KR101323886B1 - Distributed-Processing-Based Object Tracking Apparatus and Method - Google Patents

Abstract

A distributed object-based object tracking method, comprising: receiving, by a second client device, a plurality of image frames including an object to be tracked from a first client device; Predicting, by a second client device, the position of the object in an image frame (“predictive image frame”) after the plurality of image frames based on at least one of the plurality of image frames; Receiving, by a second client device, a position and a contour of the object from a computing device detected in one of the plurality of image frames; And matching, by a second client device, the contour with the predicted position of the object and the detected position of the object to calculate a tracking position and a tracking contour of the object in the predictive image frame. A distributed tracking based object tracking method is provided.

Description

Distributed-Processing-Based Object Tracking Apparatus and Method

The present invention relates to an object tracking apparatus and a tracking method based on distributed processing. More particularly, the present invention provides distributed tracking that provides an optimal tracking result by distributing a position prediction and a contour detection of a tracking target in a client and a server. Process based object tracking device and tracking method.

The technology of tracking an object in a video or a plurality of image sequences is widely studied as an essential technology for video analysis in the fields of video recognition, security surveillance system, augmented reality, and video call. For example, U.S. Patent No. 7,239,719 relates to a method for automatically detecting a target from an image data and analyzing a motion. The method detects a motion in an input video to define a region of interest and sets up a level set contour. This paper suggests a technique for extracting the contour of the object of interest.

Recently, mobile terminals such as mobile phones and tablet PCs have become popular and the number of security surveillance cameras installed for security and store management has soared, leading to the proliferation of client devices capable of producing video contents. Without considering the ding environment, it was difficult to perform heavy (i.e., many resources on the client end) techniques such as contour tracking on various client devices.

According to an embodiment of the present invention, in a state in which one or more client devices and a computing device such as a server are connected through a network, the client device predicts a tracking result of the tracked object on the client device, and the computing device detects the tracked object and thus the prediction result. And it can provide a distributed object-based object tracking method and apparatus for calculating the final tracking results by combining the detection results.

According to an embodiment of the present invention, by distributing image data acquired from a client device in a client device and a computing device, a distributed processing-based object capable of implementing high quality image recognition technology that has been difficult to implement by a client device alone. A tracking method and apparatus can be provided.

In addition, according to an embodiment of the present invention, it is possible to provide a recording medium in which a program capable of performing the distributed processing based object tracking method is recorded.

In accordance with an embodiment of the present invention, there is provided a method for tracking a distributed object based object, the method comprising: receiving, by a second client device, a plurality of image frames including a target object from a first client device; Predicting, by a second client device, the position of the object in an image frame (“predictive image frame”) after the plurality of image frames based on at least one of the plurality of image frames; Receiving, by a second client device, a position and a contour of the object from a computing device detected in one of the plurality of image frames; And matching, by a second client device, the contour with the predicted position of the object and the detected position of the object to calculate a tracking position and a tracking contour of the object in the predictive image frame. A distributed tracking based object tracking method can be provided.

According to an embodiment of the present invention, there is provided a method for tracking a distributed object based object, comprising: acquiring, by a client device, a plurality of image frames including a target object; Determining whether the client device transmits at least some image frames of the plurality of image frames to a computing device, and when determining that the client device transmits the at least some image frames; Predicting, by the client device, the position of the object in an image frame (“predictive image frame”) after the plurality of image frames based on at least one of the plurality of image frames; Receiving, by the client device, a position and a contour of the object detected in any one of the some image frames from the computing device; And matching, by the client device, the contour with the predicted position of the object and the detected position of the object to calculate the tracking position and the tracking contour of the object in the predictive image frame. A distributed tracking based object tracking method can be provided.

According to an embodiment of the present invention, a distributed processing based object tracking device, comprising: a communication unit capable of transmitting and receiving data with an external image acquisition device and a computing device through a network; A position and a contour of the object in an image frame (“predictive image frame”) after the plurality of image frames, based on at least one of a plurality of image frames received from the image acquisition apparatus, including a tracking object. A predictable object predictor; And when the communication unit receives the position and the contour of the object detected in one of the plurality of image frames from the computing device, the tracking position and the tracking contour of the object in the predicted image frame. In order to calculate, it is possible to provide a distributed processing-based object tracking device comprising a; matching unit for matching the predicted position of the object and the detected position and the contour of the object.

According to an embodiment of the present invention, there is provided a distributed object-based tracking device, comprising: an image acquisition unit capable of acquiring a plurality of image frames including a tracking object; A communication unit capable of transmitting and receiving data with an external computing device through a network; A first determining unit determining whether to transmit at least some image frames of the plurality of image frames to a computing device; An object predictor configured to predict a position and a contour of the object in an image frame (“predictive image frame”) after the plurality of image frames based on at least one of the plurality of image frames; And a calculation unit configured to calculate a tracking position and a tracking contour of the object in the predicted image frame, based on the predicted object position and the contour. Can be.

