KR100957191B1 - Method and system for generating hidden picture puzzles image - Google Patents

Abstract

Disclosed are a method and a system for producing a hidden object image. Hidden image finding method according to the present invention comprises the steps of (a) preparing a background image and the object image to be the object of the hidden object; (b) comparing the background image with the object image to determine a position at which the object image is to be inserted into the background image and an amount of rotation of the object image; And (c) inserting the object image into the background image according to the determined position and rotation amount. According to the present invention, it is possible to automatically produce a hidden object image through a comparison between the background image prepared in advance and the object image to be the object of the hidden object search.

Description

Method and system for generating hidden picture puzzles image}

The present invention relates to a hidden object puzzle, and more particularly, to a method and system for automatically producing a hidden object image through image analysis.

Hidden Object Puzzle consists of images hidden in a single background, and is used as a newspaper or magazine component for educational, advertising, and entertainment purposes. In particular, most newspapers provide new hidden object puzzles every day for the purpose of satire and play, which is a major playful element for the subscribers of the newspaper.

In the process of creating a general hidden object puzzle, a professional artist sketches a background image and an object to be hidden, analyzes the shape linkage between the two, selects a location that is difficult to find, and then performs a final drawing process. do.

In order to produce such hidden picture puzzles, the artist requires advanced image analysis skills such as basic drawing ability, form recognition ability, spatial recognition ability, and spatial memory. Therefore, the production of hidden object puzzles by non-professionals is not easy, and newspapers and magazines always publish new works while producing them. However, no attempt has been made to automatically create a hidden object image through computer image analysis.

The technical problem to be achieved by the present invention is to provide a method and system for producing a hidden object image to automatically create a hidden object image through image analysis.

In order to solve the above technical problem, the method for producing a hidden object image finding method according to the present invention comprises the steps of: (a) providing a background image and an object image to be the object of the hidden object image search; (b) comparing the background image with the object image to determine a position at which the object image is to be inserted into the background image and an amount of rotation of the object image; And (c) inserting the object image into the background image according to the determined position and rotation amount.

Here, the step (b) may include (b1) extracting feature points from the background image and the object image; (b2) generating a shape descriptor for each of the extracted feature points; And (b3) determining the position and amount of rotation to be inserted using the generated form descriptor.

The step (b) may further include converting the background image and the object image into a line drawing form, and the step (b1) may extract the feature points from the converted background image and the object image. have.

In addition, the step (b3), (b31) obtaining a similarity between the shape descriptor for each of the feature points extracted from the background image and the shape descriptor for each of the feature points extracted from the object image; And (b32) determining a position at which the object image is to be inserted into the background image by using the obtained similarity, and determining the rotation amount at the determined position.

In addition, the shape descriptor may express a distribution of feature points neighboring the feature point in a two-dimensional histogram according to distance and rotation amount around the feature point. In this case, a main axis may be selected for the corresponding feature point in the form descriptor, and the selected main axis may be a reference of the histogram according to the rotation amount, and the main axis for the corresponding feature point may be selected using a PCA technique.

In addition, the step (b32), (b321) for each of the feature points extracted from the background image to determine a pair of feature points maximizing the obtained similarity and define the similarity value at this time; And (b322) determining a position at which the object image is to be inserted using the defined similarity value, and determining the rotation amount at the determined position.

In addition, the step (b322) may include: forming a similarity map by interpolating the defined similarity values; Determining a point at which the similarity value is maximum in the formed similarity map as a position where the object image is to be inserted; And determining the amount of rotation by using a difference in principal axes between the shape descriptors of the determined pair of feature points with respect to a feature point adjacent to a point where the similarity value is maximum, and using the difference in scale between the shape descriptors of the determined pair of feature points. Determining the scale amount of.

In addition, the step (b32), for each of the feature points extracted from the background image, between the feature descriptor of the background image forming the pair of feature points and the shape descriptor of a predetermined number of feature points in the neighborhood of each feature point of the object image The method may further include modifying the defined similarity in consideration of the similarity, and the step (b322) may determine a position where an object image is to be inserted using the modified similarity value.

In addition, in the modifying the defined similarity, the defined similarity may be corrected by further considering a scale difference between the feature point of the background image constituting the feature point pair and the shape descriptor of the feature point of the object image.

