KR101945487B1 - Apparatus and method for recognizing object included in video data - Google Patents

Abstract

According to an embodiment of the present invention, a method for recognizing an object included in a video comprises the steps of: extracting a plurality of stationary images from a video composed of the plurality of stationary images obtained by photographing in a chronological order; determining a type of an object included in each of the plurality of stationary images for each of the plurality of stationary images; determining, as an unidentified image, a stationary image in which a type of an included object is not determined, among the plurality of stationary images; and estimating the type of the object included in the unidentified image based on a type of an object included in each of a first number of stationary images selected in a chronological order close to the unidentified image from the stationary image.

Description

[0001] APPARATUS AND METHOD FOR RECOGNIZING OBJECT INCLUDED IN VIDEO DATA [0002]

The present invention relates to an apparatus and a method for estimating a type of an object included in a still image constituting a moving image with a high probability with respect to a still image whose type is not clearly determined.

Image recognition technology for recognizing objects included in image data has been developed so far due to high demand and continuous research. Such image recognition technology can be applied to various fields such as a security system using face recognition and an input interface for recognizing a user's gesture.

The most important thing in the image recognition technology is to secure the recognition accuracy to a certain level or more. Currently, the technique of recognizing objects included in one still image generally shows good performance. However, in the case of a moving image composed of a plurality of still images, a still image that is difficult to accurately recognize an included object may be generated due to the influence of an object surrounding environment such as motion blur or illumination due to movement of the object . Accordingly, the reliability of recognition of all the still images constituting the moving picture is hardly guaranteed.

In order to achieve the purpose of image recognition in real situations, accurate object recognition must be performed on all the still images constituting a moving image. Therefore, it is necessary to find out objects including the still images having difficulty in recognition with high accuracy .

Korean Unexamined Patent Publication No. 10-2008-0085182 (published on September 23, 2008)

An object of the present invention is to provide an apparatus and method for finding a kind of an object included in a still image that is difficult to accurately recognize in a plurality of still images included in a moving image with high accuracy will be.

It is to be understood, however, that the present invention is not limited to the above-mentioned embodiments, but other and further objects which need not be mentioned can be clearly understood by those skilled in the art from the following description will be.

A method of recognizing an object included in a moving image according to an embodiment of the present invention includes extracting the plurality of still images from a moving image composed of a plurality of still images obtained by photographing in time sequence, Determining a type of an object included in a still image for each of the still images, determining a still image in which the type of the included object is not determined among the plurality of still images as an unidentified image, And estimating a type of the object included in the unidentified image based on the type of the object included in each of the first number of still images selected in descending order of time from the unidentified image.

An apparatus for recognizing an object included in a moving image according to an embodiment of the present invention includes a still image extracting unit for extracting the plurality of still images from a moving image composed of a plurality of still images obtained by photographing in time sequence, An object determination unit for determining a type of an object included in each of the still images with respect to each of the still images and determining a still image from which the type of the included object is not determined as an unidentified image, And an estimation unit estimating a type of an object included in the unidentified image based on a type of the object included in each of the first number of still images selected in descending order of time from the unidentified image of the still image .

According to an embodiment of the present invention, there is provided an image processing method for an unidentified image in which the type of an included object is not determined among still images constituting a moving image, based on information from still images temporally close to the unidentified image, It is possible to estimate the kinds of objects considered to contain the unidentified image with high accuracy. As a result, the reliability of the recognition of the entire video can be enhanced, and ultimately, various services based on image recognition can be provided to the users with high quality.

1 is a block diagram illustrating an apparatus for recognizing an object included in a moving picture according to an exemplary embodiment of the present invention.
FIG. 2 is a diagram illustrating each step of a method for recognizing an object included in a moving picture according to an embodiment of the present invention.
3 is a diagram for explaining details of a method for recognizing an object included in a moving picture according to an embodiment of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The advantages and features of the present invention, and the manner of achieving them, will be apparent from and elucidated with reference to the embodiments described hereinafter in conjunction with the accompanying drawings. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. To fully disclose the scope of the invention to those skilled in the art, and the invention is only defined by the scope of the claims.

In the following description of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear. The following terms are defined in consideration of the functions in the embodiments of the present invention, which may vary depending on the intention of the user, the intention or the custom of the operator. Therefore, the definition should be based on the contents throughout this specification.