According to an embodiment of the present invention, a recording medium having a computer readable program recorded thereon capable of performing the method can be provided.

According to an embodiment of the present invention, by distributing image data obtained from the client device in the client device and the computing device, it is possible to implement a high quality image recognition technology that has been difficult to implement by the client device alone.

According to an embodiment of the present invention, the tracking result optimized according to the performance of the client device, the performance of the computing device, and the network situation in tracking the object to be tracked may be provided to the user in various resolution images.

1 is a block diagram provided for explaining a distributed object-based object tracking apparatus according to a first embodiment of the present invention,
2 is a diagram schematically showing a method of matching an image and a contour according to an embodiment of the present invention;
3 is a flowchart for explaining a method for tracking objects based on distributed processing according to a first embodiment of the present invention;
4 is a block diagram provided to explain an apparatus for tracking distributed objects based on a second embodiment of the present invention;
5 is a flowchart illustrating a method for tracking a distributed object based object according to a second embodiment of the present invention;
6 is a block diagram provided to explain an apparatus for tracking distributed objects based on a third embodiment of the present invention;
7 is a flowchart illustrating a distributed processing based object tracking method according to a third embodiment of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The above and other objects, features, and advantages of the present invention will become more readily apparent from the following description of preferred embodiments with reference to the accompanying drawings. However, the present invention is not limited to the embodiments described herein but may be embodied in other forms. Rather, the embodiments disclosed herein are provided so that the disclosure can be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

In the present specification, that the first component is operated (or executed) on the second component means that the first component is operated (or executed) in an environment in which the second component is operated (or executed) or Or means in which the first component is operated (or executed) through interaction with the second component directly or indirectly.

When a component, device, or system is referred to as including a component made up of a program or software, even if not expressly stated, the component, device, or system may be a piece of hardware necessary for the program or software to execute or operate. For example, it may be understood to include a memory, a CPU, etc.) or other programs or software (eg, an operating system or a driver required to drive hardware).

It is also to be understood that unless otherwise stated in the implementation of any component, the component may be implemented in software, hardware, or any form of software and hardware.

The terminology used herein is for the purpose of illustrating embodiments and is not intended to be limiting of the present invention. In the present specification, the singular form includes plural forms unless otherwise specified in the specification. As used in the specification, 'comprises' and / or 'comprising' does not exclude the presence or addition of one or more other components in addition to the components mentioned.

Hereinafter, the present invention will be described in detail with reference to the drawings. In describing the following specific embodiments, various specific details are set forth in order to explain and understand the invention in more detail. However, those skilled in the art can understand that the present invention can be used without these various specific details. In some instances, it should be noted that portions of the invention that are well known in the description of the invention and are not significantly related to the invention do not describe confusion in describing the invention.

1 is a block diagram provided to explain a distributed tracking based object tracking apparatus according to a first embodiment of the present invention.

Referring to FIG. 1, an object tracking apparatus based on distributed processing according to the present invention may include a first client device 10, a computing device 20, and a second client device 30.

The first client device 10 (hereinafter also referred to as a “client”) is a device capable of capturing and transmitting an image of a tracking object. The first client 10 may be, for example, one of various image capturing means such as a closed circuit television (CCTV), a security camera, a surveillance camera, a security camera, and the like. As an alternative embodiment, the first client 10 may be a general portable digital camera or may be one of mobile devices (eg, portable telephones, tablet PCs, PDAs, etc.) comprising such digital camera means.

In the illustrated embodiment, the first client 10 includes an image acquisition unit 101, a determination unit 102, and a communication unit 105.

The image acquisition unit 101 is a means for receiving an image including a tracking object, and may include, for example, an image sensor such as a CCD sensor, a CMOS sensor, and a processing engine for processing an image received from the sensor. According to an embodiment, the image acquisition unit 101 may acquire an image in a frame unit, and may acquire 24 to 30 image frames per second, for example. In an alternative embodiment, the image acquisition unit 101 may include a function of importing an image already taken without directly photographing the image.

The determination unit 102 is a function module that determines whether to transmit at least some image frames of the plurality of image frames acquired by the image acquisition unit 101 to the computing device 20. For example, the image acquisition unit 101 determines whether all of the received image frames are transmitted to the computing device 20 or only some of them.

