KR102543833B1 - Apparatus for object detection based on artificial intelligence learning and method thereof - Google Patents

Abstract

The present disclosure relates to an object detection method for improving the extraction accuracy of properties of an object captured in an image and an apparatus thereof. The object detection method according to some embodiments of the present disclosure is performed by a computing device may include the step of: identifying a target object in a first frame -the first frame is a random frame included in an image representing the target object-; analyzing the first trajectory of the target object represented in the image; and extracting an optimal frame for extracting properties of the target object from the image based on the analysis of the first trajectory.

Description

Object detection method and apparatus based on artificial intelligence learning

The present disclosure relates to an object detection method and apparatus for improving extraction accuracy (e.g., object recognition) of attributes of an object captured in an image.

Object detection refers to an operation of performing classification and localization on an object photographed in an image. Here, classification may mean a task of classifying objects photographed in an image into a specific category (e.g., human, dog, car, etc.), and localization may mean a task of segmenting the location of an object in an image. .

Based on the recent development of artificial intelligence technology, a technology for dynamically detecting an object photographed in an image without human intervention has been introduced. However, since an image subject to object detection has a certain time period, there is no clear criterion as to which frame among a plurality of frames included in the image should be selected to perform object detection. For example, when an arbitrary frame included in an image is selected, an object photographed in the image is blurred and is not suitable for object attribute extraction.

Therefore, a technique for improving extraction accuracy of attributes of an object photographed in an image is required.

Korean Patent Publication No. 10-2020-0031232 (published on September 28, 2021)

A technical problem to be solved through some embodiments of the present disclosure is to provide an apparatus for improving extraction accuracy of attributes of an object captured in an image and a method performed by the apparatus.

Another technical problem to be solved through some embodiments of the present disclosure is to provide an apparatus for extracting an optimal frame from an image having a certain time period and a method performed by the apparatus.

The technical problems of the present disclosure are not limited to the above-mentioned technical problems, and other technical problems not mentioned will be clearly understood by those skilled in the art from the description below.

In order to solve the above technical problem, an object detection method according to some embodiments of the present disclosure is a method performed by a computing device, the step of identifying a target object in a first frame - the first frame, the target object Any frame included in an image in which an object is represented - Analyzing a first trajectory of the target object represented in the image and extracting an attribute of the target object based on the analysis of the first trajectory Extracting an optimal frame from the image may be included.

In some embodiments, the method may further include extracting attributes of the target object from the optimum frame.

In some embodiments, analyzing the first trajectory may include analyzing the first trajectory based on a movement of a feature point of the target object represented in the image. Here, the extracting of the optimal frame from the image may include extracting a frame having a minimum moving speed of the feature point.

In some embodiments, the identifying of the target object includes identifying an auxiliary object represented in the first frame, and the analyzing of the first trajectory includes the identification of the auxiliary object represented in the image. Analyzing a second trajectory, and extracting the optimal frame, based on the analysis of the first trajectory and the second trajectory, the image in which the first trajectory and the second trajectory are overlapped may include extracting the optimum frame for extracting the attributes of the target object and the auxiliary object in a single frame in a partial section of . Here, in the partial section of the image in which the first trajectory and the second trajectory are overlapped, the step of extracting the optimal frame may include a first moving speed of a first feature point and a second moving speed of a second feature point and extracting a frame having a minimum sum of , wherein the first feature point may be a feature point of the target object, and the second feature point may be a feature point of the auxiliary object. In addition, in the partial section of the image in which the first trajectory and the second trajectory are overlapped, the step of extracting the optimal frame includes a frame in which at least one of the first moving speed and the second moving speed is equal to or greater than a reference value. The method may further include excluding the optimum frame from extraction.

An object detection apparatus according to some embodiments of the present disclosure includes a processor, a network interface, a memory, and a computer program loaded into the memory and executed by the processor, wherein the computer program, in a first frame An instruction for identifying a target object, wherein the first frame is an arbitrary frame included in an image in which the target object is represented, an instruction for analyzing a first trajectory of the target object represented in the image, and the An instruction for extracting an optimal frame for extracting the attribute of the target object from the image based on the analysis of the first trajectory may be included.