The method for producing a hidden object image may further include transforming the inserted object image using an affine transformation.

In order to solve the above technical problem, the system for producing a hidden object image finder, the storage unit for storing the background image and the object image to be the object of the hidden object finder; A shape matching unit comparing the background image with the object image to determine a position at which the object image is to be inserted into the background image and an amount of rotation of the object image; And an image inserting unit inserting the object image into the background image according to the determined position and rotation amount.

Here, the shape matching unit extracts feature points from the background image and the object image, generates a shape descriptor for each of the extracted feature points, and determines the position and the rotation amount to be inserted using the generated shape descriptor. have.

The shape matching unit may convert the background image and the object image into a line drawing form, and then extract the feature points from the converted background image and the object image.

The shape matching unit may obtain a similarity between the shape descriptor of each of the feature points extracted from the background image and the shape descriptor of each of the feature points extracted from the object image. The position at which the object image is to be inserted can be determined, and the rotation amount at the determined position can be determined.

The image inserting unit may transform the inserted object image using an affine transformation.

In order to solve the above technical problem, there is provided a computer-readable recording medium having recorded thereon a program for executing the above-described hidden object image producing method according to the present invention.

According to the present invention described above, it is possible to automatically produce a hidden object image through the comparison between the background image prepared in advance and the object image of the object to find hidden objects.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description and the accompanying drawings, substantially the same components are denoted by the same reference numerals, and redundant description will be omitted. In addition, in the following description of the present invention, if it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted.

1 is a flowchart of a method for manufacturing a hidden object image according to an embodiment of the present invention.

Referring to FIG. 1, in operation 110, an object image that is a target of a background image and a hidden picture is prepared. Here, the background image may be any type of image, such as a live image taken by a general camera or a directly drawn image, and may be both a color image and a black and white image. A plurality of background images may be prepared in advance, and may be selected by the user for the production of hidden object images.

The object image, which is the object of the hidden object search, may also be in the form of a live image or a directly drawn image, or may be defined by separating a specific object from an arbitrary image. A plurality of object images are also provided in advance, and among them, an object image to be hidden in the background image may be selected by the user or may be automatically selected arbitrarily.

In operation 120, the background image and the object image are compared to determine a position at which the object image is to be inserted into the background image, an amount of rotation of the object image at that position, and an amount of scale of the object image. This step is to determine the position, rotation, and deformation of the object image so as to effectively hide the specific object image in the background image.

In step 130, the object image is inserted into the background image according to the position, rotation amount, and scale amount determined in step 120. Since the inserted object image portion may not be smooth in the image formed as a result of the object image being inserted in this step, the inserted portion is naturally modified by applying an appropriate deformation to the object image in step 140.

2 is a flowchart illustrating a more specific embodiment of step 120 of FIG. 1.

Referring to FIG. 2, in operation 210, the background image and the object image are converted into a line drawing form. The conversion to the line drawing form is intended to achieve a hand-drawn effect. For this purpose, we use the outline search method based on the deflection value of the image. For example, we can use the Coherent Line Drawing method. Coherent line drawing techniques are described in Henry Kang, S. Lee, and C. Chui. Coherent line drawing. In Proc. NPAR, pages 43-50, 2007.], so detailed description is omitted. Referring to FIG. 3, (a) shows a background image prepared in step 110 described above, and (b) shows a background image obtained by converting the background image into a line drawing form. The object image to be hidden is also converted to line drawing. However, if the background image and the object image are originally line drawing images, this step may be omitted.

Next, in step 220, feature points are extracted from the background image and the object image converted into the line drawing form. Here, as the feature extraction method, for example, a Harris corner detection method may be used. For Harris corner detection techniques, see C. Harris and M.J. Stephens. A combined corner and edge detector. In Alvey Vision Conference, pages 147-152, 1998.]. Referring to FIG. 4, (a) shows an object image in the form of a line drawing, and (b) shows a result of extracting feature points from the object image.

Next, in step 230, a shape descriptor is generated for each of the feature points extracted in step 220. In the present embodiment, as a shape descriptor, a method of representing a distribution of neighboring feature points around each feature point as a two-dimensional histogram according to distance and rotation amount is used. Belongie, J. Malik, and J. Puzicha. Shape matching and object recognition using shape contexts. IEEE Trans. Pat. Anal. Mach. Int., 24 (4): 509-22, 2002.