To this end, in one embodiment, the determination unit 102 performs the determination operation based on at least one of a predetermined time period, a resource of the computing device 20, and a state of the network 40. When the determination unit 102 determines based on a predetermined time period, for example, assuming that the image acquisition unit 101 receives 30 frames of images per second, the determination unit 102 determines all of these frames. 20 may be determined to be transmitted, or may be determined to transmit one frame every three frames (that is, ten frames per second). When the determination unit 102 determines based on the resource of the computing device 20 or the state of the network 40, the resource may be, for example, of the processor processing speed and available memory capacity of the computing device 20. There may be at least one.

In an embodiment, it is assumed that the first client 10 obtains and knows information about the resource of the computing device 20 in advance. In an alternative embodiment, the first client 10 may determine the first time when the determination of the determining unit 102 is necessary. The client 10 may request resource information from the computing device 20 to obtain this information. On the other hand, when the determination unit 102 determines based on the state of the network 40, the state of the network at this time may be, for example, information such as the type of the network, whether the connection state, and the transmission speed. .

The communication unit 105 The first client 10 is a function module that connects the first client 10 to an external device such as the computing device 20, the second client 30, and transmits and receives data through the network 40. The network 40 may include one or more communication networks of various types of wired or wireless communication networks such as a local area network (LAN), a wide area network (WAN), and an internet communication network. In alternative embodiments, two or more of the first client 10, the second client 30, and the computing device 20 may communicate with one another via one or more of wireless communication technologies, such as Bluetooth, WiFi, RFID, or the like. It may also be communicated by direct wired connection to each other according to the embodiment.

According to an embodiment, the communication unit 105 may further include an encoding unit that compresses the image frame by a compression algorithm such as MPEG4, and may transmit the image frame compressed by the encoding unit.

Each of the image acquisition unit 101, the determination unit 102, and the communication unit 105 described above may be implemented in software, hardware, or a combination thereof, and may be implemented in a firmware or an embedded system. In addition, although the determination unit 102 is illustrated as a separate function module in the embodiment of FIG. 1, the determination unit 102 may be included in the image acquisition unit 101 or the communication unit 105 according to the embodiment. That is, as soon as the image acquisition unit 101 acquires an image frame, it may determine whether to transmit the image frame to the computing device 20, or which image frames are transmitted immediately before transmitting to the external device through the communication unit 105. It can also be configured to determine whether or not to do so.

In the illustrated embodiment, the computing device 20 includes a communication unit 201 and an object detector 203.

The communication unit 201 is a function module for connecting the computing device 20 to an external device and transmitting / receiving data through a network 40. For example, the communication unit 201 may have the same function as the communication unit 105 of the first client 10. Can have According to an embodiment, when the communication unit 201 receives a compressed image frame, the communication unit 201 may further include a decoding unit for decompressing the compressed image frame.

The object detector 203 detects the contour of the object to be tracked in the image frame using a predetermined algorithm. As an algorithm for extracting a contour of an object, for example, an active contour model or a Kalman filtering algorithm may be used, and at least one of various existing contour extraction algorithms according to alternative embodiments. May be used.

In an embodiment, the object detector 203 may further include a function of detecting a position of the object in the image frame. Extracting the contour of an object in the image frame allows the user to immediately know the position of the object in the frame. Therefore, the position detection of the object can be performed simultaneously with the contour extraction. A separate detection algorithm may be used.

Meanwhile, in one embodiment, the computing device 20 may be a server device such as a web server. In addition, each of the communication unit 201 and the object detecting unit 203 described above may be implemented in software, hardware, or a combination thereof, or may be implemented in firmware or an embedded system.

In the illustrated embodiment, the second client 30 includes a communication unit 301, an object predictor 303, a matching unit 306, and an image output unit 307.

The communication unit 301 A functional module for transmitting and receiving data to and from the first client 10 and the computing device 20 through the network 40, for example, the communication unit 105 of the first client 10 or the communication unit 201 of the computing device 20. It can have the same function as).

The object predictor 303 is a function module that predicts a position of an object in a current image frame based on a previous image frame. That is, based on at least one of the plurality of image frames including the object to be tracked, the position of the object in the image frame after the plurality of image frames (hereinafter referred to as "predictive image frame") is predicted.

In one embodiment, the object predictor 303 uses a predictive equation model to predict the position of the object. For example, a feature based tracking algorithm that extracts feature points of an object within an image frame and predicts the object's movement based on the feature points, and template based tracking that uses an image frame as a template. ) Algorithm or Kalman filtering algorithm may be used, and various prediction algorithms may be used.

In an embodiment, the object predictor 303 may further include a function of predicting a contour of the object in the image frame. For example, when the feature-based tracking algorithm is used, the feature points of the object are extracted to predict the position of the object in the image frame, so that the contour of the object can be predicted to some extent through the connection of the feature points. However, alternative embodiments may use separate prediction algorithms for contour prediction of objects.

