WO2014013630A1 - 動画処理装置、動画処理方法、ならびに、情報記録媒体 - Google Patents
動画処理装置、動画処理方法、ならびに、情報記録媒体 Download PDFInfo
- Publication number
- WO2014013630A1 WO2014013630A1 PCT/JP2012/075417 JP2012075417W WO2014013630A1 WO 2014013630 A1 WO2014013630 A1 WO 2014013630A1 JP 2012075417 W JP2012075417 W JP 2012075417W WO 2014013630 A1 WO2014013630 A1 WO 2014013630A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- moving image
- video
- drawn
- depth information
- image processing
- Prior art date
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N5/00—Details of television systems
- H04N5/222—Studio circuitry; Studio devices; Studio equipment
- H04N5/262—Studio circuits, e.g. for mixing, switching-over, change of character of image, other special effects ; Cameras specially adapted for the electronic generation of special effects
- H04N5/265—Mixing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T13/00—Animation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T13/00—Animation
- G06T13/80—2D [Two Dimensional] animation, e.g. using sprites
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/50—Depth or shape recovery
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N13/00—Stereoscopic video systems; Multi-view video systems; Details thereof
- H04N13/10—Processing, recording or transmission of stereoscopic or multi-view image signals
- H04N13/106—Processing image signals
- H04N13/156—Mixing image signals
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N5/00—Details of television systems
- H04N5/222—Studio circuitry; Studio devices; Studio equipment
- H04N5/262—Studio circuits, e.g. for mixing, switching-over, change of character of image, other special effects ; Cameras specially adapted for the electronic generation of special effects
- H04N5/272—Means for inserting a foreground image in a background image, i.e. inlay, outlay
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2210/00—Indexing scheme for image generation or computer graphics
- G06T2210/21—Collision detection, intersection
Definitions
- the present invention relates to a moving image processing apparatus, a moving image processing method, and an information recording medium, and prevents objects from interfering when attempting to combine objects with depth information drawn in a plurality of moving images into one moving image. To do.
- the depth of the background and the depth of the object By combining these technologies, it is possible to generate a 2.5-dimensional video that distinguishes the depth of the background and the depth of the object after separating the background and the object moving in front of it.
- the surface of the object that can be seen from the camera (the front surface of the object) is drawn on the frame that constitutes the 2.5D moving image, and the pixel included in the drawing area where the object is drawn includes the object associated with the pixel.
- the depth to the surface portion front surface depth
- Such a 2.5D moving image is obtained by photographing the real world with a CCD camera and simultaneously using detection by a depth sensor, or photographing the real world with a stereo camera for stereoscopic vision to obtain a parallax for each corresponding pixel, It is obtained by calculating the depth from the parallax.
- the state of the surface of the object not drawn in the frame (the back of the object) is not photographed, and the depth information up to the back is not known.
- collision judgment is performed to prevent situations where objects moving in the virtual space do not interfere with each other or one of them penetrates the other.
- Technology is widely used.
- collision refers to the case where the areas occupied by each other overlap as well as the contact between the surfaces of the objects, or the distance between each other is actually closer than a certain threshold even though they are separated from each other. It includes cases and has a broader meaning than everyday meaning.
- any object is selected depending on the depth. It is decided whether to draw with priority. However, in the first place, if the objects interfere with each other or one of them penetrates the other, the final moving image will be unnatural.
- the present invention solves the above-described problems, and is suitable for preventing objects from interfering with each other when an object with depth information drawn in a plurality of moving images is combined into one moving image.
- An object is to provide a moving image processing apparatus, a moving image processing method, and an information recording medium.
- the moving image processing apparatus is a moving image processing apparatus that synthesizes a first moving image and a second moving image, wherein a first object is drawn on the first moving image, and a front depth of the first object With the information, the second object is drawn in the second moving image, accompanied by the front depth information of the second object, An acquisition unit for acquiring back depth information of the first object and back depth information of the second object; An occupied space that can be occupied by the first object drawn in the first moving image is obtained with reference to the front depth information and the rear depth information of the first object, and the occupied space and a frame with the second moving image are obtained.
- An interference determination unit that determines whether or not the drawn second object satisfies an interference condition with reference to front depth information and back depth information of the second object;
- a range setting unit that sets a possible range in which the second object can be positioned without interfering with the occupied space is determined from the determination result by the interference determination unit.
- the acquisition unit acquires a first depth length previously associated with the attribute of the first object from the database, and acquires a second depth length previously associated with the attribute of the second object from the database. Then, back depth information of the first object is acquired from the front depth information of the first object and the first depth length, and the second object is acquired from the front depth information of the second object and the second depth length. Can be configured to obtain back depth information.
- the range setting unit is configured to acquire an installation condition of the second object previously associated with an attribute of the second object from a database and set the possible range so as to satisfy the installation condition. be able to.
- the range setting unit may be configured to set the possible range so that the occupied space and the second object do not satisfy the interference condition for all frames of the second moving image. .
- Non-deformation transformation representing movement in a three-dimensional space is applied to one of the first video and the second video
- the possible range may be configured to be expressed by a range of values that can be taken by a conversion parameter indicating the amount of movement related to the non-deformation conversion.
- the non-deformation transformation is a parallel movement with the amount of shift in the horizontal and vertical directions as the transformation parameter
- the one video is the second video
- the horizontal and vertical shift ranges set by the range setting unit and the frame with the first moving image are displayed on the screen. It can be configured as follows.
- the non-deformation conversion is a parallel movement using the shift amount in the horizontal direction, the vertical direction, and the depth direction as the conversion parameter
- the one video is the second video
- the possible range of the horizontal and vertical shift amounts set for the current shift amount in the depth direction by the range setting unit, and A frame with the first moving image can be displayed on the screen.
- a correction unit that corrects the one moving image by selecting any conversion parameter from the possible range and performing the non-deformation conversion It can comprise so that the moving image production
- An initial value setting unit for setting an initial value of the conversion parameter If the set initial value is not included in the possible range, the correction unit can be configured to select a conversion parameter closest to the designated initial value from the possible range.
- the moving image processing method of the present invention combines the first moving image and the second moving image, the first object is drawn on the first moving image, accompanied by the front depth information of the first object, In the second video, a second object is drawn, accompanied by front depth information of the second object, An acquisition step of acquiring back depth information of the first object and back depth information of the second object; An occupied space that can be occupied by the first object drawn in the first moving image is obtained with reference to the front depth information and the rear depth information of the first object, and the occupied space and a frame with the second moving image are obtained.
- a range setting step of setting a possible range in which the second object can be positioned without interfering with the occupied space is determined from the determination result in the interference determination step.
- a computer-readable information recording medium records a program for synthesizing a first moving image and a second moving image, and the first object is drawn on the first moving image, and the first object With the front depth information of the second object, the second object is drawn on the second moving image, and the front depth information of the second object is attached.
- An interference determination unit that determines whether or not the drawn second object satisfies an interference condition with reference to front depth information and back depth information of the second object; From the determination result by the interference determination unit, the second object is configured to function as a range setting unit that sets a possible range in which the second object can be positioned without interfering with the occupied space.
- the above program can be recorded on a computer-readable non-transitory information recording medium such as a compact disk, flexible disk, hard disk, magneto-optical disk, digital video disk, magnetic tape, and semiconductor memory.
- This information recording medium can be distributed and sold independently of the computer.
- the above program can be distributed and sold via a transmission medium such as a computer communication network, independently of the computer on which the program is executed.
- a moving image processing apparatus a moving image processing method suitable for preventing objects from interfering with each other when objects with depth information drawn in a plurality of moving images are combined into one moving image
- an information recording medium can be provided.
- the moving image processing apparatus is realized by a computer executing a program.
- the program is read from a non-transitory storage device such as a ROM (Read Only Memory) or a hard disk to a temporary storage device such as a RAM (Random Access Memory).
- a CPU Central Processing Unit
- the CPU controls exchanges with input / output devices such as a keyboard, mouse, touch panel, display, and network interface under the control of the program.
- the moving image processing apparatus is not limited to that realized by a computer that executes a program.
- a dedicated electronic circuit for image processing FPGA (Field Programmable Gate array), DSP (Digital A similar function can be realized by using a signal processor.
- collision In the field of 3D graphics, interference between objects is sometimes called collision.
- collision means that objects collide violently, the surfaces touch each other, and in some cases the shape of the object deforms, but in the case of object interference or object collision. Includes a case where the surface touches, a region in the three-dimensional space occupied by the object bites or penetrates, a case where the object is separated, but the distance is sufficiently close, and the like.
- collision is used as appropriate in place of the term “interference” in order to facilitate understanding.
- FIG. 1 is an explanatory diagram showing the relationship between an object drawn in a 2.5-dimensional moving image and its specifications.
- a frame that is played back at an elapsed time t (typically expressed by a frame number) from the playback start time is denoted as X (t).
- FIG. 1 shows a situation where an object is arranged in a virtual space using a three-dimensional graphics technique as seen from the side of a moving image frame. However, it can be considered that the same situation is established.
- this figure is a side view of the shooting. Therefore, in this figure, the projection plane 12 is represented by a straight line.
- the shooting point 11 corresponds to the position of the camera or the observation position, and the distance between the shooting point 11 and the projection surface 12 is determined by the focal length of the lens and the unit of the pixel length. Further, the shooting direction 16 is a direction of a perpendicular drawn from the shooting point 11 to the projection plane 12.
- the color and front depth of the pixel X (t) [x, y] are expressed as X (t) [x, y] .color and X (t) [x, y] .fore, respectively.
- X (t) [x, y] .color is determined based on the color at the collision point 15, the lighting conditions, the distance between the shooting point 11 and the collision point 15, and the like.
- X (t) [x, y] .color represents coordinate values in various color spaces such as RGB, CYMK, and HSV, and conversion of coordinate values between color spaces is formulated. Also, an ⁇ value indicating transparency may be added to X (t) [x, y] .color.
- X (t) [x, y] .fore is the depth distance between the shooting point 11 and the portion of the object 14 that is drawn in the pixel X (t) [x, y] that hits the collision point 15.
- the depth distance the length 17 of the component in the shooting direction 16 of the vector from the shooting point 11 to the collision point 15 (corresponding to a so-called “Z distance”) is generally used, but an approximate value thereof is used.
- the distance from the shooting point 11 to the collision point 15 may be adopted.
- the depth distance can be expressed in various unit systems.
- the depth of a plurality of moving images can be normalized by using the length of the side of the pixel as a unit. The following description is based on the assumption that the depth is normalized.
- the moving image X may have a background drawn in addition to the moving object.
- the background can be thought of as an object that hardly moves.
- identification number is assigned to the identified object for easy understanding.
- the identification number of the object representing the background is assigned as 0, and the identification numbers of the other objects are assigned as 1, 2, 3,.
- the identification number of the object whose part is drawn in the pixel X (t) [x, y] is expressed as X (t) [x, y] .id
- the identification number of the object appearing in the video X The maximum value is expressed as X.maxid.
- a negative value is given as an identification number to a pixel that is no longer an object of processing in order to facilitate understanding. That is, if X (t) [x, y] .id ⁇ 0, the pixel X (t) [x, y] is interpreted as a transparent pixel.
- transparent corresponds to a blue background in chroma key composition.
- the video obtained by translating the video X by p in the horizontal direction and q in the vertical direction in the frame is expressed as move (p, q, X).
- a video obtained by moving the video X by r in the depth direction is expressed as push (r, X).
- the video obtained by each rotation is expressed as rothor ( ⁇ , X), rotver ( ⁇ , X), rotdep ( ⁇ , X).
- the coordinate value of the pixel may not be an integer value, or the coordinate value may be missing.
- anti-aliasing is performed by interpolating the values such as .color and .fore assigned to each pixel, and the value at the grid point closest to the coordinates is adopted for .id, etc. This method can be used, and where the boundary where .id changes is located by interpolation, .id can be determined based on that boundary.