1 is a block diagram illustrating an apparatus for recognizing an object included in a moving picture according to an exemplary embodiment of the present invention. The object recognition apparatus 100 includes an input unit 110, an image extraction unit 120, an object determination unit 130, an estimation unit 140, a correction unit 150, an output unit 160, ). However, since the components of the object recognition apparatus 100 of FIG. 1 are only one embodiment of the present invention, the concept of the present invention is not limited to FIG.

The input unit 110 may receive a moving image to be recognized by the object recognition apparatus 100. [ In order to perform such an input function, the input unit 110 may be implemented by an input device such as a keyboard and a mouse, a data bus for receiving data, or a wired / wireless communication module.

The image extracting unit 120 may extract a plurality of still images constituting the moving image from the moving image input through the input unit 110. [ That is, each of the still images may be regarded as an image corresponding to each of a plurality of frames of the moving image, and is generally a video captured at a predetermined interval in time. The image extracting unit 120 extracts the still images and extracts information on the order in which the still images appear in the moving image. The image extracting unit 120 may be implemented by an arithmetic unit including a microprocessor together with an object determining unit 130, an estimating unit 140, a correcting unit 150 and an output unit 160 to be described later. .

The object determining unit 130 may determine the type of object included in each still image and the position in which the object exists in the still image. On the other hand, among the still images, there may be still images in which the types of objects included therein are not easily determined. The object determination unit 130 may classify the still images into unidentified images.

The estimator 140 may estimate the type of object included in the unidentified image classified by the object determiner 130. [ Such estimation can be performed using information such as the type of the object included in the still images other than the unidentified image and the temporally different still images and the position of the corresponding object, and a detailed estimation method will be described later.

The correction unit 150 determines whether the type of the included object is a still image determined by the object determination unit 130 (hereinafter referred to as a "confirmation image" for the sake of simplicity in the sense of distinguishing it from an unidentified image) The result of the determination or estimation can be corrected for the unidentified image estimated by the type estimating unit 140. [ That is, the correcting unit 150 can improve the accuracy of the recognition result derived by the object determining unit 130 and the estimating unit 140.

The output unit 160 may output a result of determining or estimating the type and position of an object included in each still image included in the moving image. The output may be in the form of a visual output or may be in the form of an electronic output to the outside of the object recognition apparatus 100. The output unit 160 may include a visual output device such as a display or an auditory output device such as a speaker and may transmit data to the outside of the object recognition apparatus 100 A data bus for data communication, a wired / wireless communication module, and the like.

The database 170 may store the information required by the object recognizing apparatus 100. For example, the database 170 stores the still images extracted by the image extracting unit 120, their temporal order, the determination result of the object determination unit 130, and the estimation result of the estimation unit 140, And the final result of the object recognition after the correction of the object. A learning model used by the object determining unit 130 to determine the type of the object included in the still image may also be stored in the database 170. [ Of course, the information stored in the database 170 is not limited to the above example, and may be any information required for the operation of the object recognition apparatus 100. [

The database 170 may be embodied as a computer readable medium, and examples of such computer readable media include magnetic media such as hard disks, floppy disks and magnetic tape, optical media such as CD-ROMs, DVDs, A magneto-optical media such as a floppy disk, a magneto-optical media such as a floppy disk, a magneto-optical media such as a floppy disk, a magneto-optical media such as a floppy disk, a magneto-optical media such as a floppy disk, a flash memory, .

FIG. 2 is a diagram illustrating each step of a method for recognizing an object included in a moving picture according to an embodiment of the present invention. The object recognizing method of FIG. 2 can be performed by the object recognizing apparatus 100 described with reference to FIG. However, since the method shown in FIG. 2 is only an embodiment of the present invention, the concept of the present invention is not limited to FIG. 2, and each step of the method shown in FIG. And may be performed in a different order of the bar and the order.

First, the input unit 110 can acquire a moving image to be object recognized (S110). In operation S120, the image extracting unit 120 may extract a still image from the moving image. In operation S120, the object determining unit 130 may determine the type and position of the object included in each still image.

Hereinafter, a method for the object determining unit 130 to determine the type of object included in the still image will be described in more detail. In determining the types of objects included in a still image, various methods may be applied by the object determination unit 130. However, as an example, a method of using a learning model constructed through machine learning may be considered You can try.