1 illustrates an exemplary environment to which an object detection apparatus according to some embodiments of the present disclosure may be applied.
2 is an exemplary flowchart illustrating an object detection method according to some embodiments of the present disclosure.
3 to 5 are exemplary diagrams for explaining an object identification operation (S100) and an object trajectory analysis operation (S200) described with reference to FIG. 2, and are exemplary diagrams showing arbitrary frames included in an image. am.
FIG. 6 is an exemplary diagram showing a result of the trajectory analysis operation ( S200 ) described with reference to FIGS. 3 to 5 .
7 depicts an example computing device on which devices and/or systems in accordance with various embodiments of the present disclosure may be implemented.

Hereinafter, preferred embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. Advantages and features of the present disclosure, and methods of achieving them, will become clear with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, the technical idea of the present disclosure is not limited to the following embodiments and may be implemented in various different forms, and only the following embodiments complete the technical idea of the present disclosure, and in the technical field to which the present disclosure belongs. It is provided to completely inform those skilled in the art of the scope of the present disclosure, and the technical spirit of the present disclosure is only defined by the scope of the claims.

In adding reference numerals to components of each drawing, it should be noted that the same components have the same numerals as much as possible, even if they are displayed on different drawings. In addition, in describing the present disclosure, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present disclosure, the detailed description will be omitted.

Unless otherwise defined, all terms (including technical and scientific terms) used in this specification may be used with meanings commonly understood by those of ordinary skill in the art to which this disclosure belongs. In addition, terms defined in commonly used dictionaries are not interpreted ideally or excessively unless explicitly specifically defined. Terminology used herein is for describing the embodiments and is not intended to limit the present disclosure. In this specification, singular forms also include plural forms unless specifically stated otherwise in a phrase.

Also, terms such as first, second, A, B, (a), and (b) may be used in describing the components of the present disclosure. These terms are only used to distinguish the component from other components, and the nature, order, or order of the corresponding component is not limited by the term. When an element is described as being “connected,” “coupled to,” or “connected” to another element, that element is or may be directly connected to the other element, but there is another element between the elements. It will be understood that elements may be “connected”, “coupled” or “connected”.

As used herein, "comprises" and/or "comprising" means that a stated component, step, operation, and/or element is the presence of one or more other components, steps, operations, and/or elements. or do not rule out additions.

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

1 illustrates an exemplary environment to which an object detection apparatus 100 according to some embodiments of the present disclosure may be applied. 1 shows an example in which one user terminal 200 and one camera 300 are connected to the object detection device 100 through a network, but this is only for convenience of understanding, and the user terminal 200 ) and the number of cameras 300 may vary.

Meanwhile, FIG. 1 only illustrates a preferred embodiment for achieving the object of the present disclosure, and some components may be added or deleted as necessary. In addition, it should be noted that components of the exemplary environment shown in FIG. 1 represent functionally differentiated functional elements, and a plurality of components may be implemented in a form integrated with each other in an actual physical environment. For example, the object detection device 100 and the user terminal 200 may be implemented in different logic types within the same computing device. Also, for another example, the object detection device 100 and the camera 300 may be implemented in the form of different logics in the same computing device.

Hereinafter, each component shown in FIG. 1 will be described in more detail.

First, the object detection device 100 may receive an image captured by the camera 300 . Here, the object detection apparatus 100 may extract an optimal frame for extracting an attribute of an object based on an image received from the camera 300 . In this case, the object detection apparatus 100 may extract the attribute of the object represented in the extracted optimum frame.

Also, the object detection apparatus 100 may transmit results of various operations based on the image received from the camera 300 to the user terminal 200 . Here, it should be noted that the results provided to the user terminal 200 can be transformed into user-friendly forms on the user interface.

In order to exclude redundant descriptions, various operations performed by the object detection apparatus 100 will be described in more detail later with reference to the drawings below in FIG. 2 .

The object detection device 100 described above may be implemented with one or more computing devices. For example, all functions of the object detection device 100 may be implemented in a single computing device. As another example, the first function of the object detection device 100 may be implemented in a first computing device, and the second function may be implemented in a second computing device. Here, the aforementioned computing device may be a notebook, a desktop, or a laptop, but is not limited thereto and may include all types of devices equipped with a computing function. However, in an environment in which the object detection device 100 needs to detect multiple objects in conjunction with a plurality of user terminals 200 and a plurality of cameras 300, the object detection device 100 is implemented as a high-performance server-class computing device. It may be desirable to be Reference is made to FIG. 7 for an example of such a computing device.