- pixels that are outside the range of the moving image frame due to parallel movement or the like may be considered as transparent pixels, .id may be set to a negative value as described above.
- shift (d, X) a video in which the time of video X is shifted by d is denoted as shift (d, X).
- select (i, X) a movie that draws only the object with the identification number i drawn on the movie X.
- Z.maxid is shifted by shifting the .id of subsequent objects. Can also be minimized.
- a video obtained by overlaying an object other than the background of video Y on video X is denoted as superimpose (Y, X).
- the brightness and saturation of .color may be modified according to the degree.
- normalization of moving images can be performed based on settings at the time of shooting or the like, but can also be set according to the user's wishes or automatically.
- the size of the object drawn in the image is set to a desired one by simply expanding or reducing the height or width of the moving image or adjusting the resolution based on the user's instruction or the like. .
- the depth distance of the corresponding pixel is multiplied by a coefficient corresponding to the enlargement ratio, but the depth distance of the corresponding pixel is used as it is in normalization.
- the user selects a coefficient, and the depth distance of each pixel is collectively multiplied by the coefficient.
- the moving image W is a normalized moving image Y so that the object j matches the object i.
- normalization may be performed by setting c and k as desired by the user.
- the following situations can be considered when the collision between the object i and the object j can be determined relatively easily.
- the objects i and j are objects that have nothing on the back side and are planar objects that do not have a thickness and are formed only from the front side.
- This criterion can be extended in the time direction.
- the elapsed time t is expressed by a frame number, that is, a case where a frame at time t + 1 follows a frame at time t.
- the objects i and j have a shape only on the front surface.
- the thickness of the object is zero, and it can be considered that the front surface of the object matches the back surface of the object. Therefore, simple and high-speed determination is possible by performing collision determination by these methods.
- FIG. 2A is a cross-sectional view of the object 14 shown in FIG. 1, and FIG. 2B is a cross-sectional view showing the front surface of the object 14.
- FIG. 2B is a cross-sectional view showing the front surface of the object 14.
- the object 14 has a thickness and its cross section has a spread.
- the shape becomes a planar object consisting only of the front surface of the object 14 as shown in FIG. 2B.
- the planar object is expressed by the curve.
- the depth to the front surface of the object 14 is obtained from information attached to the 2.5-dimensional moving image, in order to know the thickness of the object 14, the depth to the back surface of the object 14 may be obtained.
- the first method is that the user sets the distance between the front surface and the back surface for each object in advance.
- FIG. 2C is a cross-sectional view of an object in which the shape of the back surface is estimated by making the thickness constant. As shown in the figure, the back surface of the object 14 has a shape obtained by translating the front surface of the object 14 in the shooting direction 16.
- max and avg are the maximum and average values of the body part when the variables placed before the semicolon in the subscript part are changed within the range that satisfies the conditions placed after the semicolon. Means. When “always satisfied” is adopted as the condition, only the variable is described as a subscript part.
- FIG. 2D is a cross-sectional view of the object in which the shape of the back surface of the object is estimated by obtaining a representative front depth of the object and adding the thickness to the back depth.
- the back surface of the object 14 is a plane perpendicular to the shooting direction 16, and the object 14 is approximated by a columnar shape extending in the shooting direction 16.
- the number of pixels area (X, t, i) in the area where the object i is drawn at time t in the moving image X is obtained as follows.
- the suffix of ⁇ has the same meaning as max and avg.
- the horizontal coordinate xc (X, t, i) and the vertical coordinate yc (X, t, i) of the representative point of the object i at time t are determined as follows.
- xc (X, t, i) ⁇ x , y;
- X (t) [x, y] .id i x / area (X, t, i);
- yc (X, t, i) ⁇ x , y;
- X (t) [x, y] .id i y / area (X, t, i)
- the width w (X, t, i) and height h (X, t, i) of the area where the object i is drawn at time t in the moving image X are determined as follows.
- w (X, t, i) max x, y;
- X (t) [x, y] .id i x -min x, y;
- X (t) [x, y] .id i x;
- h (X, t, i) max x, y;
- X (t) [x, y] .id i y -min x, y;
- X (t) [x, y] .id i y
- the spherical diameter D (X, t, i) can be determined in various ways, for example, as follows.
- D (X, t, i) max [w (X, t, i), h (X, t, i)];
- D (X, t, i) (w (X, t, i) 2 + h (X, t, i) 2 ) 1/2 ;
- D (X, t, i) area (X, t, i) 1/2 ;
- D (X, t, i) max x, y;
- X (t) [x, y] .id i ((x-xc (X, t, i)) 2 + (y-yc (X, t , i)) 2 ) 1/2 ;
- D (X, t, i) avg t area (X, t, i) 3/2 / area (X, t, i)
- avg t area (X, t, i) means the time average of the area in the moving image X where the object i is drawn. Therefore, avg t area (X, t, i) 3/2 corresponds to the estimated value of the volume occupied by object i. By dividing this by area (X, t, i), an estimated depth length can be obtained. Become.
- FIG. 2E is a cross-sectional view of the object in which the shape of the back surface of the object is estimated as a spherical surface.
- the back surface of the object 14 is a spherical surface centered on the representative point 21, but the front surface to the spherical surface of the object 14 is approximated by a columnar shape extending in the shooting direction 16.
- the attribute “person” is associated with a depth length of “XX cm”, the attribute “airplane” with a depth length of “XX m”, and the like. Then, the depth length of the object is acquired from the attribute estimated by image recognition from the appearance of the object drawn in the moving image or the attribute set by the user individually selecting the object.
- the depth length of the object for the attribute is the size of the product described in the product details. It is also possible to obtain from information.
- the occupying section is from the front depth X (t) [x, y] .fore to the rear depth X (t) [x, y] .back.
- overlap (X, Y, t, x, y) 0 and overlap (X, Y, t, x, y)> 0 , Objects i and j will collide.
- overlap (X, Y, t, x, y) min [X (t) [x, y] .back, Y (t) [x, y] .back] -Y (t) [x, y].
- overlap (X, Y, t, x, y) Y (t) [x, y] .back-max [X (t) [x, y] .fore, Y (t) [x, y] .fore ]
- overlap (X, Y, t, x, y) min [X (t) [x, y] .back, Y (t) [x, y] .back] -X (t) [x, y].
- FIG. 3 is an explanatory diagram showing a schematic configuration of elements that perform collision determination in the moving image processing apparatus according to the present embodiment.
- FIG. 3 is an explanatory diagram showing a schematic configuration of elements that perform collision determination in the moving image processing apparatus according to the present embodiment.
- the elements that perform the collision determination in the moving image processing apparatus 101 include a rear depth acquisition unit 102 and a collision determination unit 103.
- the moving image processing apparatus 101 sets the first moving image X and the second moving image Y as processing targets.
- first video X a 2.5-dimensional image taken of a dancer dancing in the real world is adopted.
- second video Y a video showing a character dancing with no background
- Adopt a video that shows how you are doing.
- the moving image processing apparatus 101 in the present embodiment finally synthesizes the third moving image in which the dancer is dancing with the character or another user.
- the moving image processing apparatus 101 detects a collision between the first object i drawn on the first moving image X and the second object j drawn on the second moving image Y to be superimposed on the first moving image X. judge.
- the moving image processing apparatus 101 is typically realized by executing a program on a computer.
- the first video X is accompanied by the first front depth of the first object i on the side drawn in the first video X
- the second video Y is the first video on the side drawn in the second video Y.
- the 2nd front depth of 2 objects j is the 2nd front depth of 2 objects j.
- the first video X may be composed only of a moving object or may include a background.
- i> 0 is adopted as the first object
- i ⁇ 0 is adopted as the first object.
- the object that is subject to collision determination in the second video Y does not include the background.
- j> 0 is used as the second object.
- the back depth acquisition unit 102 first depth of the first object i on the side not drawn in the first video X and the second back side of the second object j on the side not drawn in the second video Y. Get the depth and.
- the collision determination unit 103 functions as an interference determination unit that determines interference between objects. Then, the collision determination unit 103 overlaps the first moving image X and the second moving image Y, whereby the first drawing area in which the first object i is to be drawn and the second drawing in which the second object j is to be drawn. In the overlapping area where the drawing area overlaps and the first drawing area overlaps with the second drawing area, the first occupied section from the first front depth to the first back depth of the first object i, If the second occupied section from the second front depth to the second rear depth of the two objects j overlaps, it is determined that the first object and the second object collide.
- the first object i occupies the first occupied section from X (t) [x, y] .fore to X (t) [x, y] .back as the depth.
- the first object j occupies the second occupied section from Y (t) [x, y] .fore to Y (t) [x, y] .back.
- the first occupied section and the second occupied section overlap, that is, the first occupied section includes the second front depth or the second rear depth, or the second occupied section includes the first front depth or the second
- the first occupied section includes the second front depth or the second
- the back depth information is used.
- the collision determination between the first object i and the second object j in the frame at the time t is performed.
- FIG. 4 is a flowchart showing the flow of the collision determination process according to the present embodiment. Below, with reference to this figure, the collision determination process performed with the moving image processing apparatus 101 which concerns on this embodiment is demonstrated.
- the moving image processing apparatus 101 receives the first moving image X and the second moving image Y that are targets of collision determination (step S151). An object is drawn in the first moving image X and the second moving image Y, and the depth information of the object is attached.
- step S152 the process of scanning the frames of the first moving image X and the second moving image Y in order from the top is repeated.
- Step S153 it is determined whether or not there is an overlap between the region where the object is drawn in the first moving image X and the region where the object is drawn in the second moving image Y in the currently scanned frame. If there is no overlap (step S153; No), it will progress to step S157 and will repeat a process.
- step S153 If there is an overlap (step S153; Yes), the depth information of the object in the first moving image X (corresponding to the “first object” above) that overlaps the drawing area and the object in the second moving image Y Depth information (corresponding to the “second object” described above) is acquired (step S154).
- step S155 it is determined whether or not there is an overlap in the occupied section of the object obtained from the depth information. If there is no overlap, the process proceeds to step S157 and the process is repeated.
- step S155 If there is an overlap (step S155; Yes), the determination result that the object drawn in the first moving image X and the object drawn in the second moving image Y collide is output (step S156). This process ends.
- the overlapping of the drawing areas and the overlapping of the occupied sections based on the depth information are determined for all the frames. However, if the number of frames is large, thinning can be performed as appropriate.
- any object drawn in the first moving image X and any object drawn in the second moving image Y collide at any time and in any position”.
- the determination is not made, that is, when no collision occurs, even if the second moving image Y is superimposed on the first moving image X, there is no contradiction in the context of the objects. Accordingly, the third moving image superimpose (Y, X) in which the second moving image Y is superimposed on the first moving image X can be obtained.
- the user interface is devised when the user moves the second moving image Y with respect to the first moving image X by a drag-and-drop operation using a mouse and then superimposes them.
- a vertical and horizontal translation amount for designating the second moving image Y after shifting the second moving image Y with respect to the first moving image X is specified by a user's drag and drop operation.
- the operation system of the present embodiment is not necessarily limited to drag and drop, and drag and drop by adopting various instruction input operations such as a slide operation on the touch screen and an operation of repeating the click operation twice. It is also possible to change to
- FIG. 5A is an explanatory diagram showing the state of the screen before the dragging is started
- FIG. 5B is an explanatory diagram showing the state of the screen when the dragging is started
- FIG. 5C is a diagram in the middle of the dragging
- FIG. 5D is an explanatory diagram showing the state of the screen at the time of drop.
- FIG. 6 is an explanatory diagram showing a schematic configuration of a portion related to the drag-and-drop process of the moving image processing apparatus 101.
- FIG. 7 is a flowchart illustrating the flow of control of the drag and drop process executed by the moving image processing apparatus 101 according to the present embodiment.
- the shift amount setting unit 104 and the moving image generating unit 105 operate based on the user's drag and drop operation.