A learning model generated by machine learning can be generated using a plurality of learning images. Each of the plurality of learning images may be an image including a static image, and may include an image of a specific object. An object included in such a plurality of learning images is referred to as a sample object.

According to the learning model, the relationship between the feature information extracted from each learning image and the sample object included in each of the learning images can be defined. Such a learning model can be generated through various machine learning or deep learning techniques such as CNN (convolution neural network) and DNN (deep neural network), and is limited to be generated only by the above-described method no. The object determination unit 130 may directly generate such a learning model from a learning image prepared in advance, but may use a learning model that is generated and stored in the database 170. [

Using the learning model, the object determination unit 130 can determine the type of object included in each still image extracted from the moving image to be recognized. For example, the object determining unit 130 can calculate the similarity between the shape displayed on the specific still image to be determined and the shape of each sample object stored in the learning model. Then, the object determining unit 130 may determine that the sample object having the largest value of the similarity among the sample objects is an object included in the specific still image, and if the shape of the corresponding object exists in the specific still image The location of the region can also be determined.

However, even if the value of the similarity value is the largest, it may be difficult to view the object included in the specific still image if the value of the similarity value itself is less than a certain level. The object determining unit 130 determines whether the still image having the similarity value between the shape of the still image and the shape of the object included in the still image is smaller than the predetermined threshold similarity, .

For a confirmation image other than the unidentified image among the still images, a process of obtaining a blur value may be performed. The blur value may be a number indicating how blurry the object contained in the still image is. The object represented in the still image may be blurred due to the movement of the object in the shooting process and the result of object recognition may be incorrect for the still image including the blurred object. For example, it is likely that a still image captured at the moment when a finger shape of a person changes from a "V" shape representing a victory to an "O " representing an OK includes fingers photographed blurred by a switching operation And the finger shape is also likely to be intermediate between "V" and "O", the object determination unit 130 determines the type of the object included in the still image as a "V" shaped finger and an "O" shaped finger It is difficult to guarantee that the results are accurate.

Accordingly, the object determining unit 130 may obtain a blur value for each of the confirmed images in the still image, and classify the confirmed image whose blur value is equal to or greater than the predetermined threshold value to be an unidentified image. The blur value of the still image can be calculated based on the frequency analysis result of the region in which the shape of the object is determined to exist in the still image. As the shape of the object is clearly represented in the still image, the specific gravity of the high frequency component occupies a high frequency in the frequency spectrum of the still image. Therefore, the object determination unit 130 uses the principle to calculate the blur value Can be obtained.

An example of a concrete method for obtaining the blur value is as follows. According to this example, a blur value of a still image can be calculated using Laplacian. The object determining unit 130 may calculate a Laplacian value using a block having a predetermined size for all or some selected pixels belonging to an area where the shape of the object is determined to exist in the still image. Equation (1) below is an equation for calculating a laplacian value using a block having a size of 3x3.

In Equation (1), G (x, y) denotes a Laplacian value of an arbitrary pixel existing at a position corresponding to the coordinates (x, y). I (x, y) means a value of a pixel existing at (x, y). That is, the Laplacian value obtained by Equation (1) may be a value indicating the degree of difference between the pixel existing at (x, y) and the pixel adjacent to the pixel.

The object determining unit 130 may obtain the Laplacian value for each pixel in the region where the shape of the object is determined to exist in the still image and then obtain the variance of the Laplacian values. The blur value may be set to be inversely proportional to the value of the variance. According to this, if the magnitude of the dispersion is large, it means that the blur value is small, and if the dispersion size is small, it means that the blur value is large.

As described above, based on the blur value of each confirmation image, the object determination unit 130 can classify the confirmation image whose blur value is equal to or larger than the threshold value to be an unidentified image. The value of the threshold may be empirically determined through experimentation or may be determined based on the distribution of the blur values of each of the verification images. Alternatively, the value of the threshold value may be determined based on a result of analyzing the above-described learning image through machine learning. On the other hand, the object determining unit 130 may not classify the blur value separately, and may classify the still image having a high degree of blur into an unidentified image in the object determining process using the learning model. That is, until the determination of the degree of blur of the still image, the learning model can be left to machine learning. For this, the learning model may also include information on the relationship between the feature information of each learning image and the degree of blur of each learning image.

