JPH04326181A - Picture processor - Google Patents

Abstract

PURPOSE:To offer a moving object indication system by which the moving object in a moving image can be easily indicated. CONSTITUTION:A frame image constituting the moving image is successively inputted through an input terminal 301. Then, the previously inputted frame image is delayed for one frame by a frame delay equipment 303, so that the frame image is synchronized with the frame image inputted at that time. Then, a difference between the two frame images is searched by a difference equipment 305, and the result is stored in a differential buffer 306. Afterwards, an area computing element 307 searches the data stored in the differential buffer 306, and detects the existing area of the moving object. The detected existing area of the moving object is compared with an indicated coordinate position in order to judge whether or not the indicated position indicates the moving object.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus, and more particularly to an image processing apparatus for indicating a moving object in a displayed moving image.

0002

2. Description of the Related Art Database systems for storing and retrieving information have been used as a means for efficiently utilizing existing information. Many of these database systems are relational database systems in which stored information is searched by keywords, and there are problems in that the functions, capabilities, and scale of the system are largely influenced by the method of registering and operating keywords.

In recent years, alternative object-oriented database systems have been proposed that link stored information more dynamically and directly, and are commonly known as hypertext or hypermedia.

[0004] These systems link related information to all or part of a text image or a still image.

[0005] The object to which this link is created is generally called an object.

As described above, image information database systems have conventionally focused on still images. but,
The development of inexpensive large-capacity storage devices such as optical disks and magneto-optical disks, and the development of high-speed MPUs (Micro Pro
cessingUnito), DSP (Digital
L for calculations such as Signal Processor)
Along with the development of SI and highly efficient video encoding methods, the development of database systems that can also handle video images is progressing.

[0007] Such a database system is generally called a multimedia database, and although it employs the object-oriented database formation described above, the scope of abstractions is limited to text images and still images.

[0008]

[Problem to be solved by the invention] Regarding moving images, if a frame unit or a set of frames (i.e., a scene) is defined as an object for the entire image, it is possible to create links between each other and between text images and still images. is possible.

[0009] However, moving objects in moving images are generally people or animals, and in addition to spatial movement, they also involve their own movements, that is, changes in the shape of the object. The challenge is how to define and operate it as such.

[0010] The present invention has been made in view of this problem, and an object thereof is to provide an image processing device that makes it possible to easily indicate a moving object in a moving image.

[0011]

[Means for Solving the Problems] In order to solve this problem, the present invention has the configuration shown below. That is, an input means for sequentially inputting individual frame images constituting a moving image, a detection means for detecting an area where a moving object exists in a frame image by comparing input multiple frame images, and an input means for indicating a desired position. and a determining means for determining whether or not the position designated by the pointing means is within the area detected by the detecting means.

0012

[Operation] In the configuration of the present invention, the area where the moving object exists is detected by comparing adjacent moving image frames that are sequentially input by the input means. Then, by determining whether or not the position indicated by the indicating means is within the detected area, it is possible to identify whether or not the instruction indicates a moving object.

[0013]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

The concept of an object in a moving image in this embodiment will be explained with reference to FIG. In the figure, 101, 1
04 and 107 indicate certain frames in the moving image file, and indicate the n-m-th, n-th, and n+m-th frames, respectively. Also, 102 in each frame
, 105 and 108 are target animals. 103, 106
, 109 is a region containing the target moving object, which is an object defined according to this embodiment.

FIG. 2 is a diagram for explaining the relationship between a target moving body and an object. In the figure, 201 is a certain frame image, and 202 is a target moving object in the frame image. Reference numeral 203 denotes an attribute frame for recording attached information of a frame image, which is attached to each frame image;
04 indicates the area of the object recorded on the attribute frame 203.

Therefore, when actually operating on a multimedia system, the point indicated by a pointing means such as a mouse is the object 20 in the attribute frame 203.
If it is within the area indicated by 4, it is determined that the target moving object 202 has been designated. In other words, if the area indicated by the object 204 is specified, it is determined that the target moving object 202 is not specified.

FIG. 3 shows the configuration of an apparatus related to object calculation in this embodiment. In the figure, 301 is an input terminal for inputting moving image data, 302 is an A/D converter for converting the input analog image into a digital image, and 3
03 and 304 are frame delayers for one frame each, and are generally DRAM (Dynamic Random DRAM).
m Access Memory). 305 is a difference device that calculates a difference in luminance, for example, between pixels of two adjacent frames; 306 is a difference device 30;
This is a difference buffer that temporarily stores the results for one frame of 5. 307 is an area calculator that calculates an area to be defined as an object based on the data in the difference buffer 306;
308 is an attribute memory that temporarily stores the result of the area calculator 307, and 309 is a terminal that outputs image data delayed by two frames, and the object area is calculated during this two frame delay. . 310 is a terminal that outputs the contents of the attribute memory 308;
The contents of attribute memory 308 are output in synchronization with image data 309 output from.

The flow of processing in the above configuration will be explained with reference to FIG.