However, in the hidden object search, the object image to be hidden in the background image can have an arbitrary amount of rotation at the hidden position. Therefore, a shape descriptor that is independent of rotation is required to be compared with the background image. Therefore, in this embodiment, the main axis is selected for a certain feature point, and the reference is made to the histogram according to the rotation amount. In this case, a principal component analysis (PCA) technique may be used to select a major axis. Referring to FIG. 5, (a) shows a result of selecting a major axis for the feature point p _i shown in FIG. 4 using the PCA technique, and (b) generates the distance and rotation amount based on the selected axis. The shape descriptor represented by the two-dimensional histogram. To define the shape descriptor, the distance was divided into four sections and the rotation amount was divided into eight sections to form a histogram of 32 sections. In the case of the distance section, the maximum distance was defined as a section that includes all the feature points of the object image, and the range of each section was increased in the log form to increase the importance of nearby feature points. However, the number and range of sections according to the distance or the amount of rotation may be made in various forms.

Like object images, shape descriptors are created for each feature point in the background image. Referring to FIG. 6, (a) shows a background image in the form of a line drawing, and (b) shows a shape descriptor represented by a two-dimensional histogram generated for one feature point of this background image.

Referring back to FIG. 2, in step 240, the shape descriptor from the background image and the shape descriptor from the object image generated in operation 230 are used to determine the position, amount of rotation, and scale amount to be inserted.

7 is a flowchart illustrating a more specific embodiment of step 240 of FIG. 2 described above.

Hereinafter, for convenience, the object image is represented by O , the background image is represented by B , and the feature points included in the kth section ( k = 1, ..., 32) of the histogram with respect to the feature points of the object image are divided into 32 sections. Let the number be O _i (k) . Where i represents the index of the feature point of the object image. The feature points extracted from the background image are denoted by q _j , and the number of feature points included in the k th section of the histogram of the feature points of the background image is denoted by B _j (k) . Here, j represents the index of the feature point of the background image.

의 지역적 유사도가 구해질 수 있으며, 예를 들어 다음 수학식을 이용하여 구해질 수 있다.Referring to FIG. 7, in operation 710, a similarity between the shape descriptor of each feature point extracted from a background image and the shape descriptor of each feature point extracted from an object image is obtained. Each form descriptor contains local form information based on the central feature points, so each form descriptor pair

The local similarity of can be obtained, for example, by using the following equation.

는 배경 이미지 상의 특징점 j 에 대한 형태 기술자와 물체 이미지 상의 특징점 i 에 대한 형태 기술자 간의 유사도를 의미한다. here,

Denotes the similarity between the shape descriptor for the feature point j on the background image and the shape descriptor for the feature point i on the object image.

In operation 720, the most similar feature point pairs, i.e., feature point pairs O _i and B _j maximizing the similarity D , are determined for each feature point j from the background image, and the similarity value S _{j at} this time is expressed as in the following equation. Define.

를 얻게 된 것이다. If the index of the feature point on the object image that is most similar to the feature point j on the background image is I (j) , now we have the most similar feature point pair for each feature point j on the background image.

To get.

의 이웃에 있는 특징점의 형태 기술자와 B _j 의 이웃에 있는 특징점의 형태 기술자 간의 유사도를 고려하여 수정될 수 있다. 본 단계에서는, 특징점 쌍을 이루는 특징점

와 B _j 에 대하여 각각 이웃에 있는 소정 개수 T개의 특징점들 각각에 대한 형태 기술자

및

를 검출하고, 검출된 특징점의 형태 기술자 간의 유사도를 고려하여 상기 정의된 유사도 값 S _j 를 개선한다. 다음 수학식은 S _j 로부터 수정된 유사도 값

를 구하기 위한 식을 나타낸다. On the other hand, the defined similarity value, in step 730

It may be modified in consideration of the similarity between the shape descriptor of the feature point in the neighborhood of and the shape descriptor of the feature point in the neighborhood of B _j . In this step, the feature points that make up the feature point pair

Shape descriptors for each of a predetermined number of T feature points each adjacent to and B _j

And

Is detected and the similarity value S _j defined above is improved in consideration of the similarity between the shape descriptors of the detected feature points. The following equation is the similarity value modified from S _j

The formula for obtaining is shown.