For the unidentified image determined through the above-described process, the position of the included object can be predicted using a tracking technique, and the object type determination based on the learning model can be performed only on the corresponding position. The tracking technique uses a preceding image temporally preceding a specific unidentified image to find an area in which the shape of the object is considered to exist in the area of the specific unidentified image.

For the tracking, the object determining unit 130 may select a predetermined second number (for example, five) of the preceding images temporally close to the specific unidentified image in order. Next, the object determination unit 130 may detect a position where the shape of the object exists in each preceding image, with respect to each of the selected preceding images. Through this process, the object determination unit 130 can acquire information about the movement of the object, and based on the obtained information, determines an area in which the shape of the object is expected to exist in the specific unidentified image .

Next, the object determination unit 130 may apply the object type determination based on the learning model only to an area where the shape of the object is expected to exist, rather than the entire area of the specific unidentified image. This is based on the fact that the accuracy of object type judgment based on the learning model is higher when the region of the object is actually located than when the entire region of the still image is targeted.

The object determining unit 130 may perform the object type determination using the tracking as described above for each unacknowledged image, and may include an unidentified image (e.g., a shape displayed on the still image and a still image included in the still image) The image having the similarity value between the shapes of the objects determined to be equal to or greater than the threshold similarity degree) can be classified into the confirmation image.

The object determination unit 130 can determine only the image remaining as an unidentified image through the above-described process but not yet identified (S140). For each of the confirmed unidentified images, the estimator 140 may estimate the object type included in the unidentified image (S150). For the estimation operation of the estimation unit 140, a voting technique may be utilized.

For a specific unidentified image, the estimator 140 may select a first number of still images (hereinafter referred to as a "reference image") in order of temporal closeness to the specific unidentified image. Hereinafter, it is assumed that the first number is 10, and in this case, a total of 10 still images, that is, 5 still images continuously existing immediately before the specific unidentified image and 5 still images continuously following the specific unidentified image, It can be selected as a reference image.

Next, the estimator 140 may estimate an object included in the specific unidentified image based on the frequency with which the objects included in each of the reference images appear on the reference image. For example, assume that there are four types of human fingers as objects, and these four shapes are V shape, O shape, palm shape, and thumb shape. In the reference images 10, if the V shape is 5 times, the O shape is 3 times, the palm shape is 2 times, and the thumb raised shape has not appeared once, It is possible to estimate the V-shaped fingers, which have appeared many times, as objects included in the specific unidentified image. That is, the estimator 140 may estimate the object that appeared at the highest frequency in the reference image as an object included in the specific unidentified image.

Alternatively, the estimator 140 may use a weighted sum instead of simply summing the number of times each object appears in the reference image. The weight may be given to have a larger value as the temporal distance is closer to the specific unidentified image. In the above example, although the O shape appears three times and appears less frequently than the V shape, the O shape appears in reference images temporally close to the specific unidentified image, and the V shape appears in temporally distant reference images The size of the total weighted sum may be larger than that of the V shape. In this case, the estimator 140 may estimate the O-shaped finger as an object included in the specific unidentified image. Such a weight can be determined based on, for example, a Gaussian distribution.

The estimator 140 may estimate the type of object included in all unidentified images in the manner described above. In order to estimate the type of the object included in the unidentified image, a voting technique using reference images temporally close to the unidentified image is used, so that still images including the types of the included objects are not included in the conventional method It is possible to estimate the object with high accuracy.

3 is a diagram for explaining details of a method for recognizing an object included in a moving picture according to an embodiment of the present invention. That is, FIG. 3 is intended to explain steps S130 to S150 in FIG. 2 more specifically.

Referring to FIG. 3, it can be seen that the still image 210 can be extracted from the moving image 200. Referring to the three still images 211, 212, and 213 adjacent to each other in time among the still images 210, the adjacent still images 211, 212, and 213 may undergo an object recognition process using a learning model S131). If the object recognition is successful using the learning model, the still images 211, 212, and 213 may be subjected to the blur value calculation process.