The matching unit 306 may detect the detection information detected by the object detecting unit 203 of the computing device 20 and the prediction information of the object predicted by the object predicting unit 303 of the second client 30. The tracking position and the tracking contour are calculated as the prediction result for the tracking object.

As an example of the matching method, the matching unit 306 may match the position and contour of the object detected by the object detector 203 and the position of the object predicted by the object predictor 303. In this case, the tracking position of the object is calculated by weighted sum of the position predicted by the object predictor 303 and the position detected by the object detector 203 with respect to the position of the object, and then the object detector A tracking contour can be calculated by combining the contour of the object detected at 203 with the tracking position.

The image output unit 307 receives a tracking result calculated by the matching unit 306, that is, an image frame in which the tracking position and the tracking contour of the tracking object are combined and displayed to the user. As an alternative embodiment, the display may be shown to the user directly on the display, or the image frame of the matching result may be transmitted and stored in a separate storage device (not shown).

The second client 30 can be any device with display means. For example, it may be various personal computers with a monitor, or may be a portable device such as a portable telephone, a tablet PC, a PDA, a digital camera, or the like. Furthermore, in one embodiment, the first client 10 and the second client 30 may be one device.

Meanwhile, each of the communication unit 301, the object prediction unit 303, the matching unit 306, and the image output unit 307 described above may be implemented by software, hardware, or a combination thereof according to the embodiment. It can be implemented as an embedded system.

FIG. 2 is a diagram schematically illustrating a method of matching an image and a contour according to an embodiment of the present invention, and exemplarily illustrates the steps of predicting and detecting an object in an image frame and matching the two.

Referring to FIG. 2, first, in step (a), an image frame F1 including the tracking object 100 is transmitted from the first client 10 to the computer device 20 and the second client 30, respectively. Assume In step (b), the object detector 203 of the computer device 20 detects the contour C1 by applying a contour detection algorithm to the image frame F1. In this case, the contour detection algorithm or another separate algorithm may be used to detect the position P1 of the object in the image frame F1.

Meanwhile, in this step (b), the object predictor 303 of the second client 30 may perform a prediction algorithm based on at least one or more previous image frames among the previous image frames including the image frame F1. The position of the object in the current image frame F2 is predicted (P2). That is, in the illustrated example, in the previous image frame F1, although the object 100 is located at the upper left, it is predicted to move to the position P2 at the lower right through the prediction algorithm. In this case, according to the exemplary embodiment, the contour C2 of the object in the current image frame F2 may also be predicted using the position prediction algorithm or another separate algorithm.

Then, the detection information of the object (that is, the detected contour C1 information, or the detected contour C1 and the position P1 information) and the prediction information (that is, the predicted position P2 information, or the prediction). Assuming that the position (P2) and the contour (C2) information) have been delivered to the matching unit 306 of the second client 30, in step (c), the matching unit 306 may detect these detection information and the prediction information. Match

In an embodiment, assuming that the matching unit 306 uses the detected contour C1 and position P1 information and the predicted position P2 information of the object, the detected position P1 and the predicted position ( P2) is matched by weighted sum and the position P3 is calculated as a tracking position in the current frame F2 as a resultant value, and the contour C1 is combined around the tracking position P3 to combine the current frame. The tracking contour C1 in (F2) is calculated.

In another alternative embodiment, assuming that matcher 306 uses the detected contour C1 and position P1 information and the predicted contour C2 and position P2 information of the object, detected The position P1 and the predicted position P2 are matched by weighted summation to calculate the position P3 as a tracking position as a tracking position in the current frame F2 as a resultant value, and to detect the detected contour C1 and the predicted position P2. The tracking contour in the current frame F2 may be calculated by combining the value obtained by weighted addition of the contour C2 to the tracking position P3.

As described above, according to the exemplary embodiment of the present invention, since the prediction operation and the detection operation for the object to be tracked are separated and executed in separate devices, the result values are matched with each other. Depending on the situation, you can provide optimized tracking results. For example, when the second client 30 is a mobile terminal such as a smartphone, it is very difficult to show the tracking result in real time to the user after executing the prediction and detection algorithms due to the limitation of the mobile terminal performance. In the embodiment, since the contour detection function is passed to the computing device 20 and only a relatively light predictive tracking algorithm needs to be executed in the mobile terminal, an optimized tracking result can be provided to the user in real time.

3 is a flowchart illustrating a method for tracking a distributed object based object according to the first embodiment of the present invention.

Referring to FIG. 3, first, an image acquisition unit of the first client 10 obtains a plurality of image frames including a tracking object. In one embodiment, the plurality of image frames may be, for example, image frames input in real time through CCTV or surveillance cameras.