는 B _j 의 이웃하는 t 번째 특징점과

에 이웃하는 t 번째 특징점에 대하여 상기된 수학식 1에 따라 구해진 유사도를 의미한다. 따라 서 우변의 첫 번째 항은 B _j 와

간의 이웃의 유사도를 반영하게 된다. here,

Is the neighboring t th feature point of B _j

A similarity obtained according to Equation 1 described above with respect to a t-th feature point neighboring to. So the first term on the right side is B _j and

It will reflect the similarity of neighbors.

는 물체 이미지와 배경 이미지 내의 목표 영역(삽입될 영역) 간의 스케일 차이를 나타내기 위한 함수이고 w는 미리 정의된 가중치 값이다.

는 예를 들어 다음 수학식과 같이 표현될 수 있다. And the scale difference function in the second term

Is a function for representing the scale difference between the object image and the target region (the region to be inserted) in the background image and w is a predefined weight value.

For example, may be expressed as the following equation.

및

는 각각 B _j 및

내의 히스토그램 디스크의 반지름을 나타낸다. 스케일 차이 함수

는 두 반지름 간의 차이가 작아질수록 그 값이 증가하게 된다. 두 번째 항은 물체 이미지와 목표 영역 간의 크기의 차이가 매우 다를 때 서로 매칭되지 않게 함으로써 숨겨질 물체가 과도하게 스케일링될 가능성을 줄이는 역할을 하게 된다. here,

And

Are B _j and

Indicates the radius of the histogram disk in the disk. Scale difference function

The value increases as the difference between the two radii decreases. The second term serves to reduce the possibility of an overscaled object being hidden by not matching each other when the difference in size between the object image and the target area is very different.

In operation 740, the position at which the object image is inserted is determined using the similarity value modified in operation 730, and the amount of rotation and scale amount of the object image at the position is determined. Step 730 may be omitted, and in this step, the position where the object image is to be inserted is determined using the similarity value defined in step 720.

8 is a flowchart illustrating a more specific embodiment of step 740 of FIG. 7 described above.

In operation 810, the similarity values are interpolated to generate a smooth similarity map. Here, a thin-plate spline (TPS) interpolation technique may be used as the interpolation technique to minimize bending. 9 shows a similarity map generated using the TPS interpolation technique. Referring to FIG. 9, the portion marked in red means that the similarity is high, and the portion marked in blue means that the similarity is low.

Next, in step 820, the point where the similarity value is maximum in the similarity map is determined as the position where the object image is to be inserted.

라 하면,

의 주축을 B _j 에 일치되도록 회전시키는 것이다. 그리고 스케일 양은 B _j 와

의 히스토그램 디스크 간의 반지름의 비율로 결정될 수 있다. In operation 830, the amount of rotation of the object image may be determined using the difference in the principal axes between the shape descriptors of the pair of feature points at the feature point closest to the point determined in operation 820. The difference in the radius is used to determine the scale amount of the object image. For example, the amount of rotation can be determined by the angular difference between the major axes of the shape descriptor pairs and the scale amount can be determined by the ratio between the histogram disks of the shape descriptor pairs. That is, pairs of shape descriptors with B _j

Say,

It is to rotate the main axis of to match B _j . And the scale amount is B _j and

Can be determined as the ratio of the radius between the histogram disks.

Referring back to FIG. 1, in step 130, the object image is inserted into the background image according to the determined position, rotation amount, and scale amount of the object image. FIG. 10 is a detailed view illustrating a result of rotating and scaling an object image and inserting the object image in a portion shown in red in FIG. 9. Referring to FIG. 10, it can be seen that there is a discrepancy gap between the inserted object image and the background image.

In step 140, an appropriate deformation is applied to the object image to smooth the inserted portion. Hereinafter, a more specific embodiment of the 140 step will be described.

In this embodiment, the deformation of the object is defined using an affine transform in order to maintain the unique characteristics of the object image. For example, define pairs between the feature point of the object image and the background image at the currently inserted position, obtain the affine transformation matrix for each pair through iterative optimization, and then return the resulting affine transformation matrix to the object image. Apply. However, some feature pairs may generate incorrect affine transformation matrices and interfere with the optimization process, so these feature pairs are ignored. FIG. 11 illustrates pairs of feature points of an object image and feature points of a background image, and the points indicated in red represent incorrect feature point pairs. 12 illustrates a result of modifying an inserted object image using an affine transformation matrix. 12, it can be seen that the object image is more naturally inserted into the background image than in FIG. 10.