Next, the user terminal 200 may view results provided by the object detection device 100 . For example, the user can view the optimal frame of the image captured by the camera 300 through the user terminal 200 . Also, for another example, the user may view properties of objects detected through the user terminal 200 . Here, the user terminal 200 may be installed with a web browser or a dedicated application for browsing results provided by the object detection device 100 . Also, here, the user terminal 200 may be, for example, any one of a desktop, a workstation, a laptop, a tablet, and a smart phone. It is not limited to one example, and all types of devices equipped with computing functions may be the user terminal 200 .

Next, the camera 300 may be any one of all types of cameras for capturing a user's region of interest, and for example, a Closed-Circuit Television (CCTV) may be the camera 300 .

In some embodiments, the object detection device 100 , the user terminal 200 and the camera 300 may communicate through a network. The aforementioned network is implemented as all kinds of wired/wireless networks such as a Local Area Network (LAN), a Wide Area Network (WAN), a mobile radio communication network, and Wibro (Wireless Broadband Internet). It can be.

An exemplary environment to which the object detection apparatus 100 according to some embodiments of the present disclosure may be applied has been described with reference to FIG. 1 so far. Hereinafter, methods according to various embodiments of the present disclosure will be described in detail.

Each step of the above methods may be performed by a computing device. In other words, each step of the methods may be implemented as one or more instructions executed by a processor of a computing device. All of the steps involved in these methods could be performed by one physical computing device, but first steps of the method are performed by a first computing device and second steps of the method are performed by a second computing device. It could be. Hereinafter, description will be continued on the assumption that each step of the above methods is performed by the object detection apparatus 100 illustrated in FIG. 1 . However, for convenience of explanation, the description of the operating subject of each step included in the above-described methods may be omitted.

2 is an exemplary flowchart illustrating an object detection method according to some embodiments of the present disclosure. Hereinafter, with reference to FIG. 2 , operations shown in FIG. 2 will be described in more detail.

In step S100, a target object may be identified in the frame. Here, the frame may mean any frame included in the image. For example, a target object may be identified in an N-th frame (N being a natural number) of an image, where N may be appropriately changed by a user according to an environment in which an object detection device is applied. In addition, an initial frame in which a target object is to be identified may be selected according to various methods.

However, if it is an initial frame for identifying a target object, it may be preferable to mean a frame representing the target object among a plurality of frames included in the image. Here, in order to select a frame representing a target object from among a plurality of frames included in an image, various known image recognition techniques may be applied to the present disclosure. For example, a technique for identifying a moving object (i.e., a target object) through a rate of change in the position of feature points (e.g., feature points of objects fixedly present in an area captured by a camera) that is fixedly expressed in an image is the present invention. Although it may be applied to the disclosure, it should be noted that the scope of the disclosure is not limited according to this example.

If an initial frame is selected according to the above operations, a target object represented in the initial frame may be identified. More specifically, a category of the target object may be classified or a location where the target object is expressed in a frame may be specified. Here, various known image recognition techniques may be applied to the present disclosure to identify the target object represented in the frame. In particular, an artificial intelligence-based image recognition technology that classifies a category of a target object by pre-learning an image for each category of object may be applied to the present disclosure.

For a more detailed explanation related to the identification operation of the target object, it will be described with reference to FIG. 3 . 3 is an exemplary diagram illustrating a first frame 10a included in an image. A first object 20 and a second object 30 are represented in the first frame 10a of FIG. 3 . Here, according to the object identification operation, the interest boxes 23 and 33 may be displayed in the first frame 10a to correspond to the boundaries of the respective objects 20 and 30 represented in the first frame 10a. . In particular, if the classified category of each of the objects 20 and 30 is a person, the special interest boxes 21 and 31 representing the face of each of the objects 20 and 30 may be further displayed in the first frame 10a. there is.

4 and 5 are exemplary diagrams illustrating frames 10b and 10c sequentially following the first frame 10a of FIG. 3 . Specifically, FIG. 4 shows a second frame 10b following the first frame 10a of FIG. 3, and FIG. 5 shows a third frame 10c following the second frame 10b of FIG. are doing Since detailed descriptions related to FIGS. 4 and 5 can be understood by referring to the description of FIG. 3 , the description thereof will be omitted.

Referring to step S100, in some embodiments more than one object may be identified in the frame. Specifically, a target object may be identified in the frame, and an auxiliary object may be identified in the frame. Here, the auxiliary object may mean an object distinct from the target object.

If a target object is identified in an arbitrary frame included in an image according to the above-described operations, the target object may be tracked through time-sequential analysis. Again referring to FIG. 2 , operations to be described later will be described.