- the moving image processing apparatus 101 When the moving image processing apparatus 101 shifts the second moving image Y from the first moving image X and then superimposes them, the moving image processing apparatus 101 determines whether or not the position condition is satisfied so that the positional relationship of the objects satisfies the predetermined position condition.
- the interference determination part for performing is provided.
- the collision determination unit 103 is used as the interference determination unit, and the determination result is used.
- the back depth acquisition unit 102 prior to the collision determination unit 103 is not shown.
- the determination result by the collision determination unit 103 described above can also be used for determination of a position condition with reference to a positional relationship other than object interference.
- correction unit 106 illustrated in the drawing is an element that corrects the shift amount for shifting the second moving image Y from “no shift” so that interference between objects does not occur.
- the correcting unit 106 displays a representative frame of the first moving image X and a representative frame of the second moving image Y on the screen (step S181). For example, as shown in FIG. 5A, a window 411 for the first moving image X, a window 412 for the second moving image Y, and a cursor 413 are displayed in the screen 401.
- the simplest method is to adopt the first frame of each video as a representative frame.
- the user may select a frame to be synchronized and set it as a representative frame.
- the difference in the elapsed time of the frames to be synchronized corresponds to the shift amount for shifting the second moving image Y by time.
- the first video X is a video with a background
- the edge of the window 411 is drawn with a solid line.
- the second moving image Y is a moving image without a background
- the periphery of the window 412 is drawn with a transparent color
- the edge is also drawn with a dotted line. Note that the edge of the window 412 is not necessarily displayed on the screen 401. In this case, it will appear to the user as if only the non-background object drawn in the second moving image Y is displayed on the screen 401.
- the moving image processing apparatus 101 receives a click operation (start of dragging) in the window 412 of the second moving image Y by the user (step S182).
- start of dragging the click operation
- the user moves the cursor 413 using the mouse and clicks a position [s, t] in the second moving image Y.
- the moving image processing apparatus moves the display position of the window 412 of the second moving image Y accordingly (step S184). For example, as shown in FIG. 5C, when the user moves the cursor 413 while maintaining the click, the window 412 keeps the cursor 413 pointing at the position [s, t] in the second moving image Y. The whole is moved.
- the position [u, v] in the first moving image X and the position [s, t] in the second moving image Y overlap. Therefore, it is considered that the user intends to superimpose the second moving image Y on the first moving image X after performing conversion such that the second moving image Y is moved by u-s in the horizontal direction and v-t in the vertical direction.
- the shift amount setting unit 104 of the moving image processing apparatus 101 sets the horizontal direction u-s and the vertical direction vt as the shift amounts (step S185). That is, the moving image processing apparatus 101 tries to superimpose a moving image move (us, vt, Y) obtained by moving the second moving image vertically and horizontally parallel to the frame on the first moving image X. It is interpreted.
- the moving image processing apparatus 101 uses the collision determination unit 103 to change the object drawn in the first moving image X and the object drawn in the intermediate moving image move (us, vt, Y) in which the second moving image Y is shifted. Are determined to collide at any frame time and at any in-frame position or no collision occurs (step S186).
- step S186 If no collision occurs as a result of the drop operation shown in FIG. 5C (step S186; No), the moving image generating unit 105 of the moving image processing apparatus 101 performs the third moving image superimpose (move (us, vt, Y), X ) Is output (step S187), and this process ends.
- the user can obtain the third moving image positioned so that the object drawn in the moving image Y does not collide with the object in the moving image X.
- the third moving image since the objects drawn in the moving image Y do not collide with each other, it is possible to prevent an unnatural drawing in which the objects are engaged with each other.
- step S186 if a collision occurs (step S186; Yes), the moving image processing apparatus 101 issues a warning by voice, vibration, screen display, or the like (step S188). This process ends and no new video is output. That is, if the user's instruction is appropriate, the third moving image is generated, but if it is inappropriate, only a warning is generated and the third moving image is not generated.
- the conversion process is performed in a demand-driven manner including the following embodiments, the conversion for the frame is not performed until it is necessary to determine the duplication of a certain frame. Therefore, in order to obtain a conclusion that no collision occurs at all, it is necessary to convert all the frames. However, when a collision is found in a frame in the middle of a moving image, conversion processing for subsequent frames is not executed.
- the shift amount is determined by how much the position of the dropped window 412 for the second movie Y is shifted from the position of the window 411 for the first movie X. Is expressed by.
- the shift amount is an amount applied to the second moving image Y, the comparison with the first moving image X is not essential when the shift amount is input. Therefore, a method may be employed in which the user directly inputs a numerical value using a keyboard or visually inputs a shift amount using a mouse or the like.
- FIG. 8 is a flowchart showing the flow of control of the automatic adjustment processing according to the present embodiment.
- a description will be given with reference to FIG.
- the amount of movement up, down, left and right in a single trial is expressed in the horizontal direction ⁇ p and the vertical direction ⁇ q.
- the second moving image Y is adjusted while being shifted in a preset direction.
- the user clicks the position [s, t] in the second video Y and maintains the click in the position [u, v , Then click and drop it.
- the moving image processing apparatus 101 of this embodiment performs the following processing.
- the shift amount setting unit 104 of the moving image processing apparatus 101 sets u as the initial value of the variable p and v as the initial value of the variable q (step S191).
- the shift amount setting unit 104 is set to the initial value setting unit.
- the collision determination unit 103 of the moving image processing apparatus 101 determines whether the position condition “the first moving image X and the intermediate moving image move (ps, qt, Y) in which the second moving image Y is shifted by the set shift amount. In step S192, it is determined whether or not “no object collision occurs”.
- step S192 If established (step S192; Yes), the moving image generating unit 105 outputs the third moving image superimpose (move (p-s, q-t, Y), X) by synthesis (step S193), and ends this process.
- the correcting unit 106 increases the value of the variable p by ⁇ p and increases the value of the variable q by ⁇ q (step S194). As a result, the shift amount is updated minutely.
- step S195 when the abandonment condition is satisfied, such as when the parallel movement amount (ps, qt) of the video Y becomes too large or the number of repetitions exceeds a predetermined threshold (step S195; Yes), the collision Is inevitable, a warning is issued (step S196), and the process ends.
- step S195 the control of the moving image processing apparatus 101 returns to step S192.
- the direction of increasing the vertical coordinate value corresponds to the downward direction of the screen.
- This is equivalent to searching for a parallel movement amount that avoids a collision while moving from top to bottom, and if not found, this time corresponds to searching a parallel movement amount that avoids a collision while moving from bottom to top.
- the collision determination unit 103 is simply employed as the interference determination unit. This is equivalent to imposing a positional condition that no object collision occurs with respect to the positional relationship between objects. However, if the determination result by the collision determination unit 103 is used, another condition can be adopted as the position condition in the interference determination unit.
- the position condition is ““ No object collision occurs between the first movie X and the movie move (ps, qt, Y) ”and“ the first movie X and the movie move (ps, q + 1-t, Y ) ”, An object collision may occur.
- the object drawn on the second moving image Y is superimposed on the first moving image X, the object drawn on the second moving image Y is aligned so as to be substantially in contact with the ground.
- the position condition is ““ No object collision occurs between the first movie X and the movie move (ps, qt, Y) ”and“ the first movie X and the movie move (p + ⁇ p-s, q + It is also possible to change to “” that an object collision occurs with ⁇ q-t, Y).
- the ground is drawn in the first video X
- the object moving on the plane is drawn in the second video Y
- the orientation of the camera with respect to the ground in the first video X is the plane of the camera in the second video Y.
- the object depicted in the first video X is a stationary object such as the ground, ceiling, or wall.
- the user is confirmed in advance by designating an object in a certain frame, when it is known from the past editing history by the user, when it is known by comparing each frame of the first video X, etc. It is.
- setting the position condition as described above means setting a movable range in which the synthesized object can occupy in the three-dimensional space. That is, “the object may float in the air as long as it does not collide”, “the object is in contact with the floor or the ground”, “the object is in contact with the ceiling”, etc. It is possible to synthesize a video with the following conditions.
- ( ⁇ p, ⁇ q) is a vector representing the displacement of the drawing position of the object drawn in the second moving image Y, and the moving direction and the moving amount are appropriately changed by setting this direction. It is possible.
- the shift amount may be corrected so that the difference between the shift amount and the initial value set by the user is minimized while satisfying the position condition.
- the shift amount closest to the user's designation is adopted.
- the vector difference between the representative point of the first object and the representative point of the second object can be adopted as the direction of ( ⁇ p, ⁇ q).
- a parallel movement is performed in which the second object moves in a direction toward the first object or in a direction away from the first object.
- these methods may be further advanced to appropriately define an evaluation function and minimize it to adopt a method for avoiding object collision.
- it is considered to minimize the “collision amount” when the object collides.
- the collision amount between the moving image X and the moving image Y is considered as the sum of the areas of the collision areas in which the objects that collide in the moving image X and the moving image Y are drawn
- collision (Y, X) ⁇ t, x, y; X (t) [x, y] .id ⁇ 0, Y (t) [x, y] .id> 0, overlap (X, Y, t, x, y)> 0 overlap (X, Y, t, x, y) Can be adopted.
- the collision amount calculated in this way is used as an evaluation function.
- the update direction ( ⁇ p, ⁇ q) of the undetermined variable is determined by random walk, steepest descent method, conjugate gradient method, or (1,0), By preparing (0,1), (-1,0), (0, -1), etc., finding the change in the collision amount when each is adopted, and selecting the one with the smallest amount of collision, What is necessary is just to repeat the update of the value of the variable showing the grade which moves the moving image Y, such as p and q.
- the initial values of p, q, r are u, v, 0, and collision (push ( Find p, q, r that minimizes r, move (ps, qt, Y)), X).
- the moving image Y is a 2.5-dimensional moving image, it does not have information on the back side of the object. Therefore, by setting an upper limit and a lower limit in advance for the parallel movement amounts ps, qt, r and rotation amounts ⁇ , ⁇ , ⁇ , etc., it is possible to prevent the movement of the object of the video Y from appearing unnatural. be able to.
- the above collision and the value obtained by multiplying the absolute value or square value of the conversion parameter such as the translation amount ps, qt, r and the rotation amount ⁇ , ⁇ , ⁇ by a predetermined integer constant There is also a method in which the conversion parameter is continuously updated until the collision amount becomes zero while the result of adding the amount is used as an evaluation function.
- the initial value of the conversion parameter is a value specified by the user
- the shift amount closest to the value specified by the user can be found by incorporating the displacement of the conversion parameter into the evaluation function.
- the user designates the translation amount r in the depth direction by operating the mouse wheel or the like.
- a third moving image Z is generated by combining the first moving image X and the second moving image Y (or a moving image obtained by shifting the second moving image Y according to the user's instruction). Can also be adopted.
- each pixel in the third moving image Z is derived from which object, and if an object is drawn while hiding another object, the pixel is hidden.
- the pixel value and depth information in the hidden area of the object are stored separately.
- the third animation Z is corrected while referring to the stored information so that no object collision occurs.
- the depth direction of the object may change during correction.
- the pixel information in the original second moving image Y may be referred to in order to keep the accuracy of the appearance of the object as much as possible.
- An area where the user can drop that is, an area including a drop position where a moving image in which an object does not collide is generated by the user's drop operation is generated. It is intended to be presented in an easy-to-understand manner before the user's drop operation.
- the droppable area presented to the user includes information corresponding to a possible range that can be taken by a conversion parameter indicating a shift amount in the conversion of moving the video Y.
- a conversion parameter indicating a shift amount in the conversion of moving the video Y.
- an area in which the second object can exist without interfering with the first object in the space drawn in the first moving image X is expressed.
- the possible range is set by the correction unit 106 functioning as a range setting unit.
