1. A method for determining an image, comprising:

acquiring a target image, wherein the target image comprises a target object;

analyzing the target image by using a target model to determine a target integrity of a target object included in the target image, wherein the target model is trained by machine learning by using a plurality of sets of training data, and each set of data in the plurality of sets of training data includes: the image containing the object and the labeling integrity of the object contained in the image;

and determining a target sub-image from the target image based on a target integrity threshold and the target integrity, wherein the target sub-image is a sub-image of the target image.

2. The method of claim 1, wherein prior to determining a target sub-image from the target image based on a target integrity threshold and the target integrity, the method further comprises:

in response to a completeness threshold indication instruction, determining the target completeness threshold as the target completeness threshold indicating the completeness indicated by the instruction.

3. The method of claim 1, wherein analyzing the target image using a target model to determine a target integrity of a target object included in the target image comprises:

analyzing the target image by using a target model to determine the main body integrity of the target object and the sub-integrity of each part of the target object;

determining the subject integrity and the sub-integrity as the target integrity.

4. The method of claim 1, wherein prior to analyzing the target image using a target model, the method further comprises:

acquiring a plurality of images comprising objects and the labeling integrity of a target labeling object to the objects in each image, wherein the labeling integrity comprises the labeling main body integrity of the objects and the labeling sub-integrity of each part of the objects;

determining the training integrity of the object included in each image based on the labeling integrity and the target labeling object;

determining the training completeness of each image and an object included in each image as the training data;

and training an initial model by using the training data to obtain the target model.

5. The method of claim 4, wherein determining a training completeness of an object included in the each image based on the annotation completeness and the target annotation object comprises:

determining the labeling weight of each labeling object included in the target labeling object;

determining the labeling integrity of each labeling object to the object included in each image;

determining the training completeness by the sum of the products of the labeling weight and the labeling completeness of each labeling object.

6. The method of claim 5, wherein determining the annotation weight for each annotation object included in the target annotation object comprises:

determining the historical labeling accuracy of each labeled object;

determining historical average labeling accuracy of all the labeled objects;

determining the ratio of the historical annotation accuracy to the historical average annotation accuracy;

determining an average weight of each of the labeled objects;

determining the product of the average weight and the ratio as the labeling weight of each of the labeled objects.

7. The method of claim 1, wherein after analyzing the target image using a target model to determine a target integrity of a target object included in the target image, the method further comprises:

and outputting the target integrity.

8. The method of claim 1, wherein determining a target sub-image from the target image based on a target integrity threshold and the target integrity comprises:

determining a completeness threshold interval based on the target completeness threshold;

and determining the sub-image with the target integrity in the threshold integrity interval included in the target image as the target sub-image.

9. An apparatus for determining an image, comprising:

the device comprises an acquisition module, a processing module and a display module, wherein the acquisition module is used for acquiring a target image, and the target image comprises a target object;

an analysis module, configured to analyze the target image using a target model to determine a target integrity of a target object included in the target image, where the target model is trained through machine learning using multiple sets of training data, and each set of data in the multiple sets of training data includes: the image containing the object and the labeling integrity of the object contained in the image;

the determining module is used for determining a target sub-image from the target image based on a target integrity threshold and the target integrity, wherein the target sub-image is a sub-image of the target image.

10. A computer-readable storage medium, in which a computer program is stored, which computer program, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 8.

11. An electronic device comprising a memory and a processor, wherein the memory has stored therein a computer program, and wherein the processor is arranged to execute the computer program to perform the method of any of claims 1 to 8.

Background

With the continuous development of the technology, the application of the artificial intelligence technology in life is more and more extensive, the target-based intelligent analysis (such as target attribute analysis, target re-identification and target feature ratio and the like) is an important part of the artificial intelligence application, and the accuracy of the intelligent analysis has an important influence on the user experience. In order to improve the accuracy of target analysis, targets need to be screened. The integrity of the target is an important consideration dimension during screening, but the integrity of the target in the related art can only be judged whether the target is complete, i.e. the target is divided into complete and incomplete, the incomplete degree of the image is not judged, and under the condition of partial severe conditions, if the target is incomplete in the whole process, the target cannot be further screened. In addition, when determining whether the target is complete, the target integrity threshold can only be determined in advance, and the determined image cannot be dynamically adjusted during use.

