storage medium |
| CN112669321A (en) * | 2021-03-22 | 2021-04-16 | 常州微亿智造科技有限公司 | Sand blasting unevenness detection method based on feature extraction and algorithm classification |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114330221A (en) * | 2021-11-22 | 2022-04-12 | 北京百度网讯科技有限公司 | Score board implementation method, score board, electronic device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20190087490A1 (en) | Text classification method and apparatus | |
| CN114494784A (en) | Deep learning model training method, image processing method and object recognition method | |
| CN113657483A (en) | Model training method, target detection method, device, equipment and storage medium | |
| CN114882321A (en) | Deep learning model training method, target object detection method and device | |
| CN116580223A (en) | Data processing and model fine tuning method and device, electronic equipment and storage medium | |
| CN114861059A (en) | Resource recommendation method and device, electronic equipment and storage medium | |
| CN114360027A (en) | Training method and device for feature extraction network and electronic equipment | |
| CN113657538A (en) | Model training method, data classification method, device, equipment, storage medium and product | |
| CN113408632A (en) | Method and device for improving image classification accuracy, electronic equipment and storage medium | |
| CN114926322B (en) | Image generation method, device, electronic equipment and storage medium | |
| CN113554062B (en) | Training method, device and storage medium for multi-classification model | |
| CN114817476A (en) | Language model training method and device, electronic equipment and storage medium | |
| CN114707638A (en) | Model training method, model training device, object recognition method, object recognition device, object recognition medium and product | |
| CN114120416A (en) | Model training method and device, electronic equipment and medium | |
| CN113806541A (en) | Emotion classification method and emotion classification model training method and device | |
| CN113408304A (en) | Text translation method and device, electronic equipment and storage medium | |
| CN113361575A (en) | Model training method and device and electronic equipment | |
| CN114067415A (en) | Regression model training method, object evaluation method, device, equipment and medium | |
| CN113361621A (en) | Method and apparatus for training a model | |
| CN113240082A (en) | Transfer learning method and device | |
| CN112632999A (en) | Named entity recognition model obtaining method, named entity recognition device and named entity recognition medium | |
| JP7551970B2 (en) | Artificial intelligence model updating method, apparatus, electronic device and storage medium | |
| CN113553407B (en) | Event tracing method and device, electronic equipment and storage medium | |
| US20240338530A1 (en) | Generative dialog model training method and apparatus as well as generative dialog implementing method and apparatus | |
| US20220383626A1 (en) | Image processing method, model training method, relevant devices and electronic device |
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 |