- FIG. 9A is an explanatory diagram showing the state of the screen before the dragging is started
- FIG. 9B is an explanatory diagram showing the state of the screen when the dragging is started
- FIG. 9C is a diagram in the middle of the dragging
- FIG. 9D is an explanatory diagram showing the state of the screen at the time of dropping
- FIG. 9D is an explanatory diagram showing the state of the screen when dropped.
- FIG. 10 is a flowchart showing a flow of control of highlight display processing of a droppable area.
- an array F is prepared in the RAM, using as an index a candidate for an arbitrary amount of shift that can be adopted when the second moving image Y is shifted in the horizontal and vertical directions (step S501).
- [ps, qt] is adopted as a subscript, but ps, qt may take a negative value. That is, ps is an integer between the value obtained by inverting the sign of the width of the second video Y and the sum of the width of the first video X and the width of the second video Y, and qt is the height of the second video Y. An integer between the value obtained by inverting the sign and the sum of the height of the first moving image X and the height of the second moving image Y may be used.
- Step S501 a possible shift amount candidate [ps, qt] (step S501), an intermediate moving image move (ps, qt, Y) in which the first moving image X and the second moving image Y are shifted. ) (Step S502), and the process of substituting the result of the collision determination into the element F [ps, qt] of the array F (step S503) is repeated (step S504).
- the window 411 for the first moving image X and the window 412 for the second moving image Y are displayed on the screen (step S506), and the start of dragging is accepted (step S507).
- the intermediate video shifted from the second video Y is move (ps, qt , Y).
- the moving image processing apparatus 101 determines each position [p, qt] in the first moving image X according to the value of F [ps, qt] for each position [p, q] in the first moving image X (step S508).
- the process of changing the color (saturation, lightness, hue, or a combination thereof) of the pixel q] is repeated (step S509) (step S510).
- the area where the user can drop and the area where the user cannot drop are distinguished, and one of them is highlighted.
- step S183 the process may proceed to step S183.
- the first moving image X is drawn as it is in the window 411 as shown in FIG. 9A.
- the video processing device calculates a droppable area. Then, as shown in FIG. 9B, the droppable area in the window 411 is highlighted. In this figure, the highlighting is shown by hatching.
- the above method can improve the speed of emphasis after clicking, and can perform redrawing at a high speed when the click is performed again (not shown).
- the user when the user starts dragging on the second video Y, the user can easily know where to drop the first video X to obtain a new video.
- the automatic adjustment of the parallel movement amount in the third embodiment corresponds to correcting the drop position within the droppable area when the drop is made outside the droppable area.
- correcting the drop position at the boundary that separates the droppable area from the outside of the droppable area means that there is a moment when the object of the first moving image and the object of the second moving image contact each other. means.
- the drop position should be corrected to the lower limit of the dropable area.
- FIG. 9E is an explanatory diagram showing a state where the drop position is automatically corrected after the drop. As shown in the figure, the position [s, t] in the window 412 is modified so as to overlap the lower limit of the droppable area below the position [u, v] in the window 411.
- the user can easily understand the dropable position, and even if the drop is made at a position where the user cannot drop, the drop position can be easily obtained by using the already calculated information. Can be corrected.
- the user specifies the horizontal and vertical shift amounts using the drag-and-drop user interface.
- the shift amount can be taken.
- the possible range was illustrated.
- a method in which the user specifies the shift amounts in the horizontal direction, the vertical direction, and the depth direction will be described.
- the shift amount in the depth direction is fixed at 0, but in this method, the initial value of the shift amount in the depth direction is set to 0, and the shift amount depends on the rotation amount of the wheel. Increases or decreases.
- the simplest thinning-out method performs a collision determination only between the representative frame of the first moving image X displayed in the window 411 and the representative frame of the second moving image Y displayed in the window 412. It is a technique.
- the area occupied by the first object is accumulated in all the frames or in each frame thinned out every several frames, and the sum of the accumulated areas (union) is obtained.
- the same collision determination as described above is performed between the occupied space calculated in advance and the second object drawn in the representative frame of the second moving image Y displayed in the window 412. As a result, when a collision occurs, the shift amount is not included in the possible range.
- This method obtains the occupied space in advance for the moving image of the compositing partner and determines the collision with each frame of the moving image to be corrected by the shift amount, but this may be replaced.
- this is a technique of obtaining the occupancy space of the object drawn in the moving image Y in advance, and determining the possible range of the moving amount of the moving image Y by performing a collision determination between the occupied space and each frame of the moving image X.
- This method can be considered as the roles of the first moving image and the second moving image interchanged.
- the method of thinning out the collision determination can also be applied to the above-described drag and drop method.
- a person moving on the floor or ground becomes the second object
- a car moving on the road becomes the second object, or moves on the track.
- the train can be the second object.
- the attributes of the first object floor, ground, road, track, etc.
- the attributes of the second object people, cars, trains, etc.
- the installation conditions of the second object such as “contact with the occupied space” are recorded in the database as the installation conditions for the attribute
- the installation conditions for the occupied space can be set only by setting the attribute of the object in each video. Can be automatically set, and it is possible to set how the object naturally moves.
- the coordinates of the representative point of the object i drawn in the video X in the three-dimensional space are (xc (X, t, i), yc (X, t, i), zc (X, t, i )). This can be viewed as a three-dimensional position vector pos [X, i] (t) that changes with time t.
- the video Y is moved up and down.
- the video W that was translated, rotated, or shifted in time from side to side or depth was obtained and superimposed on the video X.
- the position vector pos [Y, j] (t) that is the trajectory of the object j and the position vector pos [W, j ] (t) may be different in orientation and position, but the shape is the same. That is, since non-deformation transformation that does not deform the trajectory is performed, the trajectory pos [Y, j] (t) and the trajectory pos [W, j] (t) are congruent.
- FIG. 11A is an explanatory diagram showing the states of the locus pos [X, i] (t) and the locus pos [Y, j] (t).
- FIG. 11B is an explanatory diagram showing the states of the trajectory pos [X, i] (t) and the trajectory pos [W, j] (t).
- each trajectory is shown on the horizontal axis
- the passage of time is shown on the vertical axis.
- Each locus is provided with a long and narrow rectangle indicating the horizontal extent of the object.
- the trajectory pos [W, j] (t) in FIG. 11B is obtained by moving the trajectory pos [Y, j] (t) in FIG. 11A in the horizontal direction, and the shapes of both are congruent.
- the locus pos [X, i] (t) and the locus pos [Y, i] (t) is close at time T, and rectangles representing the spread of the object in the horizontal direction overlap.
- the correction unit 106 performs a transformation transformation that transforms the trajectory of the object on the moving image, thereby preventing collision between the objects.
- trajectory pos [X, i] (t) of the object i drawn on the video X with the background is not deformed, and the trajectory pos [Y of the object j drawn on the video Y superimposed on the video X , j] (t) will be described.
- each coordinate value of the position vector is described by adding .x, .y, and .z.
- the depth becomes morph (t) .z / pos [Y, j] (t) .z times due to the deformation of the trajectory.
- This means that the object j is enlarged pos [Y, j] (t) .z / morph (t) .z times in the frame of the moving image Y at time t.
- the object j is also moved in the horizontal direction and the vertical direction.
- the amount of movement is (morph (t) .x-pos [Y, j] (t) .x) ⁇ pos [Y, j] (t) .z / morph (t) .z, (morph ( t) .y-pos [Y, j] (t) .y) ⁇ pos [Y, j] (t) .z / morph (t) .z.
- a moving image of the trajectory of a certain object can be obtained by combining object selection, parallel movement, and enlargement / reduction.
- the moving images without trajectory are drawn by overlapping the trajectory-transformed videos.
- One moving image can be generated.
- Direction of moving the position pos [Y, j] (t) of the point away from the position pos [X, i] (t) of the representative point of the object i X (t) [x, y] .id included in the video X
- the process of deforming the trajectory of the object j so as to move in the direction of the main normal vector of the trajectory pos [Y, j] (t) is repeated so that the change in the degree of deformation before and after time becomes small.
- overlap (X, Y, t, x, y) 0 is established at any time.
- FIG. 12A to FIG. 12D are explanatory diagrams showing how the trajectory is gradually deformed by repeating the processing.
- a description will be given with reference to FIG.
- the locus pos [X, i] (t) and the locus pos [Y, j] (t) are close to each other at time T, as in FIG. 11A.
- the correction amount propagated to adjacent frames is attenuated by multiplying the correction amount for its own frame by a constant of 0 or more and less than 1, and the propagation can be terminated when the correction amount becomes smaller than a predetermined threshold. It ’s fine.
- the correction may not be propagated before and after the time as described above, but smooth correction may be performed by using spline interpolation or the like.
- the second method when the trajectory pos [Y, j] (t) is deformed at time t, the size of the tangent vector (corresponding to the velocity vector) of the trajectory is not changed, and the main normal vector (velocity The vector is orthogonal to the vector and corresponds to the direction in which the direction of the velocity vector is to be bent.) Only the magnitude of the vector is changed, collision (Y, X) is minimized to zero, and the main method This is a method of minimizing the total sum (typically sum of squares) of changes in the size of the line vector.
- FIG. 8 is an explanatory diagram showing a state in which the main normal vector is adjusted in a specific portion of the locus in order to avoid a collision. This figure represents the shape of the trajectory as it is.
- locus pox [X, i] (t) collides with locus pos [Y, j] (t) at time T. Therefore, in the adjustment section Ta to T + a before and after time T, adjust the way of trajectory pos [Y, j] (t) so that collision does not occur, and the corrected trajectory morph (t) obtain.
- the locus pos [Y, j] (t) and the locus morph (t) are congruent in shape after time T + a.
- the first method is easy to calculate, but there may be a large gap between the movement and trajectory of the object j represented by the movie. At this time, the object j suddenly “moon-walked” It may appear as if.
- Whether or not the object j has moved unnaturally depends on the size and orientation of the tangent vector of the trajectory of the object j, the size of the main normal vector (this is a value corresponding to the curvature of the trajectory), and orientation. It is also possible to determine by whether or not the change in the value exceeds a predetermined threshold value.
- an upper limit may be set in advance for the amount of deformation of the trajectory when the trajectory deformation process is repeated. In this case, the movement of the object does not become unnatural, but there is a possibility that the collision cannot be avoided.
- both trajectories may be deformed.
- the object i and the object j may be moved away from each other.
- the amount of change in the main normal vector of the trajectory of the object i and the main trajectory of the object j are increased. It is only necessary to minimize both the amount of change of the normal vector.
- an upper limit is set for the amount of change in the main normal vector, or a warning is issued when the amount of change in the main normal vector exceeds a predetermined threshold value.
- both the trajectories of the objects i and j drawn in both videos X and Y may be transformed.
- the deformation parameter for the object i and the deformation parameter for the object j may be handled together to perform deformation so as to minimize the collision amount.
- the object j drawn in the video Y is an object that dances on a plane, and the background in the video X includes a floor, the object j dances on the floor of the video X It is desirable to overlap.
- the present embodiment realizes this.
- the representative point of the object j drawn in the video Y is selected.
- the representative point selected here corresponds to the toe of the character.
- the pixel position [xb (Y, t, j), yb (Y, t, j)] of the representative point can be calculated as follows, for example.
- xb (Y, t, j) min x; Y (t) [x, yb (Y, t, j)].
- id j x
- collision (Y, X) is minimized to 0, and X (t) [xb (Y, t, j), xb (Y, t, j)].
- X (t) [xb ( Y, t, j), xb (Y, t, j)].
- Y (t) [xb ( Y, t, j), xb (Y, t, j)].
- Fore difference typically sum of squares
- the floor on which the object j is placed in the movie Y and the floor in the movie X can be matched as much as possible, and the behavior of the object j superimposed on the movie X can be made as natural as possible.
- the coordinates (xc (X, t, i), yc (X, t, i)) of the representative point of the object i at the time t of the moving picture X and the depth coordinates zc of the representative point are (X, t, i) is calculated.