In step S200, the trajectory of the target object represented in the image may be analyzed. Here, the trajectory of the target object may be understood as a virtual line tracing an area where the target object is located in each of the frames included in the time-sequential images. At this time, in order to analyze the trajectory of the target object expressed in the image, various known image recognition techniques may be applied to the present disclosure.

In relation to step S200, in some embodiments, analyzing the trajectory of the target object may include analyzing the trajectory of the target object based on the movement of feature points of the target object represented in the image. Here, the feature point may mean a point corresponding to the location of the target object in the frame, and a point located within the interest boxes 23 and 33 of FIG. 3 may be understood as a point corresponding to the location of the target object. For example, a feature point may refer to a point located within the interest boxes 23 and 33 of FIG. 3 . At this time, various methods for selecting points inside the boxes of interest 23 and 33 may be applied to the present disclosure, and the midpoint of the boxes of interest 23 and 33 may be determined as a feature point, but the scope of the present disclosure is limited to this example. is not limited.

For another example, the feature point may mean a point located within the special interest boxes 21 and 31 of FIG. 3 . In this case, a point located within the special interest boxes 21 and 31 can be understood as a point corresponding to the location of the target object, and is located in an area that the user may be more interested in (e.g., a human face area). can be understood as points. According to this example, when points located in the special interest boxes 21 and 31 are used as feature points, accuracy of extracting attributes of the target object according to operations to be described later can be further improved.

Regarding step S200, in some other embodiments, trajectories of two or more objects represented in a frame may be analyzed respectively. Specifically, the trajectory of the target object may be analyzed, and the trajectory of the auxiliary object may also be analyzed.

In order to explain the above-described trajectory analysis in more detail, it will be described with reference to FIG. 6 . FIG. 6 shows an example of analyzing the trajectories of the objects 20 and 30 shown in FIGS. 3 to 5 . The analysis frame 10d of FIG. 6 is not a frame included in the image, but can be understood as a frame obtained by analyzing the frame included in the image.

The analysis frame 10d of FIG. 6 shows an interest box 23a corresponding to the first object, a feature point 27 located in the interest box 23a, and an identifier 29 of the first object. An interest box 33a corresponding to two objects, a feature point 37 located within the interest box 33a, and an identifier 39 of the second object are shown. In particular, the analysis frame 10d also shows the first trajectory 25 corresponding to the first object and the second trajectory 35 corresponding to the second object.

According to the above-described operations, the trajectory of the target object represented in the image may be analyzed. Again referring to FIG. 2 , operations to be described later will be described.

In step S300, based on the analysis of the trajectory, an optimal frame for extracting attributes of the target object may be extracted from the image. Here, the optimal frame may mean a frame most suitable for extracting the property of the target object among a plurality of frames included in the image. In this case, the property of the target object means a feature that can be extracted through image analysis. For example, the face image of the target object may be included in the property of the target object, but in this example, the present disclosure The scope of is not limited.

In relation to step S300 , in some embodiments, extracting an optimal frame from an image may include extracting a frame having a minimum moving speed of a feature point. That is, since the frame with the minimum movement speed of the feature point is likely to be a frame in which the target object is not blurred, the frame with the minimum movement speed of the feature point may be a suitable frame as an optimal frame.

Regarding step S300, in some other embodiments, extracting the optimal frame may include extracting the optimal frame for extracting attributes of the target object and the auxiliary object in a single frame. Specifically, one optimal frame suitable for extracting each attribute may be extracted instead of individually trajectory analyzing the target object and the auxiliary object, which are the user's interest, and individually extract two optimal frames. In this case, it is possible to save computing resources according to individual operations on the object of interest (e.g., target object, auxiliary object), save computing resources more effectively in an environment with a large number of objects of interest, and optimize frames suitable for attribute extraction. can be extracted.

Here, the step of extracting an optimal frame from a partial section of an image in which the first trajectory of the target object and the second trajectory of the auxiliary object are overlapped represents the first movement speed of the first feature point and the movement speed of the second feature point. A step of extracting a frame having a minimum sum may be included. In this case, the first feature point may be a feature point of the target object, and the second feature point may be a feature point of the auxiliary object. That is, since the frame in which the sum of the moving speeds of feature points is minimum is likely to be a frame in which both the target object and the auxiliary object are not blurred, the frame in which the sum of the moving speeds of feature points is minimal is suitable as an optimal frame. can be

Also, here, the step of extracting an optimal frame from a partial section of an image in which the first trajectory of the target object and the second trajectory of the auxiliary object are overlapped is the first movement speed (i.e., the movement speed of the first feature point of the target object) ) and the second moving speed (i.e., the moving speed of the second feature point of the auxiliary object) may include excluding frames having a reference value or higher from extraction of the optimal frame. In this case, noise (i.e., a phenomenon in which a specific object is blurred) that may be generated according to the sum operation of the moving speed may be excluded.