Furthermore, once all object images have been inserted into the background image, the edge detection can be performed once again using the Coherent Line Drawing technique to connect discontinuous edges and generate a more uniform hidden object image.

13 is a block diagram of a hidden object image production system according to an embodiment of the present invention. Referring to FIG. 13, the hidden object image producing system according to the present exemplary embodiment includes a storage unit 10, a shape matching unit 20, an image inserting unit 30, and a background image selecting unit 40.

The storage unit 10 stores a background image and an object image that is a target of hidden object search. Since the object image of the background image stored in the storage unit 10 is the same as the image described in step 110 described above, a detailed description thereof will be omitted.

The image selector 40 selects a background image and an object image to be inserted into the background image from the storage unit 10 according to a user's input. The hidden object image producing system may provide a background image and an object image stored in the storage unit 10 to be viewed by a user through a display unit (not shown) for selecting a background image and an object image of the user.

The shape matching unit 20 compares the background image selected by the image selecting unit 40 with the object image to determine a position where the object image is to be inserted in the background image, the amount of rotation of the object image, and the scale amount of the object image. The more detailed operation of the shape matching unit 20 is the same as described with respect to step 120 described above, and thus a detailed description thereof will be omitted.

When the position, rotation amount and scale amount to be inserted into the object image are determined by the shape matching unit 20, the image insertion unit 30 inserts the object image into the background image according to the determined position, rotation amount and scale amount, and inserts the image. The inserted object image is transformed using affine transformation to naturally process the inserted part. More specific operations of the image inserter 30 are the same as those described above with respect to operations 130 and 140, and thus a detailed description thereof will be omitted.

Meanwhile, in the hidden object image manufacturing system according to the present embodiment, the shape matching unit 20 takes a plurality of object images from the storage unit 10 and compares the background images with the shape descriptors of the object images for comparison. When the image selection unit 40 selects a predetermined number of object images in the order of high similarity among them, the shape matching unit 20 determines the position, amount of rotation and scale to be inserted for each selected object image. To determine the amount, the image inserter 30 may be implemented to insert each object image into a background image.

14 and 15 show a hidden object image generated by hiding object images selected by the image selector 40. 16 illustrates a hidden object image generated by hiding alphabets G, R, A, P, H, I, C, and S as object images.

According to the present invention described above, it is possible to automatically create a hidden image finder image that has been made by hand with a pre-arranged background image and an object image, and a non-expert who is not good at drawing can easily produce a hidden object image. Can be.

Meanwhile, the above-described embodiments of the present invention can be written as a program that can be executed in a computer, and can be implemented in a general-purpose digital computer that operates the program using a computer-readable recording medium. The computer-readable recording medium may include a magnetic storage medium (eg, ROM, floppy disk, hard disk, etc.), an optical reading medium (eg, CD-ROM, DVD, etc.) and a carrier wave (eg, internet Storage medium).

So far I looked at the center of the preferred embodiment for the present invention. Those skilled in the art will appreciate that the present invention can be implemented in a modified form without departing from the essential features of the present invention. Therefore, the disclosed embodiments should be considered in descriptive sense only and not for purposes of limitation. The scope of the present invention is shown in the claims rather than the foregoing description, and all differences within the scope will be construed as being included in the present invention.

1 is a flowchart of a method for manufacturing a hidden object image according to an embodiment of the present invention.

2 is a flowchart illustrating a more specific embodiment of step 120 of FIG. 1.

3 shows a background image prepared in advance and a background image obtained by converting the background image into a line drawing form.

4 shows an object image in the form of a line drawing and a result of extracting feature points from the object image.

5 shows a form descriptor according to an embodiment of the present invention.

6 shows a background image in the form of a line drawing and a shape descriptor generated for one feature point of this background image.

7 is a flowchart illustrating a more specific embodiment of step 240 of FIG. 2.

8 is a flowchart illustrating a more specific embodiment of step 740 of FIG. 7.

9 illustrates a similarity map generated according to an embodiment of the present invention.

10 is a view showing in detail the result of inserting by rotating and scaling the object image to the background image.