Therefore, the problem that the determined image cannot satisfy the dynamic scene because the image can be determined only based on the predetermined integrity threshold exists in the related art.

In view of the above problems in the related art, no effective solution has been proposed.

Disclosure of Invention

The embodiment of the invention provides an image determining method, an image determining device, a storage medium and an electronic device, which are used for at least solving the problem that the determined image cannot meet the dynamic scene because the image can only be determined based on a predetermined complete threshold value in the related art.

According to an embodiment of the present invention, there is provided a method of determining an image, including: acquiring a target image, wherein the target image comprises a target object; analyzing the target image by using a target model to determine a target integrity of a target object included in the target image, wherein the target model is trained by machine learning by using a plurality of sets of training data, and each set of data in the plurality of sets of training data includes: the image containing the object and the labeling integrity of the object contained in the image; and determining a target sub-image from the target image based on a target integrity threshold and the target integrity, wherein the target sub-image is a sub-image of the target image.

According to another embodiment of the present invention, there is provided an image determination apparatus including: the device comprises an acquisition module, a processing module and a display module, wherein the acquisition module is used for acquiring a target image, and the target image comprises a target object; an analysis module, configured to analyze the target image using a target model to determine a target integrity of a target object included in the target image, where the target model is trained through machine learning using multiple sets of training data, and each set of data in the multiple sets of training data includes: the image containing the object and the labeling integrity of the object contained in the image; the determining module is used for determining a target sub-image from the target image based on a target integrity threshold and the target integrity, wherein the target sub-image is a sub-image of the target image.

According to yet another embodiment of the invention, there is also provided a computer-readable storage medium having a computer program stored therein, wherein the computer program, when executed by a processor, implements the steps of the method as set forth in any of the above.

According to yet another embodiment of the present invention, there is also provided an electronic device, including a memory in which a computer program is stored and a processor configured to execute the computer program to perform the steps in any of the above method embodiments.

According to the invention, after the target image is obtained, the target image is analyzed by using the target model to determine the target integrity of the target object included in the target image, and the target sub-image is determined from the target image according to the target integrity threshold and the target integrity.

Drawings

Fig. 1 is a block diagram of a hardware configuration of a mobile terminal of an image determination method according to an embodiment of the present invention;

FIG. 2 is a flow chart of a method of determining an image according to an embodiment of the invention;

FIG. 3 is a diagram of a target model structure according to an exemplary embodiment of the present invention;

FIG. 4 is a schematic flow chart illustrating the process of determining a target sub-image from a target image based on a target integrity threshold and a target integrity according to an exemplary embodiment of the present invention;

FIG. 5 is a diagram of an apparatus for determining an image according to an embodiment of the present invention;

fig. 6 is a diagram of an image determination apparatus according to an embodiment of the present invention.

Detailed Description

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings in conjunction with the embodiments.

It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order.

The method embodiments provided in the embodiments of the present application may be executed in a mobile terminal, a computer terminal, or a similar computing device. Taking an example of the method running on a mobile terminal, fig. 1 is a block diagram of a hardware structure of the mobile terminal of the method for determining an image according to the embodiment of the present invention. As shown in fig. 1, the mobile terminal may include one or more (only one shown in fig. 1) processors 102 (the processor 102 may include, but is not limited to, a processing device such as a microprocessor MCU or a programmable logic device FPGA), and a memory 104 for storing data, wherein the mobile terminal may further include a transmission device 106 for communication functions and an input-output device 108. It will be understood by those skilled in the art that the structure shown in fig. 1 is only an illustration, and does not limit the structure of the mobile terminal. For example, the mobile terminal may also include more or fewer components than shown in FIG. 1, or have a different configuration than shown in FIG. 1.

