JP2008084171A - Method for improving sense of dynamic depth - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method for improving the sense of dynamic depth, improving the sense of depth by detecting a mobile body in a frame by comparing a plurality of frames in a moving image, and highlighting the circumferential shadow of the mobile body, thereby separating the mobile body from the background. <P>SOLUTION: In response to input of an image, a frame concerned is compared with an internally retained previous frame, motion vectors are detected in a pixel unit, and adjacent pixels having the same motion vector quantity are determined to be the same mobile body. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention detects a moving object in a frame by comparing a plurality of frames in a moving image, and enhances a sense of depth by separating a background and a moving object by enhancing shadows around the corresponding moving object. The present invention relates to a method for improving a dynamic depth feeling characterized by:

  Conventionally, there has been a method of using shading to express the stereoscopic effect of a still image, for example, a method of shading in order to display a building, a three-dimensional intersection, or the like in a three-dimensional manner.

For example, Patent Document 1 describes these methods.
JP 2003-141555 A

  However, the above-described conventional technique is not characterized by the presentation of a sense of depth by separating the moving object and the background. In general, shading is a method for shading a model based on the shape of the light source and the model.

  The present invention has been made in consideration of the above-described circumstances, and dynamic depth sensation that enables separation of moving objects and the background by enhancing shadows in units of moving objects, not in units of pixels and blocks, thereby improving the feeling of depth. An object is to provide an improvement method.

  In order to solve the above problem, in the dynamic depth enhancing method according to the first aspect of the present invention, a plurality of video frames constituting a moving image composed of a background image and a moving object are compared, and the moving object in the video frame is detected. The moving object is displayed three-dimensionally from the background image by enhancing the shadow of the boundary portion between the detected moving object and the background image, thereby improving the sense of depth of the moving image.

  Further, in the dynamic depth sensation improving method according to the second aspect of the present invention, the enhancement processing is performed on a peripheral portion of the moving body.

  Furthermore, in the dynamic depth improving method of the invention according to claim 3, the enhancement processing is performed on an inner portion of the moving body.

  According to the present invention, focusing on a moving object (moving object) in an image, separating the moving object from the background, and emphasizing shadows (luminance change), the moving object is lifted in the image and provides a sense of depth. can do.

  FIG. 1 shows an image when a conventional frame is displayed. When 101 is the background of the corresponding frame and 102 is the target moving object, 103 indicates the luminance value of the corresponding frame in a graph. For comparison, stepwise shadow processing as shown in FIG. 2 is not performed.

  FIG. 2 shows an image diagram of the present invention. 201 is the background of the corresponding frame as in 101, and 202 is a moving object as in 102. When the background 204 adjacent to 202 is changed in brightness more sharply than 201, the shadow is emphasized, and the moving object 202 is set as the background. It is characterized by a sense of depth improvement by separating from 201. In FIG. 2, the background 204 schematically shows the stage of shading, but actually it may be a gentle display as shown in the graph 203.

  FIG. 3 shows an example of the operation flow of the present invention. In the present embodiment, it is assumed that there is only one moving object in the corresponding frame. Here, the present invention is not compatible with only one moving object, and even when there are a plurality of moving objects, the moving object can be selected and detected by means such as setting a threshold value of a motion vector. The same processing as in the figure is possible.

  In this flowchart, first, an input of a video is received (S1), and the immediately preceding frame held inside is compared with the corresponding frame (S2), for example, a motion vector in pixel units is detected, and the motion of adjacent pixels Moving object detection is performed by determining that the same vector amount is the same moving object (S3).

  When the moving object detection is completed, the corresponding moving object and the surrounding background are separated and each luminance value is detected (S4). Next, a brightness value for shadow enhancement is arbitrarily determined (S5), and a depth improvement process characterized by separating a moving object and a background as shown in the following example is performed (S6), and an image display is performed. Perform (S7).

  An example of the depth feeling improving process will be described with reference to FIG. In this example, the moving object and the background are assumed to be separated, and an example of drawing a circle is used. However, in the present invention, this patent is applied not only to drawing of CGI etc. but also to images of TV broadcasting etc. It is also possible to apply.

  As an example, as shown in FIG. 2, the periphery of 202 is displayed in black as 204, and the shadow is emphasized to display 202 as if it was lifted from 201. When this luminance value is expressed in a graph, it becomes 203, and when compared with 103, it can be confirmed that the portion 205 sharply falls and emphasizes the shadow.

  As an example of a function for expressing this, the luminance value of 202 is Y0, the luminance value of 201 is Y1, the luminance value of 204 for shadow enhancement is Y2, the radius of 202 is r, and the center coordinate of 202 is ( If x, y) and the density of the shadow are α and the steepness of the shadow is β, the condition of f (d) is as follows.

(1) When d increases to ∞, f (d) = Y1
(2) When d approaches r from ∞, f (d) = Y2 (d)
f (d) = Y1 (1 + ((α-1) + (1-α) (Y0 / Y1)) (r / d) ^ β) (r ≤ d)
Can be expressed as

  In the above example, the shadow around the moving object is emphasized, but the depth feeling can be improved by enhancing the shadow on the inner side of the moving object.

  As described above, a moving object in a frame is detected by comparing a plurality of frames in a moving image, and a shadow around the corresponding moving object is emphasized, thereby separating the background and the moving object and improving the sense of depth. It is possible to provide a method for improving the dynamic depth feeling.

Explanatory drawing at the time of displaying the conventional flame | frame. Explanatory drawing which shows the dynamic depth feeling improvement method which concerns on embodiment of this invention. The flowchart which shows the dynamic depth feeling improvement method which concerns on embodiment of this invention.

Explanation of symbols

201 ... Background 202 ... Moving object 203 ... Graph 204 ... Background showing shading

Claims (3)

By comparing a plurality of video frames constituting a moving image composed of a background image and a moving object, detecting the moving object in the video frame, and emphasizing a shadow of a boundary portion between the detected moving object and the background image, A dynamic depth feeling improving method, wherein the moving object is three-dimensionally displayed from the background image to improve the depth feeling of the moving image. The dynamic depth improving method according to claim 1, wherein the enhancement processing is performed on a peripheral portion of the moving body. The dynamic depth improving method according to claim 1, wherein the enhancement processing is performed on an inner portion of the moving body.