In operation S303, the first client 10 determines whether to transmit at least some of the plurality of image frames to the computing device 20. In one embodiment, the determination is performed by the determination unit 102, and the determination may be performed based on at least one of a predetermined time period, a resource of the computing device 20, and a state of the network 40.

As an example of the determination according to the time period, the determination unit 102 may set how many frames per second to transmit according to a predetermined time period, and determine whether to transmit each image frame according to this setting value. The resource of the computing device 20 may be, for example, at least one of the processor processing speed and the available memory capacity of the computing device, and the state of the network may be, for example, the type of the network, whether it is connected, and the transmission speed. Information, and so on.

Thereafter, in operation S305, the first client 10 may transmit an image frame to the second client 30, and at the same time, transmit the image frame to the computing device 20 according to the determination result of operation S303. have. At this time, according to the embodiment, the image frame to be transmitted to the computing device 20 and the second client 30 may be compressed and transmitted.

If the image frame is transmitted to the computing device 20, in operation S309, the computing device 20 detects the contour of the object to be tracked in the image frame using a contour detection algorithm. According to an embodiment, the computing device 20 may detect not only the contour of the object but also the position in the frame of the object in this step S309. The computing device 20 then sends the contour (or detected contour and location) of the detected object to the second client 30.

Meanwhile, the second client 30 predicts the position of the tracking target object in the frame with respect to the image frame received from the first client 10 in step S305 (S313). In an embodiment, the predicting step may be performed by the object predictor 303, and predicts the position of the object in the current image frame based on at least one of the plurality of past image frames using a predetermined prediction algorithm. do. In an alternative embodiment, the object predictor 303 may also predict the contour of the object. That is, the method may further include predicting a contour of the object in the current image frame based on at least one of the plurality of past image frames.

Next, in step S317, the matching unit 306 of the second client device 30 matches the predicted position of the object and the detected position of the object with a contour, thereby tracking the position of the object in the predicted image frame. And the tracking contour. As an embodiment of the specific matching method, for example, the tracking position of the object is calculated by weighted summation of the predicted and detected positions of the object, and the detected contour is combined with the tracking cone of the object. Can calculate the tour.

In another alternative embodiment, the matching unit 306 weights the detected position and the predicted position to calculate the tracking position of the object, and adds the weighted sum of the detected contour and the predicted contour to the tracking position. It is also possible to calculate the tracking contour of the object by combining the values.

Meanwhile, not all image frames are transmitted to the computing device 20 due to the determination operation of step S303. Therefore, the image frames transmitted to the second client 30 in step S311 may also be part of the entire frame. will be. Therefore, in the matching step S317, a situation may occur in which a matching operation is not performed for some frames. In this case, for example, when the communication unit 301 receives the position and the contour of the detected object for the predetermined image frame from the computing device 20, the matching unit 306 receives the corresponding image received from the first client 10. If the communication unit 301 does not receive the detected position and contour of the object from the computing device 20 for a predetermined image frame, the matching unit 306 predicts the object. The position and the contour of the object predicted by the unit 303 are output as it is. In this case, the predicted position and the contour become the tracking position and the tracking contour.

Thereafter, in operation S319, the image output unit 307 receives the image frame in which the tracking position and the tracking contour of the object calculated by the matching unit 306 are combined, and displays the image frame to the user.

4 is a block diagram provided to explain an apparatus for tracking a distributed object based on a second embodiment of the present invention. Referring to FIG. 4, in comparison with FIG. 1, the second client 30 further includes a determination unit 305. The determination unit 305 determines whether to perform the operation of the matching unit 306 and controls it. In an embodiment, the determination unit 305 may control the matching unit 306 based on at least one of a resource and a predetermined time period of the second client 30.

Here, based on the resources of the second client, for example, to determine the execution of the operation of the matching unit 306 based on whether or not the resources are sufficient to execute the matching algorithm, the resource is the second It may include at least one of a processor processing speed and available memory capacity of the client 30. Also, based on a predetermined time period, the matching operation is performed every predetermined time. For example, the matching operation may be performed only for a predetermined number of image frames per second.

Therefore, by adding the determination unit 305 as described above, the determination unit 305 and the matching unit 306 may be collectively referred to as a calculation unit, and the calculation unit may operate by the determination unit 305 to operate the matching unit 306. Calculate the tracking position and tracking contour of the object while controlling it. That is, if the matching unit 306 is operated by the determining unit 305, the calculating unit calculates the tracking position and the tracking contour according to the execution result of the matching unit 306, and the matching unit by the determining unit 305. If 306 does not operate, the object predictor 303 outputs the position and contour of the predicted object as a result of the calculator.