The memory 104 may be used to store a computer program, for example, a software program and a module of application software, such as a computer program corresponding to the image determination method in the embodiment of the present invention, and the processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, so as to implement the method described above. The memory 104 may include high speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include memory located remotely from the processor 102, which may be connected to the mobile terminal over a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The transmission device 106 is used to receive or transmit data via a network. Specific examples of the network described above may include a wireless network provided by a communication provider of the mobile terminal. In one example, the transmission device 106 includes a Network adapter (NIC) that can be connected to other Network devices through a base station to communicate with the internet. In one example, the transmission device 106 may be a Radio Frequency (RF) module, which is used to communicate with the internet in a wireless manner.

In the present embodiment, an image determining method is provided, and fig. 2 is a flowchart of an image determining method according to an embodiment of the present invention, as shown in fig. 2, the flowchart includes the following steps:

step S202, acquiring a target image, wherein the target image comprises a target object;

step S204, analyzing the target image by using a target model to determine a target integrity of a target object included in the target image, wherein the target model is trained by machine learning using a plurality of sets of training data, and each set of data in the plurality of sets of training data includes: the image containing the object and the labeling integrity of the object contained in the image;

step S206, determining a target sub-image from the target image based on a target integrity threshold and the target integrity, wherein the target sub-image is a sub-image of the target image.

In the above-described embodiments, the target image may be an image including a target object, and the target object may be a person, a vehicle, an animal, or the like.

In the above embodiment, the target image may be an image obtained through preprocessing. Firstly, an input original image is obtained, then the original image is preprocessed and scaled to the size matched with the input of a target model, the size is determined before training, and the model is kept consistent during training and deployment.

In the above embodiment, the target model may be implemented by using a deep convolutional neural network, where a preprocessed target image is input, feature extraction is performed through a series of convolutional operations, and then a full link layer is connected, and a regression score (i.e., target integrity) of integrity of the target predicted by the model is regressed and output. The target model structure diagram can refer to fig. 3, and as shown in fig. 3, the integrity score (target integrity) of the target image can be obtained after the target image sequentially passes through the input layer, the feature extraction layer and the integrity score regression layer (i.e., the full connection layer).

In the above embodiment, the target image may include a plurality of images, that is, the integrity of the object in the image may be determined in batches. After the target integrity is determined, the target sub-image can be determined from the target image according to the target integrity and the target integrity threshold. The target integrity threshold may be a threshold dynamically adjusted according to an application scenario. When a plurality of images are included in the target image, the target sub-image may be a partial image included in the plurality of images or an image of a partial region of one image included in the target image. When the target image is an image, the target sub-image may be an image of a part of the area in the target image. For example, when the target image is one sheet, and the target image is an image including a human face, the target sub-image may be an image of human eyes, or an image of human mouth.

Optionally, the main body of the above steps may be a background processor, or other devices with similar processing capabilities, and may also be a machine integrated with at least an image acquisition device and a data processing device, where the image acquisition device may include a graphics acquisition module such as a camera, and the data processing device may include a terminal such as a computer and a mobile phone, but is not limited thereto.

According to the invention, after the target image is obtained, the target image is analyzed by using the target model to determine the target integrity of the target object included in the target image, and the target sub-image is determined from the target image according to the target integrity threshold and the target integrity.

In an exemplary embodiment, before determining a target sub-image from the target image based on a target integrity threshold and the target integrity, the method further comprises: in response to a completeness threshold indication instruction, determining a target completeness indicated by the target completeness threshold indication instruction as the target completeness threshold. In this embodiment, a user may input an integrity threshold indication instruction through a control interface or a control window, and after receiving the integrity threshold indication instruction, determine a target integrity indicated by the integrity threshold indication instruction as a target integrity threshold, and determine a target sub-image according to the target integrity threshold and the target integrity.

In the above embodiment, the target integrity threshold indication instruction may further include an output instruction, for example, output an image greater than the target integrity threshold, and output an image smaller than the target integrity threshold. The integrity threshold indication instructions may also indicate a plurality of integrity thresholds. When the integrity threshold indication instruction indicates multiple integrity thresholds, for example, when two integrity thresholds are included in the target integrity threshold indication instruction, the output instruction may be to output an image that is between the two thresholds.