In FIG. 4A, 401 is a frame image at a certain time, and 402 is a target moving object in the frame image 401. In FIG. Similarly, 403 in FIG. 4(b)
is a frame image adjacent to (following) the frame image 401, and 404 is a target moving object in the frame image 403. In other words, assume that the frame image 401 is first input from the input terminal, and then the frame image 403 is input. Therefore, as a matter of course, the target animal 4
04 is the same as the target animal 402.

Now, at this time, the frame image 401 will be stored in the frame delay unit 303, so the frame image 401 and the frame image 403 that has just been input will be input to the subtractor 305. Both of these frame images 401 and 403 are input to the subtractor 305. In reality, each frame image is input pixel by pixel in synchronization with a pixel clock (not shown), but the subtractor 305 calculates the difference between the two input pixel data (luminance in the case of a monochrome image). A process is performed to calculate and write it to the corresponding position of the difference buffer 306.

Since the difference in the area other than the target object is essentially "0", the difference buffer 306 ends up containing data 405 (pixel distribution with values above a certain level) as shown in FIG. 4(c). will be constructed. In other words, by taking the difference between adjacent frames that make up a video image,
It becomes possible to detect moving objects.

Now, the area calculator 307 examines the data stored in the difference buffer 306, and calculates the top and bottom points in the vertical direction and the rightmost and leftmost points in the horizontal direction of the area where the difference is large. Detect points. That is, a rectangular area 412 specified by points 408, 409, 410, and 411 in FIG. 4(d), that is, an area to be an object is detected. Data representing the detected object area is stored in the attribute memory 308 and sent to the system side (not shown). The system side, for example, compares the coordinate position indicated by the mouse with data indicating the object area, determines whether or not the point indicated by the mouse points to an object, and performs the corresponding processing. . Note that the means for specifying the coordinate position may be any other means, and whether or not the specified coordinate position is within the detected area can be determined by a simple comparison.
I will omit the explanation here.

[Description of Other Embodiments] In the present embodiment described above, the object area is defined as a circumscribed rectangle, but the object area is not limited to this.

FIG. 5 illustrates another example of defining an object area. In the figure, 501 is a frame image, and 502 is a target moving object in the frame image.

Now, FIG. 5(a) shows the area definition by the circumscribed circle, and 503 is the result.

Similarly, FIG. 5(b) shows area definition using circumscribed polygons, and 504 is the result.

Further, FIG. 5(c) shows the area definition by the maximum inscribed circle centered on the center of gravity of the target animal body, and FIG.
d) shows region definition by the largest inscribed quadrilateral centered on the center of gravity.

The reason why the center of gravity is set as the center point here is based on the idea that when a human follows the line of sight of an animal, he or she is following the center of gravity of the animal.

Further, although the difference between frame images is calculated between two frames, the difference is not limited to this, and if the difference is calculated between three frames including the front and rear of the target frame, an area more adapted to movement can be defined as an object. Furthermore, if the calculation is performed not between adjacent frames but between frames several frames apart, it is possible to define a wider area as an object. However, compared to this embodiment, the hardware scale is larger.

In this embodiment, the attribute memory is used to store the object area, but the invention is not limited to this, and it can also be realized by inserting it in a flag format into a part of the image data. It is also possible to implement a method of managing the data in a batch completely independently from the data and reading it out at the appropriate time.

Furthermore, although this embodiment has been described using a bitmap method in which the object area spatially corresponds to image data, the object area is not limited to this, and it is also possible to have the object area as vector data.

Furthermore, in this embodiment, a method of synchronizing with image data was explained when calculating the object area, but this is not always the case in the case of a system that supplies image data using package media (tape, optical disk, magnetic disk, etc.). There is no need to synchronize; area calculations can be performed individually and synchronized during packaging.

[0033]Although the present invention is realized by hardware, it goes without saying that it can also be applied to a case where it is achieved by supplying a program to an apparatus.

As explained above, according to this embodiment, when a moving object in a moving image is defined as an object, a wider area including the moving object itself and the moving range is defined as an object. It becomes possible to reliably and easily point out the animal body inside.

[0035]

As described above, according to the present invention, it becomes possible to easily indicate a moving object in a moving image.

[Brief explanation of drawings]

FIG. 1 is a diagram for explaining the concept of an object in a moving image in an embodiment.

FIG. 2 is a diagram illustrating the relationship between a target moving body and an object in an example.

FIG. 3 is a diagram showing a specific block configuration of this embodiment.

FIG. 4 is a diagram showing the process of object area detection in the embodiment.

FIG. 5 is a diagram illustrating another principle of object area detection in the embodiment.

[Explanation of symbols]

101 Frame image 102 Target animal 103 Object area 301 Input terminal 302 A/D converter 303 and 304 Frame delay device 305 Differentiator 306 Difference buffer 307 Area calculator 308 Attribute memory

Claims (1)

[Claims] 1. An input means for sequentially inputting individual frame images constituting a moving image, a detection means for detecting an area where a moving object exists in a frame image by comparing the input plurality of frame images, An image processing apparatus comprising an instruction means for indicating a position, and a determination means for determining whether the position specified by the instruction means is within the area detected by the detection means.