FIG. 5 is a flowchart illustrating a method for tracking a distributed object based object according to the second embodiment of the present invention shown in FIG. 4. Compared to FIG. 3, a matching determination step (S515) according to the operation of the determination unit 305 is further added, and all remaining steps are the same.

That is, in step S515, the determination unit 305 may determine whether to go to the matching step S517 or directly output the prediction result in step S519 without matching.

If the determination result of the determination unit 305 determines to perform the matching, in the matching step (S517), the position of the object predicted by the object predictor 304 and the position and contour of the object detected by the object detector 203 Match If the determination unit 305 decides not to perform the matching step S517, only the predicted position of the object by the prediction unit 304 is output as the tracking result (step S519). That is, the predicted position and the contour of the object by the object predictor 304 are calculated and output as the tracking position and the tracking contour of the object.

6 is a block diagram provided to explain an apparatus for tracking distributed objects based on a third embodiment of the present invention. Referring to FIG. 6, in comparison with FIG. 4, the first client 10 further includes an image preprocessor 103 and an image encoder 104.

The image preprocessor 103 converts a resolution of an image frame to be transmitted to the computing device 20 and / or the second client 30. In order to rapidly transmit image data through the network 40 and to calculate the contour of the object in the computing device 20, it is desirable to lower the resolution of the image. Therefore, in the illustrated embodiment, the object detector 203 of the computing device 20 may lower the resolution of the image frame in order to increase the transmission speed and the detection processing speed in the range of the minimum resolution that can detect the contour. Regarding the video frame transmitted to the second client 30, the second client such as the processing speed of the object predictor 304 of the second client 30, the processing speed of the matching unit 306, the resolution of the video output screen, and the like. The resolution of the image frame may be lowered based on the resource of. In some cases, an image frame to be transmitted to the computing device 20 or the second client 30 may be converted to a higher resolution.

In addition, according to the addition of the image preprocessor 103, the first client 10 may further include a determiner configured to determine whether to perform image preprocessing based on resources of the computing device or the second client as described above. Can be. In one embodiment, the determination unit 102 may further have the additional determination function for determining whether to convert the resolution of the image frame. In this case, the determination unit 102 further includes a function of determining whether to change the resolution of an image frame to be transmitted to the computing device based on at least one of a resource of the computing device 20 and a state of the network 40. . In this case, the resource may include at least one of a processor processing speed and available memory capacity of a computing device, and the state of the network may include at least one of a connection state and a transmission speed with the network. .

The determination unit 102 may further include a function of determining whether to change the resolution of an image frame to be transmitted to the second client based on the resources of the second client 30. In this case, the resource may include information on at least one of a processor processing speed, available memory capacity, and resolution of an image output screen of the second client.

In this embodiment, it is assumed that the first client 10 acquires and knows information about the resource of the computing device 20 or the resource of the second client 30 in advance. That is, when the first client, the second client, and the computing device are connected to each other through a network, the first client may acquire information about each remaining device in advance. In an alternative embodiment, the first client may obtain the information by requesting the resource information from the computing device or the second client when the determination of the determining unit 102 is needed.

The image encoder 104 has a function of compressing an image frame before transmission of the image frame. For example, the encoder unit 104 may compress an image frame by a standard compression algorithm such as MPEG4.

4, the computing device 20 of FIG. 6 further includes an image decoder 202, a determiner 204, a contour preprocessor 205, and a contour encoder 206.

The image decoder 202 corresponds to the image encoder 104 of the first client 10. When the computing device 20 receives the image frame compressed by the encoder 104, the image decoder 202 is provided. Decompress this image.

The contour encoder unit 206 has a function of compressing the contour information before transmitting the contour of the object detected by the object detector 203 to the second client 30. For example, the contour encoder 206 may compress the contour information by, for example, an Arithmetic Coding scheme.

The contour preprocessor 205 converts the resolution of the contour to be transmitted to the second client 30. Since the computing device 20 receives the image converted to the low resolution when the image is transmitted from the first client, when the contour is transmitted to the second client 30, it may be necessary to convert the image to the high resolution again.

According to the addition of the contour preprocessing unit 205, the computing device 20 further includes a determination unit 204 that determines whether to perform the contour preprocessing based on resources of the second client. In one embodiment, the determination unit 204 determines whether to change the resolution of the contour to be transmitted to the second client based on the resources of the second client 30. In this case, the resource may include at least one of a processor processing speed and an available memory capacity of the second client.

In this case, it is assumed that the computing device 20 acquires and knows information about a resource of the second client 30 in advance. That is, when the first client, the second client, and the computing device are connected to each other through a network, the computing device may obtain the information from the second client 30 in advance. In alternative embodiments, the computing device may receive information of the second client 30 from the first client 10.