Referring to fig. 4, a schematic flow chart of determining a target sub-image from a target image based on a target integrity threshold and a target integrity may be shown, as shown in fig. 4, after the target integrity is determined, whether an integrity score threshold (corresponding to the integrity threshold indication instruction) is input may be determined, and if not, the target integrity is output. And if so, determining the target sub-image according to the target integrity threshold.

In the above embodiment, the integrity score may be directly output or a threshold integrity score may be input, and the target is output according to the threshold integrity. The integrity score threshold value can be determined by a downstream user module according to actual task requirements, a higher integrity score threshold value is set for task requirements with high integrity requirements, and a lower integrity score threshold value is set for task requirements with low integrity requirements, so that complete or incomplete targets can be screened out according to different actual task requirements.

It should be noted that the target integrity threshold instruction is an instruction input by the user, and therefore, the user can input the target integrity threshold according to the application scene, so as to achieve the effect of determining the image meeting the dynamic use scene.

In one exemplary embodiment, analyzing the target image using a target model to determine a target integrity of a target object included in the target image comprises: analyzing the target image by using a target model to determine the main body integrity of the target object and the sub-integrity of each part of the target object; determining the subject integrity and the sub-integrity as the target integrity. In this embodiment, the target image is analyzed by using the target model, and the determined target integrity includes a main body integrity of the target object and sub-integrity of each part of the target object. That is, each set of data in the training data includes an image of the object and the subject training integrity and each partial training integrity of the object included in the image. After the target model is trained, the target image is input, and the main image quality attribute (corresponding to the main integrity) and the image quality attribute of each sub-part (corresponding to the sub-integrity) can be output through network model operation.

In an exemplary embodiment, before analyzing the target image using the target model, the method further comprises: acquiring a plurality of images comprising objects and the labeling integrity of a target labeling object to the objects in each image, wherein the labeling integrity comprises the labeling main body integrity of the objects and the labeling sub-integrity of each part of the objects; determining the training integrity of the object included in each image based on the labeling integrity and the target labeling object; determining the training completeness of each image and an object included in each image as the training data; and training an initial model by using the training data to obtain the target model. In this embodiment, the model is trained by inputting pre-prepared data pairs, and a data pair is composed of an image with a fixed size and corresponding label information. And in the training process, comparing the output result of the model with the label result, calculating a loss value based on a loss function, then training by a gradient descent method, and updating the parameters of the model.

In the above embodiment, the target annotation object may annotate the object included in each image to obtain the annotation integrity of each object. Before labeling can be performed, labeling information of integrity scores (corresponding to the target integrity) of the target object can be obtained first, different integrity scores are set for different visibility degrees according to actual requirements, for example, 100% visible, 75% visible, 50% visible, 25% visible, and several classes of invisible targets are set, and the corresponding integrity scores are 1,0.75,0.5,0.25, and 0 respectively. It should be noted that the corresponding relationship is only an exemplary illustration, and the invention is not limited thereto.

After the corresponding relation between the visibility degree and the integrity degree is determined, the images in each image can be labeled according to the corresponding relation to obtain the labeling integrity degree, then the training integrity degree of the object included in each image is determined according to the labeling integrity degree and the target labeling correspondence, and each image and the training integrity degree of the object included in each image are determined as training data. And training the initial model according to the training data. The initial block structure diagram can also be seen in fig. 3.

In one exemplary embodiment, determining the training completeness of the object included in each image based on the annotation completeness and the target annotation object comprises: determining the labeling weight of each labeling object included in the target labeling object; determining the labeling integrity of each labeling object to the object included in each image; determining the training completeness by the sum of the products of the labeling weight and the labeling completeness of each labeling object. In this embodiment, the labeling may be performed in a multi-user independent labeling manner, each target labeled object is labeled with a score, after all the labeled scores are obtained, the labeled data is subjected to anomaly detection, the abnormal labeled data is eliminated, then the effective labeled score is subjected to weighted average according to the labeling level of the target labeled object, and the result of the weighted average is used as the target integrity. That is, the target integrity can be expressed as S (∑ (Pi × Si)). Wherein, S is the final labeling score (i.e. training integrity), Pi is the weight of the ith labeling object, and Si is the labeling score (i.e. labeling integrity) of the ith labeling object.