- the depth coordinates of each part of the object i can be approximated by zc (X, t, i). Therefore, when the depth of the representative point changes from zc (X, t, i) to zc (X, t, i) + r, the depth at each position is [zc (X, t, i) + r] / It approximates that it has become zc (X, t, i) times.
- Such approximation is effective for parallel movement of a moving image depth method, and rotation around a horizontal axis and a vertical axis.
- the movement destination of only the representative point of the object i that is the object of translation, rotation around the horizontal axis, and the vertical axis is obtained by the same calculation as in the above embodiment. Then, with the moving destination as the center, the above calculation is simplified by arranging the moving image in an enlarged or reduced manner according to the change in the depth of the representative point.
- the depth method is translated and rotated around the horizontal axis and the vertical axis. Even in this case, this is equivalent to moving the plate while being perpendicular to the depth direction.
- the plate In parallel movement in the depth direction, the plate remains perpendicular to the depth direction.
- the rotation angle is limited to about several degrees, and the position of the plate is changed by rotation, but the direction of the plate remains perpendicular to the depth direction. I think.
- a moving image obtained by moving the object i drawn on the moving image X by r in the depth direction is obtained by performing the following processing.
- A) Select only object i from video X,
- B) Translate along the frame so that (xc (X, t, i), yc (X, t, i)) is the origin,
- C) Scale the video to zc (X, t, i) / [zc (X, t, i) + r] times,
- D A translation is performed along the frame so that the origin is (xc (X, t, i), yc (X, t, i)).
- the coordinates of the representative point of the object in the video are (xc (X, t, i), yc (X, t, i)) To (xc (X, t, i), zc (X, t, i) ⁇ cos ⁇ -yc (X, t, i) ⁇ sin ⁇ ), the depth is zc (X, t, i) to zc Move to (X, t, i) ⁇ sin ⁇ + yc (X, t, i) ⁇ cos ⁇ . That is, the following processing may be performed.
- the processing speed can be increased.
- the conversion shift (d, X) is performed to shift the time of the moving image X by d.
- a video X that shot a person (object i) who moves the heel in the mortar with a hand
- a video Y that shot a person (object j) who moves the hand in the same mortar with a hand
- a person A and a person B want to generate a moving image that expresses how they are working together.
- Parameters for shift and fastforward to minimize the evaluation function that represents the amount of collision when the situation of the mortar and surroundings, and the position and orientation of the camera where the image was taken are common to video X and video Y If you ask for.
- these conversions can also be applied when it is desired to determine from a movie whether furniture or electrical appliances can be placed in a house.
- the following process may be performed.
- an electrical product manufacturer or an electrical product seller provides a video X on the web or the like taken from the front of the washing machine door opening and closing periodically.
- a user considering purchasing a washing machine prepares a video Y taken from the front of a room door or furniture door that opens and closes periodically in the room where the washing machine is to be installed. To do.
- the user After normalization is performed so that the distance from the camera to the washing machine in the video X and the distance from the camera to the candidate location for the washing machine in the video Y are substantially the same, the user performs the washing machine in the video X. Are dragged and dropped to the location of the installation candidate in the video Y.
- the collision amount of the object in the videos X and Y is obtained. If there is no collision, it can be estimated that the washing machine can be installed at a desired location. Even if a collision occurs, if the parameters for shift and fastforward that minimize the amount of collision can be obtained so that the collision does not occur, the washing machine can be installed at the desired location. Can be estimated.
- the shooting direction may be limited when the user takes a picture of the room.
- it is supposed to provide videos taken from various directions, such as shooting from the left side, shooting from the right side, etc.
- a moving image shot in the same direction as the direction in which the user shot the room may be selected.
- animation can be synthesized so that the objects do not collide with each other, and the object surface touches the desired moment or the object bottom keeps touching the floor. Can be synthesized.
- the first application example is to allow an exhibitor to easily prepare a video showing the state of a product in an electronic market such as an Internet auction.
- This video material can be prepared by sellers and bidders in the electronic market, or a video distribution system that uploads videos that can be used publicly and allows other users to browse and distribute this video. It is good also as reusing the animation accumulated in the system.
- Exhibitors can either shoot a video showing the background in the room at home, or prepare a decorative table or folding screen to make the product look good. You can keep your items organized and cleaned so that you don't know your private information.
- the background moving image may be selected from moving image materials.
- the photograph can be taken at any place in the room without organizing or cleaning the items.
- the exhibitor drags and drops the video of the new product that has been photographed onto the background video that was previously photographed. Then, by the above-described collision avoidance and automatic position adjustment, a moving image representing a state where the product does not collide with an article (eg, folding screen) of the background image and is in contact with the floor of the background image (eg, the top surface of the decoration table). Can be synthesized.
- an article eg, folding screen
- the floor of the background image eg, the top surface of the decoration table
- composition of the product video with the background video may be made not only by the exhibitor but also by the bidder.
- a party dress is exhibited, and a product moving image representing a state in which a model such as an exhibitor is walking or turning around wearing the dress is provided.
- Bidders can shoot their own background videos taken inside the venue where the party they are going to attend wearing a dress is held, or select from the video materials. Then, the bidder synthesizes the product video of the dress worn by the model with the background video of the party venue.
- the model moves on the floor, but a moving image that does not collide with other people, furniture, and equipment in the venue is synthesized.
- the bidder can check in advance whether the dress displayed as a product matches the atmosphere of the party venue before bidding.
- the second application example is the further use of moving picture material.
- a background video showing how a famous celebrity is dancing to music is provided as a video material
- the user can shoot a singer and synthesize a video that the user dances by himself. It is possible to generate a video that shows how you are dancing with.
- a background video showing how the merry-go-round is moving with vacant seats is provided as video material
- the user moves on a different vehicle, such as a bicycle or cart, and moves almost the same track as the merry-go-round.
- a moving image processing apparatus a moving image processing method suitable for preventing objects from interfering with each other when objects with depth information drawn in a plurality of moving images are combined into one moving image