In comparison with FIG. 4, the second client 30 of FIG. 6 further includes a decoder 302 and a contour preprocessor 303.

The decoder unit 302 corresponds to the image encoder unit 104 of the first client 10 and the contour encoder unit 206 of the computing device, and is transmitted from the first client 10 and the computing device 20, respectively. Compressed image frame and contour information is decompressed.

The contour preprocessor 303 converts the resolution of the contour detected by the object detector 203 as the same functional module as the contour preprocessor 205 of the computing device. In one embodiment, if the determining unit 302 of the computing device determines that the resource of the second client 30 is not sufficient and the contour preprocessing unit 205 does not perform the preprocessing, the contour preprocessing of the second client 30 is performed. The unit 303 performs preprocessing. On the contrary, if the determination unit 302 determines that the computing device performs the contour preprocessing due to insufficient resources of the second client, the contour preprocessing unit 205 performs the preprocessing, and in this case, the second client 30 The contour preprocessor 303 does not need to perform preprocessing.

7 is a flowchart illustrating a method for tracking a distributed object based object according to a third embodiment of the present invention of FIG. 6. In comparison with FIG. 5, FIG. 7 further includes image preprocessing steps S703 and S709 and contour preprocessing steps S715, S717 and S722, and all remaining steps are the same.

In steps S703 and S709, the first client 10 converts the resolution of an image frame to be transmitted to each device based on respective resources of the second client and the computing device, respectively. If the first client 10 and the second client 30 are one device, the resolution conversion process may be omitted for the image frame to be transmitted to the second client.

The computing device 20 then determines whether to perform preprocessing of the contour based on the resources of the second client in step S715. If it is determined that the resources of the second client 30 are not sufficient, the contour preprocessing is performed in the computing device 20 (step S717), and if it is determined that the resources of the second client 30 are sufficient, the second client 30 is determined. Contour preprocessing is performed (step S722).

The technique related to the distributed processing-based object tracking apparatus and method of the present invention described above may be implemented in hardware or software, or a suitable combination of the two. Thus, for example, all or part of an object tracking method may take the form of program code (ie, instructions) implemented on a tangible medium such as a floppy diskette, CD-ROM, hard drive, or any other computer readable storage medium. When such program code is loaded on or executed by a machine such as a computer, the machine may be part of the object tracking device.

While the present invention has been described with reference to the particular embodiments and drawings, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims. This is possible.

10: first client 101: image acquisition unit
102: determination unit 103: image preprocessing unit
20: computing device 203: object detection unit
204: Decision unit 205,303: Contour preprocessor
30: second client 306: matching unit

Claims (24)