In relation to step S300, in some other embodiments, extracting an optimal frame may include excluding frames having a gradient of a trajectory along a moving direction of a feature point equal to or greater than a threshold value from extraction of the optimal frame. That is, a frame in which the gradient of the trajectory according to the moving direction of the feature point is greater than or equal to the threshold value is highly likely to be a frame in which the target object is blurred, such as when the target object rapidly changes the moving direction, and thus can be excluded from extraction of the optimal frame. .

An optimal frame included in an image may be extracted according to the operations described above. Again referring to FIG. 2 , operations to be described later will be described.

In step S400, attributes of the target object may be extracted from the optimum frame. Here, since the optimum frame may be a frame of an image in which the property of the target object is most easily extracted, the accuracy of object property extraction may be improved by extracting the property of the target object from the optimum frame.

So far, an object detection method according to some embodiments of the present disclosure has been described with reference to FIG. 2 and exemplary drawings. According to the methods described above, a frame (i.e., optimal frame) in an image most suitable for extracting an attribute of an object may be selected. In particular, in the case of extracting and analyzing the attributes of two or more objects, the user can effectively extract the attributes of the objects according to a method with improved extraction accuracy while saving computing resources.

Hereinafter, an exemplary computing device 1500 capable of implementing an object detection device according to some embodiments of the present disclosure will be described in more detail with reference to FIG. 7 .

The computing device 1500 includes one or more processors 1510, a bus 1550, a communication interface 1570, a memory 1530 for loading a computer program 1591 executed by the processor 1510, and a computer A storage 1590 for storing the program 1591 may be included. However, only components related to the embodiment of the present disclosure are shown in FIG. 7 . Accordingly, those skilled in the art to which the present disclosure belongs may know that other general-purpose components may be further included in addition to the components shown in FIG. 7 .

The processor 1510 controls the overall operation of each component of the computing device 1500 . The processor 1510 includes a central processing unit (CPU), a micro processor unit (MPU), a micro controller unit (MCU), a graphics processing unit (GPU), or any type of processor well known in the art of the present disclosure. It can be. Also, the processor 1510 may perform an operation for at least one application or program for executing a method according to embodiments of the present disclosure. Computing device 1500 may include one or more processors.

Memory 1530 stores various data, commands and/or information. Memory 1530 may load one or more programs 1591 from storage 1590 to execute a method according to embodiments of the present disclosure. The memory 1530 may be implemented as a volatile memory such as RAM, but the technical scope of the present disclosure is not limited thereto.

The bus 1550 provides a communication function between components of the computing device 1500 . The bus 1550 may be implemented as various types of buses such as an address bus, a data bus, and a control bus.

The communication interface 1570 supports wired and wireless Internet communication of the computing device 1500 . Also, the communication interface 1570 may support various communication methods other than Internet communication. To this end, the communication interface 1570 may include a communication module well known in the art of the present disclosure.

According to some embodiments, communication interface 1570 may be omitted.

The storage 1590 may non-temporarily store one or more programs 1591 and various data described above.

The storage 1590 may be non-volatile memory, such as read only memory (ROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, etc., a hard disk, a removable disk, or as is well known in the art. It may be configured to include any known type of computer-readable recording medium.

Computer program 1591 may include one or more instructions that when loaded into memory 1530 cause processor 1510 to perform methods/operations in accordance with various embodiments of the present disclosure. That is, the processor 1510 may perform methods/operations according to various embodiments of the present disclosure by executing one or more of the above-described instructions.

So far, various embodiments of the present disclosure and effects according to the embodiments have been described with reference to FIGS. 1 to 7 . Effects according to the technical idea of the present disclosure are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description of the specification.

The technical idea of the present disclosure described with reference to FIGS. 1 to 7 so far may be implemented as computer readable code on a computer readable medium. The aforementioned computer-readable recording medium is, for example, a removable recording medium (CD, DVD, Blu-ray disc, USB storage device, removable hard disk) or a fixed recording medium (ROM, RAM, computer-equipped hard disk). can be A computer program recorded on a computer-readable recording medium may be transmitted to another computing device through a network such as the Internet, installed in the other computing device, and thereby used in the other computing device.