- an information recording medium can be provided.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Processing Or Creating Images (AREA)
Abstract
Description
前記第1オブジェクトの背面奥行情報と、前記第2オブジェクトの背面奥行情報と、を取得する取得部、
前記第1動画に描かれた第1オブジェクトが占有しうる占有空間を、前記第1オブジェクトの前面奥行情報ならびに背面奥行情報を参照して求め、前記占有空間と、前記第2動画のあるフレームに描かれた前記第2オブジェクトと、が、干渉条件を満たすか否かを、前記第2オブジェクトの前面奥行情報ならびに背面奥行情報を参照して判定する干渉判定部、
前記干渉判定部による判定結果から、前記第2オブジェクトが前記占有空間に干渉せずに位置しうる可能範囲を設定する範囲設定部
を備えるように構成する。
前記取得部は、データベースから、前記第1オブジェクトの属性にあらかじめ対応付けられた第1奥行長を取得し、前記データベースから、前記第2オブジェクトの属性にあらかじめ対応付けられた第2奥行長を取得し、前記第1オブジェクトの前面奥行情報と前記第1奥行長とから前記第1オブジェクトの背面奥行情報を取得し、前記第2オブジェクトの前面奥行情報と前記第2奥行長とから前記第2オブジェクトの背面奥行情報を取得する
ように構成することができる。
前記範囲設定部は、データベースから、前記第2オブジェクトの属性にあらかじめ対応付けられた前記第2オブジェクトの設置条件を取得し、前記可能範囲を、前記設置条件を満たすように設定する
ように構成することができる。
前記範囲設定部は、前記第2動画のすべてのフレームについて、前記占有空間と、前記第2オブジェクトと、が前記干渉条件を満たさないように、前記可能範囲を設定する
ように構成することができる。
前記第1動画と前記第2動画とのいずれか一方の動画には、3次元空間内の移動を表す非変形変換が施され、
前記可能範囲は、前記非変形変換に係る移動の量を表す変換パラメータがとりうる値の範囲により表現される
ように構成することができる。
前記非変形変換は、水平方向、垂直方向のずらし量を前記変換パラメータとする平行移動であり、
前記一方の動画は、前記第2動画であり、
水平方向ならびに垂直方向のずらし量をユーザに修正させるため、前記範囲設定部により設定された水平方向ならびに垂直方向のずらし量の可能範囲と、前記第1動画のあるフレームと、が画面に表示される
ように構成することができる。
前記非変形変換は、水平方向、垂直方向、奥行方向のずらし量を前記変換パラメータとする平行移動であり、
前記一方の動画は、前記第2動画であり、
水平方向、垂直方向、奥行方向のずらし量をユーザに修正させるため、前記範囲設定部により現在の奥行方向のずらし量に対して設定された水平方向ならびに垂直方向のずらし量の可能範囲と、前記第1動画のあるフレームと、が画面に表示される
ように構成することができる。
前記可能範囲からいずれかの変換パラメータを選択して、前記非変形変換を施すことにより、前記一方の動画を修正する修正部、
前記第1動画と、前記第2動画と、を合成した第3動画を生成する動画生成部
をさらに備えるように構成することができる。
前記変換パラメータの初期値を設定する初期値設定部、
前記修正部は、前記設定された初期値が前記可能範囲に含まれなければ、前記可能範囲から、前記指定された初期値に最も近い変換パラメータを選択する
ように構成することができる。
前記第1オブジェクトの背面奥行情報と、前記第2オブジェクトの背面奥行情報と、を取得する取得工程、
前記第1動画に描かれた第1オブジェクトが占有しうる占有空間を、前記第1オブジェクトの前面奥行情報ならびに背面奥行情報を参照して求め、前記占有空間と、前記第2動画のあるフレームに描かれた前記第2オブジェクトと、が、干渉条件を満たすか否かを、前記第2オブジェクトの前面奥行情報ならびに背面奥行情報を参照して判定する干渉判定工程、
前記干渉判定工程における判定結果から、前記第2オブジェクトが前記占有空間に干渉せずに位置しうる可能範囲を設定する範囲設定工程
を備えるように構成する。
前記第1オブジェクトの背面奥行情報と、前記第2オブジェクトの背面奥行情報と、を取得する取得部、
前記第1動画に描かれた第1オブジェクトが占有しうる占有空間を、前記第1オブジェクトの前面奥行情報ならびに背面奥行情報を参照して求め、前記占有空間と、前記第2動画のあるフレームに描かれた前記第2オブジェクトと、が、干渉条件を満たすか否かを、前記第2オブジェクトの前面奥行情報ならびに背面奥行情報を参照して判定する干渉判定部、
前記干渉判定部による判定結果から、前記第2オブジェクトが前記占有空間に干渉せずに位置しうる可能範囲を設定する範囲設定部
として機能させるように構成する。
Signal Processor)等を利用すれば、同様の機能を実現することが可能である。
以下では、理解を容易にするため、2.5次元動画に描かれるオブジェクトを表現するための記法について整理する。図1は、2.5次元動画に描かれるオブジェクトとその諸元の関係を表す説明図である。以下、本図を参照して説明する。
以下では、ある動画から、別の動画を生成するための変換手法について説明する。このような変換手法には、以下のようなものがある。
(1)動画のフレームに沿った上下左右の平行移動
(2)動画の拡大縮小
(3)動画の奥行方向の平行移動
(4)動画の水平軸、垂直軸、奥行軸周りの回転
(5)動画の時間方向の移動
(6)動画から特定のオブジェクトのみを選択
(7)ある動画に別の動画を重ねる合成
以下、順に説明する。
Y(t)[x+p,y+q] = X(t)[x,y]
ならびに、
Y(t)[x,y] = X(t)[x-p,y-q]
が成立する。ここで、上記の等号は、.colorや.foreなど、各画素に割り当てられた値が、いずれも互いに等しいことを意味する。
Y(t)[c×x,c×y] = X(t)[x,y]
すなわち
Y(t)[x,y] = X(t)[x/c,y/c]
が成立する。
Y(t)[c×x,c×y].fore = X(t)[x,y].fore/c
すなわち
Y(t)[x,y].fore = X(t)[x/c,y/c].fore/c
が成立する。
Y(t)[x/k,y/k] = X(t)[x,y]
が成立し、.foreについては、
Y(t)[x/k,y/k].fore = X(t)[x,y].fore + r
が成立する。
Y(t)[x,y] = X(t)[x,X(t)[x,y].fore×cosθ-y×sinθ]
とし、.foreについて、
Y(t)[x,y].fore = X(t)[x,y].fore×sinθ+y×cosθ
が成立する。
Y(t)[x,y] = X(t)[X(t)[x,y].fore×cosφ-x×sinφ,y]
とし、.foreについて、
Y(t)[x,y].fore = X(t)[x,y].fore×sinφ+x×cosφ
が成立する。
Y(t)[x,y] = X(t)[x×cosψ-y×sinψ,x×sinψ+y×cosψ]
が成立する。
Y(t)[x,y] = X(t-d)[x,y]
が成立する。
Y(t)[x,y] = X(t)[x,y], if
X(t)[x,y].id = i;
が成立する。また、.idについては、
Y(t)[x,y].id = 1, if
X(t)[x,y].id = iかつi>0;
Y(t)[x,y].id = 0, if
X(t)[x,y].id = iかつi=0;
Y(t)[x,y].id = -1, otherwise
とすれば、動画Yにおける識別番号の最大値Y.maxidをできるだけ小さくすることができる。
Z(t)[x,y] = X(t)[x,y], if
Y(t)[x,y].id≦0;
Z(t)[x,y] = X(t)[x,y], if
X(t)[x,y].fore<Y(t)[x,y].fore;
Z(t)[x,y] = Y(t)[x,y], otherwise
が成立する。また、.idについては、
Z(t)[x,y].id = X(t)[x,y].id, if
Y(t)[x,y].id≦0;
Z(t)[x,y].id = X(t)[x,y].id, if
X(t)[x,y].fore<Y(t)[x,y].fore;
Z(t)[x,y].id = Y(t)[x,y].id+X.maxid, otherwise
とすれば、オブジェクトの識別番号の重複を避けることができる。
move(p,q,scale(c,move(-p,-q,X)));
move(p,q,push(r,move(-p,-q,X)))
などを利用すれば良い。
Z(t)[x,y] = Y(t)[x/c,y/c]
とする。
W(t)[x,y] = Z(t)[x,y]
とし、.foreについては、
W(t)[x,y].fore = k×Z(t)[x,y].fore
とする。
X(t)[x,y].id = i;
Y(t)[x,y].id = j;
であれば、位置[x,y]は、オブジェクトiが描画されるべき領域ならびにオブジェクトjが描画されるべき領域の両方に重複して含まれることになる。
X(t)[x,y].fore = Y(t)[x,y].fore
が成立すれば、オブジェクトiとオブジェクトjが衝突する、と判定できる。
X(t)[x,y].id = X(t+1)[x,y].id = i;
Y(t)[x,y].id = Y(t+1)[x,y].id = j;
X(t)[x,y].fore < Y(t)[x,y].fore;
X(t+1)[x,y].fore > Y(t+1)[x,y].fore
が成立すれば、時刻tから時刻t+1にかけてオブジェクトjがオブジェクトiの後から前にすり抜けて出てきた、と考えられる。また、
X(t)[x,y].id = X(t+1)[x,y].id = i;
Y(t)[x,y].id = Y(t+1)[x,y].id = j;
X(t)[x,y].fore > Y(t)[x,y].fore;
X(t+1)[x,y].fore < Y(t+1)[x,y].fore
が成立すれば、時刻tから時刻t+1にかけてオブジェクトiがオブジェクトjの後から前にすり抜けて出てきた、と考えられる。
X(t)[x,y].back = X(t)[x,y].fore + thick(X(t)[x,y].id)
とするものである。
(1)オブジェクトiの前面奥行の最大値を採用する。
repfore(X,t,i) = maxx,y;X(t)[x,y].id=i X(t)[x,y].fore
(2)オブジェクトiの前面奥行の平均値を採用する。
repfore(X,t,i) = avgx,y;X(t)[x,y].id=i X(t)[x,y].fore
X(t)[x,y].back = repfore(X,t,i) + thick(i)
あるいは、
X(t)[x,y].back = max〔repfore(X,t,i)
+ thick(i),X(t)[x,y].fore〕
のように定めることができる。ここで、max〔...〕は、括弧内に並べられた値の最大値を意味する。
area(X,t,i) = Σx,y;X(t)[x,y].id=i 1
ここで、Σの添字は、maxやavgと同じ意味を有する。
xc(X,t,i) = Σx,y;X(t)[x,y].id=i x/area(X,t,i);
yc(X,t,i) = Σx,y;X(t)[x,y].id=i y/area(X,t,i)
w(X,t,i) = maxx,y;X(t)[x,y].id=i x - minx,y;X(t)[x,y].id=i x;
h(X,t,i) = maxx,y;X(t)[x,y].id=i y - minx,y;X(t)[x,y].id=i y
D(X,t,i) = max〔w(X,t,i),h(X,t,i)〕;
D(X,t,i) = (w(X,t,i)2+h(X,t,i)2)1/2;
D(X,t,i) = area(X,t,i)1/2;
D(X,t,i) = maxx,y;X(t)[x,y].id=i ((x-xc(X,t,i))2+(y-yc(X,t,i))2)1/2;
D(X,t,i) = avgt area(X,t,i)3/2/area(X,t,i)
area(X,t,i)3/2は、オブジェクトiが占める体積の推定値に相当し、これをarea(X,t,i)で割ることで、奥行長の推定値が得られることになる。
zc(X,t,i) = X(t)[xc(X,t,i),yc(X,t,i)].fore +
D(X,t,i)/2
zc(X,t,i) = X(t)[xc(X,t,i),yc(X,t,i)].fore
X(t)[x,y].back = max〔zc(X,t,i)+〔max〔(D(X,t,i)/2)2-(x-xc(X,t,i))2-(y-yc(X,t,i))2〕,0〕1/2,X(t)[x,y].fore〕
のように定めれば、球面により近似したことになるし、
X(t)[x,y].back = max〔zc(X,t,i)+D,X(t)[x,y].fore〕
のように定めれば、円柱により近似したことになる。
X(t)[x,y].id = 0
が成立する場合には、
X(t)[x,y].back = ∞
とする。
X(t)[x,y].id = i;
Y(t)[x,y].id = j
であり、かつ、
(1)X(t)[x,y].fore≦Y(t)[x,y].fore≦X(t)[x,y].back;
(2)X(t)[x,y].fore≦Y(t)[x,y].back≦X(t)[x,y].back;
(3)Y(t)[x,y].fore≦X(t)[x,y].fore≦Y(t)[x,y].back;
(4)Y(t)[x,y].fore≦X(t)[x,y].back≦Y(t)[x,y].back
の4つのいずれか少なくとも1つが成立すれば、動画Xと動画Yを重ねたときに、動画Xに描かれるオブジェクトiと動画Yに描かれるオブジェクトjとが衝突する、と判定することができる。
overlap(X,Y,t,x,y) = min〔X(t)[x,y].back,Y(t)[x,y].back〕-Y(t)[x,y].fore
であり、上記(2)の場合は、
overlap(X,Y,t,x,y) = Y(t)[x,y].back-max〔X(t)[x,y].fore,Y(t)[x,y].fore〕
であり、上記(3)の場合は、
overlap(X,Y,t,x,y) = min〔X(t)[x,y].back,Y(t)[x,y].back〕-X(t)[x,y].fore
であり、上記(4)の場合は、
overlap(X,Y,t,x,y) = X(t)[x,y].back-max〔X(t)[x,y].fore,Y(t)[x,y].fore〕
である。ここで、min〔...〕は、max〔...〕とは逆に、括弧内の値の最小値を返す。
(a)第1動画Xとして、現実世界においてダンサーがダンスをしている様子を撮影した2.5次元画像を採用し、
(b1)第2動画Yとして、キャラクターが無背景で踊っている様子を表す動画を採用したり、
(b2)第2動画Yとして、他のユーザがダンスをしている様子を撮影した後に当該他のユーザが編集を行い、背景の情報をすべて除去して、当該他のユーザが無背景で踊っている様子を表す動画を採用する。
= jを満たすものについて、上記の手法により、第2背面奥行Y(t)[x,y].backを求める。
= iを満たす位置[x,y]の集合であり、ある時刻tにおける第2描画領域とは、Y(t)[x,y].id = jを満たす位置[x,y]の集合である。これらの集合の共通部分が空集合でなければ、時刻tにおいて、第1描画領域と、第2描画領域と、が重なることになり、これらの集合の共通部分が時刻tにおける重複領域に相当する。
X(t)[x,y].id = i;
Y(t)[x,y].id = j
が成立する。
collision(Y,X) =Σt,x,y;X(t)[x,y].id≧0,Y(t)[x,y].id>0,overlap(X,Y,t,x,y)>0 1
によって計算が可能である。
collision(Y,X) =Σt,x,y;X(t)[x,y].id≧0,Y(t)[x,y].id>0,overlap(X,Y,t,x,y)>0 overlap(X,Y,t,x,y)
を採用することができる。このように計算される衝突量を評価関数とする。
yb(Y,t,j) = minx,y;Y(t)[x,y].id=j y;
xb(Y,t,j) = minx;Y(t)[x,yb(Y,t,j)].id=j x
X(t)[xb(Y,t,j),xb(Y,t,j)].id = 0;
Y(t)[xb(Y,t,j),xb(Y,t,j)].id = j;
X(t)[xb(Y,t,j),xb(Y,t,j)].fore =
Y(t)[xb(Y,t,j),xb(Y,t,j)].fore
が成立するはずである。
(a)動画Xからオブジェクトiのみを選択し、
(b)フレームに沿って(xc(X,t,i),yc(X,t,i))が原点になるように平行移動し、
(c)zc(X,t,i)/〔zc(X,t,i)+r〕倍に動画を拡大縮小し、
(d)フレームに沿って原点が(xc(X,t,i),yc(X,t,i))になるように平行移動する。
push'(i,r,X) = move(xc(X,t,i),yc(X,t,i),
scale(zc(X,t,i)/〔zc(X,t,i)+r〕,
move(-xc(X,t,i),-yc(X,t,i),
select(i,X))))
となる。
(a)動画Xからオブジェクトiのみを選択し、
(b)フレームに沿って(xc(X,t,i),yc(X,t,i))を(xc(X,t,i),zc(X,t,i)×cosθ-yc(X,t,i)×sinθ)へ移動し、
(c)zc(X,t,i)/〔zc(X,t,i)×sinθ+yc(X,t,i)×cosθ〕倍に動画を拡大縮小する。
rothor'(i,θ,X) =
scale(zc(X,t,i)/〔zc(X,t,i)×sinθ+yc(X,t,i)×cosθ〕,
move(0,zc(X,t,i)×cosθ-yc(X,t,i)×sinθ-yc(X,t,i),
select(i,X)))
となる。
rotver'(i,φ,X) =
scale(zc(X,t,i)/〔zc(X,t,i)×sinθ+xc(X,t,i)×cosθ〕,