The still image 210 having the blur value less than the threshold is determined as the confirmation image in which the object determination is completed, but the still image 210 having the blur value of the threshold value or more, even if the object recognition fails or the object recognition succeeds, (S134), the object recognition is performed using the learning model only for the region in which the shape of the object is assumed to exist. Here, the still image 210 that has succeeded in recognizing the object may be a confirmation image in which the object determination is completed. However, the still image 210 that has failed to be recognized is classified as an unidentified image and subjected to the object type estimation process by voting (S150). When the still image 211 is classified as an unidentified image, the adjacent still images 212 and 213 can provide estimation data by voting as described above.

After the steps up to S150, the determination or estimation of the included object types is completed for each still image. However, there may also be errors in the results of these judgments and estimates. For example, a certain still image may have been determined to include a V-shaped finger as an object. Let us suppose that an object included in two still images adjacent to each other in time before and after the specific still image is an O-shaped finger. Then, the shape of the finger is sequentially changed from O shape to V shape to O shape over three adjacent still images. Generally, the time interval between adjacent still images is extremely short, about 30 ms. The likelihood of such a change over such a short period of time is too low. Therefore, in this case, the type of the object determined to be included in the specific still image can be modified from the V-shaped finger to the O-shaped finger.

In other words, the correction unit 150 may be configured such that the correction unit 150 corrects a specific still image by using the same type of objects included in two still images adjacent to the specific still image temporally, If the type of the object included in the still image is different from the type of the included object, the type of the object included in the specific still image may be modified to the type of the object included in the two still images (S160). The operation of the correcting unit 150 may be performed only on the confirmed image, but may also be performed on the unidentified image on which the estimation is completed.

In addition, the corrector 150 may use only two still images adjacent to a specific still image for correction, or may select a still image closer in time to a specific still image as more than two to use for correction. When more than two still images are selected for correction, the types of the objects included in the selected still images are all the same, and the types of objects included in the specific still images are the types of the objects included in the selected still images The correcting unit 150 may modify the type of object included in the specific still image.

When the correction by the correction unit 150 is completed, the output unit 160 may output the result of object recognition to the outside of the object recognition apparatus 100 (S170). Of course, the result of the object recognition may be stored in the database 170.

According to the object recognition method of the present invention described above, it is possible to provide a high-reliability object recognition result even in the presence of an external factor that lowers the recognition rate of an object such as a motion blur or a sudden change in illumination.

Combinations of each step of the flowchart and each block of the block diagrams appended to the present invention may be performed by computer program instructions. These computer program instructions may be embedded in an encoding processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, so that the instructions, performed through the encoding processor of a computer or other programmable data processing apparatus, Thereby creating means for performing the functions described in each step of the flowchart. These computer program instructions may also be stored in a computer usable or computer readable memory capable of directing a computer or other programmable data processing apparatus to implement the functionality in a particular manner so that the computer usable or computer readable memory It is also possible for the instructions stored in the block diagram to produce a manufacturing item containing instruction means for performing the functions described in each block or flowchart of the block diagram. Computer program instructions may also be stored on a computer or other programmable data processing equipment so that a series of operating steps may be performed on a computer or other programmable data processing equipment to create a computer- It is also possible that the instructions that perform the processing equipment provide the steps for executing the functions described in each block of the block diagram and at each step of the flowchart.

Also, each block or each step may represent a module, segment, or portion of code that includes one or more executable instructions for executing the specified logical function (s). It should also be noted that in some alternative embodiments, the functions mentioned in the blocks or steps may occur out of order. For example, two blocks or steps shown in succession may in fact be performed substantially concurrently, or the blocks or steps may sometimes be performed in reverse order according to the corresponding function.

The above description is merely illustrative of the technical idea of the present invention, and various modifications and changes may be made by those skilled in the art without departing from the essential characteristics of the present invention. Therefore, the embodiments disclosed in the present invention are intended to illustrate rather than limit the scope of the present invention, and the scope of the technical idea of the present invention is not limited by these embodiments. The scope of protection of the present invention should be construed according to the following claims, and all technical ideas within the scope of equivalents thereof should be construed as falling within the scope of the present invention.

According to an embodiment of the present invention, since it is possible to estimate a type of an object included in an unidentified image, which is hard to recognize among still images constituting a moving image, with high accuracy, the reliability of recognition of the entire moving image can be enhanced, Can provide various services based on image recognition to users with high quality.