In the distributed tracking object tracking method,
Receiving, by the second client device, a plurality of image frames including the tracked object from the first client device;
Predicting, by a second client device, the position of the object in an image frame (“predictive image frame”) after the plurality of image frames based on at least one of the plurality of image frames;
Receiving, by a second client device, a position and a contour of the object from a computing device detected in one of the plurality of image frames; And
And matching, by the second client device, the contour with the predicted position of the object and the detected position of the object to calculate the tracking position and the tracking contour of the object in the predictive image frame. ,
Prior to the matching step,
Determining whether a second client device performs the matching step;
And if the matching step is not performed, the predicted position and the contour of the object are calculated as the tracking position and the tracking contour of the object. The method of claim 1, wherein the step of matching,
Calculating a tracking position of the object by weighted summing up the predicted position and the detected position of the object; And
And calculating the tracking contour of the object by combining the detected contour of the object with the tracking position. The method of claim 1, wherein predicting the position of the object comprises:
And predicting a contour of the object in the predicted image frame based on at least one of the plurality of image frames. delete The method of claim 1, wherein the determining comprises: determining based on at least one of a resource and a predetermined time period of the second client device.
In this case, the resource comprises at least one of the processor processing speed and the available memory capacity of the second client device. 6. The distributed processing method according to claim 5, wherein the first client device and the second client device are integrated into one device, and the computing device is a server device connected through the network with the integrated device. Based object tracking method. In the distributed tracking object tracking method,
Acquiring, by the client device, a plurality of image frames including the object to be tracked;
Determining whether the client device transmits at least some image frames of the plurality of image frames to a computing device, and when determining that the client device transmits the at least some image frames;
Predicting, by the client device, the position of the object in an image frame (“predictive image frame”) after the plurality of image frames based on at least one of the plurality of image frames;
Receiving, by the client device, a position and a contour of the object detected in any one of the some image frames from the computing device; And
And matching, by the client device, a contour with the predicted position of the object and the detected position of the object to calculate a tracking position and a tracking contour of the object in the predictive image frame.
Prior to the matching step,
Determining whether the client device performs the matching step;
And if the matching step is not performed, the predicted position and the contour of the object are calculated as the tracking position and the tracking contour of the object. 8. The method of claim 7, wherein determining whether to send to the computing device is based on at least one of a resource and a predetermined time period of the computing device,
And wherein said resource comprises at least one of processor processing speed and available memory capacity of said computing device. The method of claim 7, wherein the step of matching,
Calculating a tracking position of the object by weighted summing up the predicted position of the object and the detected position of the object; And
And calculating the tracking contour of the object by combining the detected contour of the object with the tracking position. The method of claim 7, wherein the step of predicting the position of the object,
And predicting a contour of the object in the predicted image frame based on at least one of the plurality of image frames. delete The method of claim 7, wherein the determining of whether to perform the matching step comprises: determining based on at least one of a resource and a predetermined time period of the client device;
And wherein the resource comprises at least one of a processor processing speed and an available memory capacity of the client device. The method of claim 7, prior to transmitting the partial image frame to the computing device,
Changing, by the client device, a resolution of an image frame to be sent to the computing device based on a resource of the computing device;
The resource tracking method of claim 1, wherein the resource comprises at least one of a processor processing speed and an available memory capacity of the computing device. The method of claim 13, wherein prior to the matching step,
Changing, by the client device or the computing device, the resolution of the contour of the detected object based on a resource of the client device;
In this case, the resource of the client device may include at least one of a processor processing speed of the client device, available memory capacity, and a resolution of an image output device included in the client device. In the distributed tracking object tracking device,
A communication unit capable of transmitting and receiving data with an external image acquisition device and a computing device through a network;
A position and a contour of the object in an image frame (“predictive image frame”) after the plurality of image frames, based on at least one of a plurality of image frames received from the image acquisition apparatus, including a tracking object. A predictable object predictor; And
When the communication unit receives a position and a contour of the object detected in one of the plurality of image frames from the computing device, the tracking position and the tracking contour of the object in the predicted image frame are calculated. And a matching unit capable of matching a contour with a predicted position of the object and the detected position of the object.
When the communication unit does not receive the detected position and the contour of the object from the computing device, the matching unit determines the position and the contour of the object predicted by the object predictor in the tracking position in the predicted image frame. Distributed processing based object tracking device, characterized in that for calculating and as a tracking contour, respectively. The method of claim 15, wherein the matching unit,
Calculating a tracking position of the object by weighted summing up the predicted position of the object and the detected position of the object,
And a tracking contour of the object by combining the detected contour of the object with the tracking position. delete In the distributed tracking object tracking device,
An image acquisition unit configured to acquire a plurality of image frames including a tracking object;
A communication unit capable of transmitting and receiving data with an external computing device through a network;
A first determining unit determining whether to transmit at least some image frames of the plurality of image frames to a computing device;
An object predictor configured to predict a position and a contour of the object in an image frame (“predictive image frame”) after the plurality of image frames based on at least one of the plurality of image frames; And
And a calculator configured to calculate a tracking position and a tracking contour of the object in the predicted image frame based on the predicted position and the contour of the object.
The calculation unit includes a matching unit for matching the position and the contour of the object detected in the image frame with the predicted position of the object,
The calculation unit further includes a second determination unit for controlling the operation of the matching unit based on at least one of a resource and a predetermined time period of the tracking device,
When the matching unit is not operated by the second determination unit, the calculator calculates the position and the contour of the object predicted by the object predictor as the tracking position and the tracking contour in the predicted image frame, respectively. Distributed processing-based object tracking device, characterized in that. 19. The apparatus of claim 18, wherein the first determiner is based on at least one of a resource and a predetermined time period of the computing device, wherein the resource includes at least one of a processor processing speed and available memory capacity of the computing device. Distributed processing-based object tracking device, characterized in that. The method of claim 18,
The matching unit, when the communication unit receives the position and the contour of the object detected in any one of the image frame of the image from the computing device, the predicted position of the object and the detected position of the object and Distributed processing-based object tracking device, characterized in that matching the contour. The method of claim 20, wherein the matching unit,
Calculating a tracking position of the object by weighted summing up the predicted position of the object and the detected position of the object,
And a tracking contour of the object by combining the detected contour of the object with the tracking position. delete The method of claim 18, wherein the dispersion-based object tracking device,
A third determining unit determining whether to change a resolution of an image frame to be transmitted to the computing device based on a resource of the computing device; And
And an image preprocessor configured to change the resolution of an image frame to be transmitted to the computing device according to the determination result of the third determination unit. A computer-readable recording program for causing a computer to execute the distributed processing based object tracking method according to any one of claims 1 to 3, 5 to 10, and 12 to 14. Possible recording media.

Images