move(zc(X,t,i)×cosθ-xc(X,t,i)×sinθ-xc(X,t,i),0
select(i,X)))
となる。
Y(t)[x,y] = X(t/a)[x,y]
が成立する。以下、この変換をfastforward(a,X)と表記する。
(1)2012年7月20日にされた日本国特許出願2012-161924
(2)2012年9月7日にされた特許協力条約に基づく国際出願PCT/JP2012/072988
(3)2012年9月7日にされた特許協力条約に基づく国際出願PCT/JP2012/072989
を基礎とする優先権を主張するものとし、指定国の法令が許す限り、当該基礎出願の内容を本願に取り込むものとする。
12 投影面
13 半直線
14 オブジェクト
15 衝突点
17 Z距離
21 代表点
101 動画処理装置
102 背面奥行取得部
103 衝突判定部
104 ずらし量設定部
105 動画生成部
106 修正部
401 画面
411 ウィンドウ
412 ウィンドウ
413 カーソル
Claims (11)
- 第1動画と、第2動画と、を合成する動画処理装置であって、前記第1動画には、第1オブジェクトが描かれ、前記第1オブジェクトの前面奥行情報を伴い、第2動画には、第2オブジェクトが描かれ、前記第2オブジェクトの前面奥行情報を伴い、
前記第1オブジェクトの背面奥行情報と、前記第2オブジェクトの背面奥行情報と、を取得する取得部、
前記第1動画に描かれた第1オブジェクトが占有しうる占有空間を、前記第1オブジェクトの前面奥行情報ならびに背面奥行情報を参照して求め、前記占有空間と、前記第2動画のあるフレームに描かれた前記第2オブジェクトと、が、干渉条件を満たすか否かを、前記第2オブジェクトの前面奥行情報ならびに背面奥行情報を参照して判定する干渉判定部、
前記干渉判定部による判定結果から、前記第2オブジェクトが前記占有空間に干渉せずに位置しうる可能範囲を設定する範囲設定部
を備えることを特徴とする動画処理装置。 - 請求項1に記載の動画処理装置であって、
前記取得部は、データベースから、前記第1オブジェクトの属性にあらかじめ対応付けられた第1奥行長を取得し、前記データベースから、前記第2オブジェクトの属性にあらかじめ対応付けられた第2奥行長を取得し、前記第1オブジェクトの前面奥行情報と前記第1奥行長とから前記第1オブジェクトの背面奥行情報を取得し、前記第2オブジェクトの前面奥行情報と前記第2奥行長とから前記第2オブジェクトの背面奥行情報を取得する
ことを特徴とする動画処理装置。 - 請求項1に記載の動画処理装置であって、
前記範囲設定部は、データベースから、前記第2オブジェクトの属性にあらかじめ対応付けられた前記第2オブジェクトの設置条件を取得し、前記可能範囲を、前記設置条件を満たすように設定する
ことを特徴とする動画処理装置。 - 請求項1に記載の動画処理装置であって、
前記範囲設定部は、前記第2動画のすべてのフレームについて、前記占有空間と、前記第2オブジェクトと、が前記干渉条件を満たさないように、前記可能範囲を設定する
ことを特徴とする動画処理装置。 - 請求項1に記載の動画処理装置であって、
前記第1動画と前記第2動画とのいずれか一方の動画には、3次元空間内の移動を表す非変形変換が施され、
前記可能範囲は、前記非変形変換に係る移動の量を表す変換パラメータがとりうる値の範囲により表現される
を備えることを特徴とする動画処理装置。 - 請求項5に記載の動画処理装置であって、
前記非変形変換は、水平方向、垂直方向のずらし量を前記変換パラメータとする平行移動であり、
前記一方の動画は、前記第2動画であり、
水平方向ならびに垂直方向のずらし量をユーザに修正させるため、前記範囲設定部により設定された水平方向ならびに垂直方向のずらし量の可能範囲と、前記第1動画のあるフレームと、が画面に表示される
ことを特徴とする動画処理装置。 - 請求項5に記載の動画処理装置であって、
前記非変形変換は、水平方向、垂直方向、奥行方向のずらし量を前記変換パラメータとする平行移動であり、
前記一方の動画は、前記第2動画であり、
水平方向、垂直方向、奥行方向のずらし量をユーザに修正させるため、前記範囲設定部により現在の奥行方向のずらし量に対して設定された水平方向ならびに垂直方向のずらし量の可能範囲と、前記第1動画のあるフレームと、が画面に表示される
ことを特徴とする動画処理装置。 - 請求項5に記載の動画処理装置であって、
前記可能範囲からいずれかの変換パラメータを選択して、前記非変形変換を施すことにより、前記一方の動画を修正する修正部、
前記第1動画と、前記第2動画と、を合成した第3動画を生成する動画生成部
をさらに備えることを特徴とする動画処理装置。 - 請求項8に記載の動画処理装置であって、
前記変換パラメータの初期値を設定する初期値設定部、
前記修正部は、前記設定された初期値が前記可能範囲に含まれなければ、前記可能範囲から、前記指定された初期値に最も近い変換パラメータを選択する
ことを特徴とする動画処理装置。 - 第1動画と、第2動画と、を合成する動画処理方法であって、前記第1動画には、第1オブジェクトが描かれ、前記第1オブジェクトの前面奥行情報を伴い、第2動画には、第2オブジェクトが描かれ、前記第2オブジェクトの前面奥行情報を伴い、
前記第1オブジェクトの背面奥行情報と、前記第2オブジェクトの背面奥行情報と、を取得する取得工程、
前記第1動画に描かれた第1オブジェクトが占有しうる占有空間を、前記第1オブジェクトの前面奥行情報ならびに背面奥行情報を参照して求め、前記占有空間と、前記第2動画のあるフレームに描かれた前記第2オブジェクトと、が、干渉条件を満たすか否かを、前記第2オブジェクトの前面奥行情報ならびに背面奥行情報を参照して判定する干渉判定工程、
前記干渉判定工程における判定結果から、前記第2オブジェクトが前記占有空間に干渉せずに位置しうる可能範囲を設定する範囲設定工程
を備えることを特徴とする動画処理方法。 - 第1動画と、第2動画と、を合成するプログラムであって、前記第1動画には、第1オブジェクトが描かれ、前記第1オブジェクトの前面奥行情報を伴い、第2動画には、第2オブジェクトが描かれ、前記第2オブジェクトの前面奥行情報を伴い、前記プログラムは、コンピュータを、
前記第1オブジェクトの背面奥行情報と、前記第2オブジェクトの背面奥行情報と、を取得する取得部、
前記第1動画に描かれた第1オブジェクトが占有しうる占有空間を、前記第1オブジェクトの前面奥行情報ならびに背面奥行情報を参照して求め、前記占有空間と、前記第2動画のあるフレームに描かれた前記第2オブジェクトと、が、干渉条件を満たすか否かを、前記第2オブジェクトの前面奥行情報ならびに背面奥行情報を参照して判定する干渉判定部、
前記干渉判定部による判定結果から、前記第2オブジェクトが前記占有空間に干渉せずに位置しうる可能範囲を設定する範囲設定部
として機能させることを特徴とするプログラムを記録したコンピュータ読取可能な情報記録媒体。
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
ES12881371.4T ES2675514T3 (es) | 2012-07-20 | 2012-10-01 | Dispositivo de procesamiento de imágenes en movimiento, método de procesamiento de imágenes en movimiento, y medio de registro de información |
JP2013509348A JP5270052B1 (ja) | 2012-07-20 | 2012-10-01 | 動画処理装置、動画処理方法、ならびに、情報記録媒体 |
EP12881371.4A EP2775454B1 (en) | 2012-07-20 | 2012-10-01 | Moving-image processing device, moving-image processing method, and information recording medium |
US14/361,477 US9723225B2 (en) | 2012-07-20 | 2012-10-01 | Moving-image processing device, moving-image processing method, and information recording medium |
Applications Claiming Priority (6)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2012-161924 | 2012-07-20 | ||
JP2012161924 | 2012-07-20 | ||
PCT/JP2012/072988 WO2014013627A1 (ja) | 2012-07-20 | 2012-09-07 | 動画処理装置、動画処理方法、ならびに、情報記録媒体 |
JPPCT/JP2012/072989 | 2012-09-07 | ||
JPPCT/JP2012/072988 | 2012-09-07 | ||
PCT/JP2012/072989 WO2014013628A1 (ja) | 2012-07-20 | 2012-09-07 | 動画処理装置、動画処理方法、ならびに、情報記録媒体 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2014013630A1 true WO2014013630A1 (ja) | 2014-01-23 |
Family
ID=49948477
Family Applications (4)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2012/072988 WO2014013627A1 (ja) | 2012-07-20 | 2012-09-07 | 動画処理装置、動画処理方法、ならびに、情報記録媒体 |
PCT/JP2012/072989 WO2014013628A1 (ja) | 2012-07-20 | 2012-09-07 | 動画処理装置、動画処理方法、ならびに、情報記録媒体 |
PCT/JP2012/075416 WO2014013629A1 (ja) | 2012-07-20 | 2012-10-01 | 動画処理装置、動画処理方法、ならびに、情報記録媒体 |
PCT/JP2012/075417 WO2014013630A1 (ja) | 2012-07-20 | 2012-10-01 | 動画処理装置、動画処理方法、ならびに、情報記録媒体 |
Family Applications Before (3)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2012/072988 WO2014013627A1 (ja) | 2012-07-20 | 2012-09-07 | 動画処理装置、動画処理方法、ならびに、情報記録媒体 |
PCT/JP2012/072989 WO2014013628A1 (ja) | 2012-07-20 | 2012-09-07 | 動画処理装置、動画処理方法、ならびに、情報記録媒体 |
PCT/JP2012/075416 WO2014013629A1 (ja) | 2012-07-20 | 2012-10-01 | 動画処理装置、動画処理方法、ならびに、情報記録媒体 |
Country Status (5)
Country | Link |
---|---|
US (4) | US9819878B2 (ja) |
EP (4) | EP2779107B1 (ja) |
JP (1) | JP5451955B1 (ja) |
ES (4) | ES2676719T3 (ja) |
WO (4) | WO2014013627A1 (ja) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11450352B2 (en) | 2018-05-29 | 2022-09-20 | Sony Corporation | Image processing apparatus and image processing method |
Families Citing this family (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9429912B2 (en) * | 2012-08-17 | 2016-08-30 | Microsoft Technology Licensing, Llc | Mixed reality holographic object development |
US9823824B2 (en) * | 2013-08-19 | 2017-11-21 | Kodak Alaris Inc. | Context sensitive adaptable user interface |
US10311595B2 (en) * | 2013-11-19 | 2019-06-04 | Canon Kabushiki Kaisha | Image processing device and its control method, imaging apparatus, and storage medium |
JP6427332B2 (ja) * | 2014-04-08 | 2018-11-21 | 株式会社ミツトヨ | 画像測定機 |
JP2015207802A (ja) * | 2014-04-17 | 2015-11-19 | ソニー株式会社 | 画像処理装置および画像処理方法 |
EP3073726B1 (en) * | 2015-03-24 | 2017-03-01 | Axis AB | Method for configuring a camera |
US10600169B2 (en) * | 2015-03-26 | 2020-03-24 | Sony Corporation | Image processing system and image processing method |
US20170068863A1 (en) * | 2015-09-04 | 2017-03-09 | Qualcomm Incorporated | Occupancy detection using computer vision |
CN105243268B (zh) * | 2015-09-18 | 2018-06-22 | 网易(杭州)网络有限公司 | 一种游戏地图定位方法、装置及用户终端 |
GB2550854B (en) | 2016-05-25 | 2019-06-26 | Ge Aviat Systems Ltd | Aircraft time synchronization system |
CN106095294B (zh) * | 2016-06-08 | 2019-03-05 | 福建天泉教育科技有限公司 | 不同平台之间同步笔画的方法及其系统 |
JP6488329B2 (ja) * | 2016-09-29 | 2019-03-20 | 株式会社リクルート | 順番管理システム、順番管理装置、およびプログラム |
US10609332B1 (en) | 2018-12-21 | 2020-03-31 | Microsoft Technology Licensing, Llc | Video conferencing supporting a composite video stream |
CN112232170B (zh) * | 2020-10-10 | 2024-08-02 | 浙江大华技术股份有限公司 | 对象行为的确定方法及装置、存储介质、电子装置 |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH07322137A (ja) * | 1994-05-30 | 1995-12-08 | Nec Eng Ltd | 映像信号合成装置 |
JP2004145448A (ja) * | 2002-10-22 | 2004-05-20 | Toshiba Corp | 端末装置、サーバ装置および画像加工方法 |
JP2009080660A (ja) | 2007-09-26 | 2009-04-16 | Rakuten Inc | 物体領域抽出処理プログラム、物体領域抽出装置、および物体領域抽出方法 |
Family Cites Families (37)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5087981A (en) * | 1990-01-02 | 1992-02-11 | Eastman Kodak Company | Error diffusion of overlapping dots |
JP3244798B2 (ja) * | 1992-09-08 | 2002-01-07 | 株式会社東芝 | 動画像処理装置 |
US6151009A (en) * | 1996-08-21 | 2000-11-21 | Carnegie Mellon University | Method and apparatus for merging real and synthetic images |
US6456289B1 (en) * | 1999-04-23 | 2002-09-24 | Georgia Tech Research Corporation | Animation system and method for a animating object fracture |
US6738066B1 (en) * | 1999-07-30 | 2004-05-18 | Electric Plant, Inc. | System, method and article of manufacture for detecting collisions between video images generated by a camera and an object depicted on a display |