11 is a diagram illustrating pairs of feature points of an object image and feature points of a background image.

12 illustrates a result of modifying an inserted object image using an affine transformation matrix.

13 is a block diagram of a hidden object image production system according to an embodiment of the present invention.

14 to 16 show hidden object images generated by hiding selected object images.

Claims (18)

(a) providing a background image and an image of an object to be hidden object search; (b) comparing the background image with the object image to determine a position at which the object image is to be inserted into the background image and an amount of rotation of the object image; And (c) inserting the object image into the background image according to the determined position and amount of rotation, In step (b), (b1) extracting feature points from the background image and the object image; (b2) generating a shape descriptor for each of the extracted feature points; And and (b3) determining the position to be inserted and the amount of rotation using the generated shape descriptor. delete The method of claim 1, In step (b), Converting the background image and the object image into a line drawing form; The step (b1) is a method for producing a hidden object image, characterized in that for extracting the feature points from the converted background image and the object image. The method of claim 1, Step (b3), (b31) obtaining a similarity between the shape descriptor for each feature point extracted from the background image and the shape descriptor for each feature point extracted from the object image; And and (b32) determining a position at which the object image is to be inserted into the background image by using the obtained similarity, and determining the rotation amount at the determined position. The method of claim 4, wherein And the shape descriptor expresses a distribution of feature points neighboring the feature points in a two-dimensional histogram according to distance and rotation amount based on the feature points. The method of claim 5, The method for producing a hidden object image, characterized in that the main axis is selected with respect to the corresponding feature point in the form descriptor and the selected main axis is a reference of the histogram according to the rotation amount. The method of claim 6, The method for producing a hidden object image, characterized in that the principal axis for the corresponding feature point is selected using a PCA technique. The method of claim 6, Step (b32), (b321) determining a feature point pair that maximizes the obtained similarity with respect to each of the feature points extracted from the background image, and defining similarity values at this time; And and (b322) determining a position at which an object image is to be inserted using the defined similarity value, and determining the amount of rotation at the determined position. The method of claim 8, The step (b322), Interpolating the defined similarity values to form a similarity map; Determining a point at which the similarity value is maximum in the formed similarity map as a position where the object image is to be inserted; And The amount of rotation is determined by using a difference between major axes of the shape descriptors of the determined pair of feature points with respect to a feature point adjacent to a point where the similarity value is the maximum, and the scale difference between the shape descriptors of the determined feature point pairs is used to determine an object image. A method for producing a hidden object image comprising determining a scale amount. The method of claim 8, Step (b32), For each of the feature points extracted from the background image, the defined similarity is determined in consideration of the similarity between the feature points of the background image constituting the feature point pairs and the shape descriptors of a predetermined number of feature points adjacent to each of the feature points of the object image. Further includes modifying The step (b322) is a method for producing a hidden object image, characterized in that for determining the position to insert the object image using the modified similarity value. The method of claim 10, Modifying the similarity defined above, The method for manufacturing a hidden picture search image, characterized in that for modifying the defined similarity by further considering the scale difference between the feature point of the background image constituting the feature point pair and the shape descriptor of the feature point of the object image. The method of claim 1, And transforming the inserted object image using an affine transformation. A computer-readable recording medium having recorded thereon a program for executing the method for producing a hidden object image according to any one of claims 1 to 13. A storage unit for storing the background image and the object image to be hidden object search; A shape matching unit comparing the background image with the object image to determine a position at which the object image is to be inserted into the background image and an amount of rotation of the object image; And An image inserting unit inserting the object image into the background image according to the determined position and rotation amount; The shape matching unit extracts feature points from the background image and the object image, generates a shape descriptor for each of the extracted feature points, and determines the position and amount of rotation to be inserted using the generated shape descriptor. Hidden Objects Image Creation System. delete The method of claim 14, And the shape matching unit extracts the feature points from the converted background image and the object image after converting the background image and the object image into a line drawing form. The method of claim 14, The shape matching unit obtains a similarity between a shape descriptor for each feature point extracted from the background image and a shape descriptor for each feature point extracted from the object image, and uses the obtained similarity to add the object image to the background image. And determine the position to be inserted, and determine the rotation amount at the determined position. The method of claim 14, The image inserting unit, the hidden object image creation system, characterized in that for modifying the inserted object image using affine transformation.

Images