In one exemplary embodiment, an annotation weight package is determined for each annotation object included in the target annotation objectComprises the following steps: determining the historical labeling accuracy of each labeled object; determining historical average labeling accuracy of all the labeled objects; determining the ratio of the historical annotation accuracy to the historical average annotation accuracy; determining an average weight of each of the labeled objects; determining the product of the average weight and the ratio as the labeling weight of each of the labeled objects. In this embodiment, the labeling weight may be determined according to the historical labeling accuracy, the average weight is 1/n, and the weight value of the label is the product of the average weight and the ratio of the label exceeding the average weight. Wherein the labeling weight can be expressed asn is the number of labels of the valid labels.

In an exemplary embodiment, after analyzing the target image using a target model to determine a target integrity of a target object included in the target image, the method further includes: and outputting the target integrity. In this embodiment, after the target model analyzes the target image, if the input output threshold is not received, the main body integrity and the sub-integrity of the object included in the target image may be directly output.

In an exemplary embodiment, determining a target sub-image from the target image based on a target integrity threshold and the target integrity comprises: determining a completeness threshold interval based on the target completeness threshold; and determining the sub-image with the target integrity in the threshold integrity interval included in the target image as the target sub-image. In this embodiment, the integrity threshold interval may be determined according to the target integrity threshold, and the sub-image in which the target integrity is in the integrity interval included in the target image is determined as the target sub-image. After the target sub-image is determined, the target sub-image can be output.

The following describes a method for determining an image in accordance with an embodiment:

fig. 5 is a diagram of an apparatus for determining an image according to an embodiment of the present invention, as shown in fig. 5, the apparatus including:

the image preprocessing module is used for firstly acquiring an input original image, then preprocessing the original image and scaling the original image to a size matched with the input of a subsequent target integrity score regression model, wherein the size is determined before training, and the consistency of the model during training and deployment is kept.

A target integrity score regression module: the target integrity degree regression module is realized by adopting a deep convolution neural network, a preprocessed target image is input, feature extraction is carried out through a series of convolution operations, then a full connection layer is connected, and the integrity regression score of the target predicted by the model is regressed and output.

A target integrity post-processing module: the target integrity post-processing module firstly obtains the returned image integrity score, then can select to directly output the integrity score or select to input an integrity score threshold, and outputs whether the target is complete or not according to the threshold. The integrity score threshold value can be determined by a downstream user module according to actual task requirements, a higher integrity score threshold value is set for task requirements with high integrity requirements, and a lower integrity score threshold value is set for task requirements with low integrity requirements, so that complete or incomplete targets can be screened out according to different actual task requirements.

In the foregoing embodiment, regression is introduced into the target integrity judgment, so that the integrity score of the target can be obtained, and the degree of the target being blocked or truncated can be described more accurately. The target integrity threshold value can be dynamically adjusted during use based on the regressed target integrity score, so that complete or incomplete targets can be screened out according to different actual task requirements, deployment flexibility is improved, and model training workload is reduced.

Through the above description of the embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but the former is a better implementation mode in many cases. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which is stored in a storage medium (e.g., ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal device (e.g., a mobile phone, a computer, a server, or a network device) to execute the method according to the embodiments of the present invention.

In this embodiment, an image output apparatus is further provided, and the apparatus is used to implement the above embodiments and preferred embodiments, and the description of the apparatus is omitted. As used below, the term "module" may be a combination of software and/or hardware that implements a predetermined function. Although the means described in the embodiments below are preferably implemented in software, an implementation in hardware, or a combination of software and hardware is also possible and contemplated.