In the above, even though all the components constituting the embodiments of the present disclosure have been described as being combined or operated as one, the technical idea of the present disclosure is not necessarily limited to these embodiments. That is, within the scope of the purpose of the present disclosure, all of the components may be selectively combined with one or more to operate.

Although actions are shown in a particular order in the drawings, it should not be understood that the actions must be performed in the specific order shown or in a sequential order, or that all shown actions must be performed to obtain a desired result. In certain circumstances, multitasking and parallel processing may be advantageous. Moreover, the separation of the various components in the embodiments described above should not be understood as requiring such separation, and the described program components and systems may generally be integrated together into a single software product or packaged into multiple software products. It should be understood that there is

Although the embodiments of the present disclosure have been described with reference to the accompanying drawings, those skilled in the art may implement the present disclosure in other specific forms without changing the technical spirit or essential features. can understand that there is Therefore, the embodiments described above should be understood as illustrative in all respects and not limiting. The protection scope of the present disclosure should be construed by the following claims, and all technical ideas within the equivalent range should be construed as being included in the scope of rights of the technical ideas defined by the present disclosure.

Claims (8)

An object detection method performed by a computing device,
identifying a target object in a first frame, wherein the first frame is an arbitrary frame included in an image in which the target object is expressed;
analyzing a first trajectory of the target object represented in the image; and
Based on the analysis of the first trajectory, extracting an optimal frame for extracting the attribute of the target object from the image,
The step of identifying the target object,
Identifying an auxiliary object represented in the first frame, wherein the auxiliary object is an object distinct from the target object;
Analyzing the first trajectory,
analyzing the first trajectory based on movement of a first feature point corresponding to the target object expressed in the image; and
Analyzing a second trajectory of the auxiliary object represented in the image,
Analyzing the second trajectory,
Analyzing the second trajectory based on movement of a second feature point corresponding to the auxiliary object represented in the image;
Extracting the optimal frame,
Based on the analysis of the first trajectory and the second trajectory, in a partial section of the image in which the first trajectory and the second trajectory are overlapped, the properties of the target object and the auxiliary object are extracted in a single frame Extracting the optimal frame for
The step of extracting the optimal frame from a partial section of the image in which the first trajectory and the second trajectory are overlapped,
Extracting a frame in which the sum of a first moving speed of a first feature point and a second moving speed of a second feature point is minimum,
object detection method. According to claim 1,
Further comprising extracting attributes of the target object from the optimal frame,
object detection method. delete delete delete delete According to claim 1,
The step of extracting the optimal frame from a partial section of the image in which the first trajectory and the second trajectory are overlapped,
Excluding a frame in which at least one of the first moving speed and the second moving speed is equal to or greater than a reference value from the extraction of the optimal frame,
object detection method. processor;
network interface;
Memory; and
A computer program loaded into the memory and executed by the processor,
The computer program,
an instruction for identifying a target object in a first frame, wherein the first frame is an arbitrary frame included in an image in which the target object is expressed;
an instruction for analyzing a first trajectory of the target object expressed in the image; and
Based on the analysis of the first trajectory, an instruction for extracting an optimal frame for extracting the attribute of the target object from the image,
The instruction for identifying the target object,
An instruction for identifying an auxiliary object represented in the first frame, wherein the auxiliary object is an object distinct from the target object;
The instruction for analyzing the first trajectory,
an instruction for analyzing the first trajectory based on movement of a first feature point corresponding to the target object expressed in the image; and
An instruction for analyzing a second trajectory of the auxiliary object expressed in the image;
The instruction for analyzing the second trajectory,
An instruction for analyzing the second trajectory based on movement of a second feature point corresponding to the auxiliary object represented in the image;
The instruction for extracting the optimal frame,
Based on the analysis of the first trajectory and the second trajectory, in a partial section of the image in which the first trajectory and the second trajectory are overlapped, the properties of the target object and the auxiliary object are extracted in a single frame Includes instructions for extracting the optimal frame for
In a partial section of the image in which the first trajectory and the second trajectory are overlapped, the instruction for extracting the optimal frame,
Including an instruction for extracting a frame in which the sum of the first moving speed of the first feature point and the second moving speed of the second feature point is minimum.
object detection device.

Images