Browse Prior Art Database

Method of Dissolve Detection in Moving Picture

IP.com Disclosure Number: IPCOM000108666D
Original Publication Date: 1992-Jun-01
Included in the Prior Art Database: 2005-Mar-22
Document File: 4 page(s) / 113K

Publishing Venue

IBM

Related People

Ioka, M: AUTHOR

Abstract

This article describes a method of detecting dissolve scene changes in moving pictures. In dissolving scene changes, fade-out and fade-in occur simultaneously. And it is very important to detect these kinds of changes, because it is often used in order to change scenes effectively.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Method of Dissolve Detection in Moving Picture

       This article describes a method of detecting dissolve
scene changes in moving pictures.  In dissolving scene changes,
fade-out and fade-in occur simultaneously.  And it is very important
to detect these kinds of changes, because it is often used in order
to change scenes effectively.

      Fig. 1 shows the task flow for detecting dissolve scene
changes.
1.   One frame is input.  In this step, the inputted frame is
transformed analog signal to digital data.
2.   If there are no more frames, execution is terminated.
3.   It is not necessary to calculate all the pixels in the frame.
Pixels are subsampled.  Rectangular sampling is adopted.
4.   A moving picture usually consists of color frame sequences.  The
transformed digital image may be in RGB, YIQ or some other format.
In this disclosure, the intensity of the frame is adopted as basic
information. Thus, if the transformed image is in RGB style, its
intensity value is extracted from its RGB value.
5.   The frame difference (FD) between frame t and frame t-1 at each
subsampled point is calculated by the following equation:
      FD(i,j,t) = I(i,j,t) - I(i,j,t-1)
      where I(i,j,t): intensity of pixel(i,j) at frame t.
6.   The FD(i,j,t) is divided by the previously calculated
FD(i,j,t-1) at each corresponding subsampled point. The ratio of
pixels, which show positive value in procedure 5), to the number of
all subsampled pixels is calculated.  The series of the ratio is
plotted as shown in Fig. 2.
7.   The series of the calculated ratio is filtered by the median
filter. The series of the filtered ratio is plotted as shown in Fig.
3.
8.   Whether the current frame is located at local maximum or minimum
in the series of the filtered ratio (procedure 5)-7)) is checked.  If
the following e...