Fig. 6 is a diagram of an image determination apparatus according to an embodiment of the present invention, as shown in fig. 6, the apparatus including:

an obtaining module 62, configured to obtain a target image, where the target image includes a target object;

an analysis module 64, configured to analyze the target image using a target model to determine a target integrity of a target object included in the target image, where the target model is trained through machine learning using multiple sets of training data, and each set of the multiple sets of training data includes: the image containing the object and the labeling integrity of the object contained in the image;

a determining module 66, configured to determine a target sub-image from the target image based on a target integrity threshold and the target integrity, where the target sub-image is a sub-image of the target image.

The obtaining module 62 corresponds to the image preprocessing module, the analyzing module 64 corresponds to the target integrity score regression module, and the determining module 66 corresponds to the target integrity post-processing module.

In an exemplary embodiment, the apparatus may be configured to determine, in response to a completeness threshold indication instruction, a completeness indicated by the target completeness threshold indication instruction as the target completeness threshold before determining a target sub-image from the target image based on a target completeness threshold and the target completeness.

In an exemplary embodiment, analysis module 64 may implement the analysis of the target image using a target model to determine a target integrity of a target object included in the target image by: analyzing the target image by using a target model to determine the main body integrity of the target object and the sub-integrity of each part of the target object; determining the subject integrity and the sub-integrity as the target integrity.

In an exemplary embodiment, the apparatus may be configured to, before analyzing the target image using a target model, obtain a plurality of images including an object and an annotation integrity of the target annotation object for the object included in each image, where the annotation integrity includes an annotation main body integrity of the object and an annotation sub-integrity of each portion of the object; determining the training integrity of the object included in each image based on the labeling integrity and the target labeling object; determining the training completeness of each image and an object included in each image as the training data; and training an initial model by using the training data to obtain the target model.

In an exemplary embodiment, the apparatus may enable determining a training completeness of an object included in the each image based on the annotation completeness and the target annotation object by: determining the labeling weight of each labeling object included in the target labeling object; determining the labeling integrity of each labeling object to the object included in each image; determining the training completeness by the sum of the products of the labeling weight and the labeling completeness of each labeling object.

In an exemplary embodiment, the apparatus may determine the annotation weight of each annotation object included in the target annotation object by: determining the historical labeling accuracy of each labeled object; determining historical average labeling accuracy of all the labeled objects; determining the ratio of the historical annotation accuracy to the historical average annotation accuracy; determining an average weight of each of the labeled objects; determining the product of the average weight and the ratio as the labeling weight of each of the labeled objects.

In an exemplary embodiment, the apparatus may be further configured to output the target integrity after analyzing the target image using a target model to determine the target integrity of a target object included in the target image.

In an exemplary embodiment, the determining module 66 may determine the target sub-image from the target image based on the target integrity threshold and the target integrity by: determining a completeness threshold interval based on the target completeness threshold; determining a sub-image included in the target image and having the target integrity within the integrity threshold interval as the target sub-image; and outputting the target sub-image.

It should be noted that, the above modules may be implemented by software or hardware, and for the latter, the following may be implemented, but not limited to: the modules are all positioned in the same processor; alternatively, the modules are respectively located in different processors in any combination.

Embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method as set forth in any of the above.

In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to: various media capable of storing computer programs, such as a usb disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic disk, or an optical disk.

Embodiments of the present invention also provide an electronic device comprising a memory having a computer program stored therein and a processor arranged to run the computer program to perform the steps of any of the above method embodiments.

In an exemplary embodiment, the electronic apparatus may further include a transmission device and an input/output device, wherein the transmission device is connected to the processor, and the input/output device is connected to the processor.

For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and exemplary embodiments, and details of this embodiment are not repeated herein.

It will be apparent to those skilled in the art that the various modules or steps of the invention described above may be implemented using a general purpose computing device, they may be centralized on a single computing device or distributed across a network of computing devices, and they may be implemented using program code executable by the computing devices, such that they may be stored in a memory device and executed by the computing device, and in some cases, the steps shown or described may be performed in an order different than that described herein, or they may be separately fabricated into various integrated circuit modules, or multiple ones of them may be fabricated into a single integrated circuit module. Thus, the present invention is not limited to any specific combination of hardware and software.

The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the principle of the present invention should be included in the protection scope of the present invention.