100: Object recognition device
110: input unit
120:
130:
140:
150:
160: Output section
170: Database

Claims (11)

Extracting the plurality of still images from a moving image composed of a plurality of still images obtained by photographing in time sequence;
Determining a type of an object included in each of the plurality of still images, for each of the plurality of still images;
Determining a still image in which the type of an included object is not determined among the plurality of still images as an unidentified image; And
Estimating a type of an object included in the unidentified image based on a type of an object included in each of the first number of still images selected in descending order of time from the unidentified image among the plurality of still images,
Wherein the determining comprises:
Sequentially selecting a temporally preceding image and a temporally subsequent image based on a still image for which the type of the object is not judged to be a second number;
Detecting, for each of the selected second number of images, a location where the shape of the included object is present;
Determining an area in which an object is expected to exist in a still image in which the type of the object is not determined based on a result of the detection;
Estimating a type of an object existing in the expected area; And
And determining the still image to which the type of the object is not estimated as the unidentified image if the type of the object included in the predicted area is not estimated,
Wherein the determining step comprises:
Using a learning model that is learned based on learning data for which a plurality of learning images are input and the type of sample objects included in each of the plurality of learning images and the blur value of each of the plurality of learning images are correct, The type of object included in each still image and the blur value of each of the inputted still images
How to recognize objects contained in a movie. delete The method according to claim 1,
Wherein the step of determining a still image in which the type of an included object among the plurality of still images is not determined as an unidentified image,
Performing a step of obtaining similarity between a shape displayed on one of the still images and a shape of an object determined to be included in the image based on the learning model for each of the still images; And
Determining, as the unidentified image, a still image whose similarity is less than a predetermined threshold similarity among the plurality of still images
How to recognize objects contained in a movie. The method of claim 3,
Wherein the step of determining a still image in which the type of an included object among the plurality of still images is not determined as an unidentified image,
Determining a still image whose blur value is equal to or greater than a predetermined threshold among the still images having the similarity degree equal to or higher than the threshold similarity as the unidentified image,
The blur value is calculated based on a frequency analysis result for a region in which the shape of the object included in the specific image exists in the specific image for which the blur value among the still images is to be obtained
How to recognize objects contained in a movie. delete The method according to claim 1,
Estimating a type of an object included in the unidentified image,
Estimating a type of an object included in the unidentified image based on a frequency of each of the objects included in each of the first number of still images appearing in the first number of still images
How to recognize objects contained in a movie. The method according to claim 6,
The type of object included in the unidentified image is estimated based on a weight and a frequency assigned to each of the first number of still images and the weight is assigned to have a larger value as the temporal distance is closer to the unidentified image
How to recognize objects contained in a movie. The method according to claim 1,
Wherein the type of the object included in the image is determined based on the type of the object included in the still image when the types of the objects included in the two still images adjacent to each other are temporally identical to each other, The method further comprises modifying the type of the object included in the image into the type of the object included in the two still images
How to recognize objects contained in a movie. A still image extracting unit for extracting the plurality of still images from a moving image composed of a plurality of still images obtained by photographing in time sequence;
An object determination unit for determining a type of an object included in each of the still images with respect to each of the still images and determining a still image from which the type of the included object is not determined as an unidentified image, ; And
And an estimation unit estimating a type of an object included in the unidentified image based on a type of the object included in each of the first number of still images selected in descending order of time from the unidentified image of the still image,
The object judging unit,
Selecting a second number of images in order of a total number of temporally preceding images and temporally subsequent images based on a still image for which the type of the object is not judged; Determining an area in which the object is expected to exist in the still image in which the type of the object is not determined based on a result of the detection, And determining a type of the object included in the predicted area as the unidentified image, wherein the type of the object is not estimated,
The object judging unit,
Using a learning model that is learned based on learning data for which a plurality of learning images are input and the type of sample objects included in each of the plurality of learning images and the blur value of each of the plurality of learning images are correct, The type of object included in each still image and the blur value of each of the inputted still images
A device that recognizes objects contained in a movie. A program stored in a computer-readable medium for performing the method according to any one of claims 1, 3, 4 and 6 to 8. A computer-readable medium having recorded thereon a program comprising instructions for performing steps comprising the steps of any one of claims 1, 3, 4, and 6 to 8.

Images