DE10106023A1 (de) * | 2001-02-09 | 2002-08-29 | Fraunhofer Ges Forschung | Verfahren und Vorrichtung zur Kollisionserkennung von Objekten |
US6791549B2 (en) | 2001-12-21 | 2004-09-14 | Vrcontext S.A. | Systems and methods for simulating frames of complex virtual environments |
US7982738B2 (en) * | 2004-12-01 | 2011-07-19 | Microsoft Corporation | Interactive montages of sprites for indexing and summarizing video |
US7788071B2 (en) * | 2004-12-03 | 2010-08-31 | Telekinesys Research Limited | Physics simulation apparatus and method |
JP4335160B2 (ja) * | 2005-03-02 | 2009-09-30 | 任天堂株式会社 | 衝突判定プログラムおよび衝突判定装置 |
CN101228549B (zh) * | 2005-03-31 | 2012-10-17 | 本田技研工业株式会社 | 文件管理系统、文件管理方法 |
JP4345737B2 (ja) * | 2005-11-14 | 2009-10-14 | セイコーエプソン株式会社 | デジタルコンテンツ作成システム、デジタルコンテンツ作成方法およびプログラム |
JP4754364B2 (ja) * | 2006-01-20 | 2011-08-24 | 三菱電機株式会社 | 画像重ね合わせ装置 |
KR100718157B1 (ko) * | 2006-03-08 | 2007-05-14 | 삼성전자주식회사 | 충돌 감지 장치 및 방법 |
US8016653B2 (en) * | 2007-02-01 | 2011-09-13 | Sportvision, Inc. | Three dimensional virtual rendering of a live event |
TW200844899A (en) * | 2007-05-04 | 2008-11-16 | Sunplus Mmobile Inc | An interactive image and graphical system and method that is capable of detecting collisions of graphical objects |
JP5253936B2 (ja) * | 2008-09-08 | 2013-07-31 | 株式会社エヌ・ティ・ティ・ドコモ | 情報処理装置及びプログラム |
US8730245B2 (en) * | 2008-12-01 | 2014-05-20 | Naturalmotion Ltd. | Defining an animation of a virtual object within a virtual world |
GB0901351D0 (en) * | 2009-01-28 | 2009-03-11 | Univ Dundee | System and method for arranging items for display |
TWI378718B (en) * | 2009-06-05 | 2012-12-01 | Univ Nat Taiwan | Method for scaling video content according to bandwidth rate |
DE102009037835B4 (de) * | 2009-08-18 | 2012-12-06 | Metaio Gmbh | Verfahren zur Darstellung von virtueller Information in einer realen Umgebung |
US8749588B2 (en) * | 2009-09-15 | 2014-06-10 | HNTB Holdings, Ltd. | Positioning labels in an engineering drawing |
US20110199302A1 (en) | 2010-02-16 | 2011-08-18 | Microsoft Corporation | Capturing screen objects using a collision volume |
US20110316972A1 (en) * | 2010-06-29 | 2011-12-29 | Broadcom Corporation | Displaying graphics with three dimensional video |
US20120044259A1 (en) * | 2010-08-17 | 2012-02-23 | Apple Inc. | Depth management for displayed graphical elements |
US8860766B2 (en) * | 2010-09-30 | 2014-10-14 | Nvidia Corporation | System, method, and computer program product for determining one or more contact points between a pair of objects |
JP5643617B2 (ja) | 2010-11-18 | 2014-12-17 | 任天堂株式会社 | 画像処理プログラム、画像処理装置、画像処理方法、画像処理システム |
US8401225B2 (en) * | 2011-01-31 | 2013-03-19 | Microsoft Corporation | Moving object segmentation using depth images |
JP5602040B2 (ja) | 2011-02-03 | 2014-10-08 | ミツカワ株式会社 | 積層体布帛 |
JP5678710B2 (ja) * | 2011-02-14 | 2015-03-04 | ソニー株式会社 | 画像処理装置、および画像処理方法、並びにプログラム |
JP5708196B2 (ja) * | 2011-04-21 | 2015-04-30 | セイコーエプソン株式会社 | 衝突検出システム、ロボットシステム、衝突検出方法及びプログラム |
US10109097B2 (en) | 2011-08-01 | 2018-10-23 | Autodesk, Inc. | Dynamic collision avoidance for crowd simulation over structured paths that intersect at waypoints |
US20130215230A1 (en) * | 2012-02-22 | 2013-08-22 | Matt Miesnieks | Augmented Reality System Using a Portable Device |
US9183676B2 (en) * | 2012-04-27 | 2015-11-10 | Microsoft Technology Licensing, Llc | Displaying a collision between real and virtual objects |
US9147221B2 (en) * | 2012-05-23 | 2015-09-29 | Qualcomm Incorporated | Image-driven view management for annotations |
US9741145B2 (en) | 2012-06-29 | 2017-08-22 | Disney Enterprises, Inc. | Augmented reality simulation continuum |
JP2014238731A (ja) * | 2013-06-07 | 2014-12-18 | 株式会社ソニー・コンピュータエンタテインメント | 画像処理装置、画像処理システム、および画像処理方法 |
-
2012
- 2012-09-07 ES ES12881432.4T patent/ES2676719T3/es active Active
- 2012-09-07 ES ES12881218.7T patent/ES2675513T3/es active Active
- 2012-09-07 WO PCT/JP2012/072988 patent/WO2014013627A1/ja active Application Filing
- 2012-09-07 WO PCT/JP2012/072989 patent/WO2014013628A1/ja active Application Filing
- 2012-09-07 EP EP12881218.7A patent/EP2779107B1/en active Active
- 2012-09-07 US US14/361,332 patent/US9819878B2/en active Active
- 2012-09-07 EP EP12881432.4A patent/EP2775452B1/en active Active
- 2012-09-07 JP JP2013552056A patent/JP5451955B1/ja active Active
- 2012-09-07 US US14/360,840 patent/US9876965B2/en active Active
- 2012-10-01 ES ES12881371.4T patent/ES2675514T3/es active Active
- 2012-10-01 US US14/361,477 patent/US9723225B2/en active Active
- 2012-10-01 EP EP12881371.4A patent/EP2775454B1/en active Active
- 2012-10-01 US US14/361,028 patent/US9374535B2/en active Active
- 2012-10-01 EP EP12881482.9A patent/EP2775455B1/en active Active
- 2012-10-01 WO PCT/JP2012/075416 patent/WO2014013629A1/ja active Application Filing
- 2012-10-01 ES ES12881482.9T patent/ES2673545T3/es active Active
- 2012-10-01 WO PCT/JP2012/075417 patent/WO2014013630A1/ja active Application Filing
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH07322137A (ja) * | 1994-05-30 | 1995-12-08 | Nec Eng Ltd | 映像信号合成装置 |
JP2004145448A (ja) * | 2002-10-22 | 2004-05-20 | Toshiba Corp | 端末装置、サーバ装置および画像加工方法 |
JP2009080660A (ja) | 2007-09-26 | 2009-04-16 | Rakuten Inc | 物体領域抽出処理プログラム、物体領域抽出装置、および物体領域抽出方法 |
Non-Patent Citations (1)
Title |
---|
"Article on "2.5D"", WIKIPEDIA, Retrieved from the Internet <URL:http://en.wikipedia.org/wiki/2.5D> [retrieved on 20120718] |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11450352B2 (en) | 2018-05-29 | 2022-09-20 | Sony Corporation | Image processing apparatus and image processing method |
Also Published As
Publication number | Publication date |
---|---|
WO2014013627A1 (ja) | 2014-01-23 |
ES2675514T3 (es) | 2018-07-11 |
EP2775452A1 (en) | 2014-09-10 |
JP5451955B1 (ja) | 2014-03-26 |
EP2775454A1 (en) | 2014-09-10 |
EP2775452B1 (en) | 2018-05-02 |
EP2779107A4 (en) | 2015-03-11 |
US20140321736A1 (en) | 2014-10-30 |
ES2676719T3 (es) | 2018-07-24 |
WO2014013628A1 (ja) | 2014-01-23 |
JPWO2014013628A1 (ja) | 2016-06-30 |
EP2779107A1 (en) | 2014-09-17 |
EP2775455A1 (en) | 2014-09-10 |
ES2675513T3 (es) | 2018-07-11 |
US9723225B2 (en) | 2017-08-01 |
EP2775454A4 (en) | 2015-04-01 |
EP2775454B1 (en) | 2018-05-09 |
EP2775452A4 (en) | 2015-03-04 |
EP2775455B1 (en) | 2018-05-09 |
ES2673545T3 (es) | 2018-06-22 |
EP2779107B1 (en) | 2018-05-16 |
WO2014013629A1 (ja) | 2014-01-23 |
US20150201132A1 (en) | 2015-07-16 |
US9876965B2 (en) | 2018-01-23 |
EP2775455A4 (en) | 2015-04-22 |
US20140340477A1 (en) | 2014-11-20 |
US20140347560A1 (en) | 2014-11-27 |
US9819878B2 (en) | 2017-11-14 |
US9374535B2 (en) | 2016-06-21 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP5451955B1 (ja) | 動画処理装置、動画処理方法、ならびに、情報記録媒体 | |
JP6950013B2 (ja) | 画像処理装置、画像処理方法及びプログラム | |
US20130113701A1 (en) | Image generation device | |
KR20070119018A (ko) | 3d 카메라 및 3d비디오를 위한 자동 씬 모델링 | |
KR20200043458A (ko) | 3d 장면의 이미지를 생성 및 수정하기 위한 방법 | |
WO2018051747A1 (ja) | 画像処理装置、画像生成方法及びプログラム | |
CN113347373B (zh) | 通过ar空间定位实时制作特效视频的图像处理方法 | |
JP2022077148A (ja) | 画像処理方法、プログラムおよび画像処理システム | |
JP6089145B2 (ja) | カメラワーク生成方法、カメラワーク生成装置及びカメラワーク生成プログラム | |
JP5254505B1 (ja) | 動画処理装置、動画処理方法、ならびに、情報記録媒体 | |
KR20190074911A (ko) | 실감형 영상 콘텐츠 제공 방법 및 이를 이용한 서버 | |
JP5270052B1 (ja) | 動画処理装置、動画処理方法、ならびに、情報記録媒体 | |
JP5270051B1 (ja) | 動画処理装置、動画処理方法、ならびに、情報記録媒体 | |
US20240233297A1 (en) | Image processing apparatus, image processing method, and computer-readable storage medium | |
Albuquerque et al. | An overview on virtual sets | |
JP2014078861A (ja) | 表示制御装置 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
ENP | Entry into the national phase |
Ref document number: 2013509348 Country of ref document: JP Kind code of ref document: A |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 12881371 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 14361477 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2012881371 Country of ref document: EP |
|
NENP | Non-entry into the national phase |
Ref country code: DE |