Television – Format conversion
Reexamination Certificate
1998-09-25
2001-05-08
Eisenzopf, Reinhard (Department: 2614)
Television
Format conversion
C348S699000, C348S416100, C348S445000, C348S459000, C375S240160
Reexamination Certificate
active
06229570
ABSTRACT:
BACKGROUND OF THE INVENTION
I. Field of the Invention
The present invention relates to the processing of source video signals for use with high definition televisions (HDTV). More particularly, the present invention pertains to a method of converting an existing source video signal to a signal having an increased frequency rate (number of frames per second) for use with HDTV.
II. Description of the Related Art
Presently available source material like motion-picture film is recorded at a temporal rate of 24 frames/second and video source material is recorded at a rate of 30 frames/second. This temporal rate of 24 or 30 Hz is too low and causes detrimental motion artifacts during camera panning, especially with large screen viewing. In addition, the Grand Alliance High-Definition Television (GA-HDTV) system developed for the United States can provide a video rate of 60 Hz or frames per second. Thus, a frame rate-up conversion from 24/30 Hz to 60 Hz is needed for utilizing information such as motion-picture films as a high quality source for HDTV.
Motion information is necessary for a high quality frame rate conversion. Known techniques such as frame repetition or linear interpolation do not suffice for conversion of data having high spatial resolution as in the case of HDTV. Although many motion compensated video coding techniques are presently known, the motion information required for frame rate conversion is not the same as that required for video coding. High quality frame rate conversion requires motion estimation that more closely reflects the true physical motion of a moving object in a video sequence.
Recently, many motion compensated frame/field conversion techniques have also become known. Techniques such as hierarchical block matching are used to generate displacement vectors that more closely reflect the true motion of the displaced objects in a video sequence, and a motion compensated filter (spatio-temporal filter) is utilized to reconstruct the continuous spatio-temporal signal. However, since some level of aliasing is always present in the sampled or scanned video signal, perfect reconstruction of the continuous spatio-temporal signal is not always possible through the use of such a motion-compensated reconstruction filter.
In addition, these known techniques for performing motion-compensated frame/field rate conversion have in general been proposed for implementation at the receiver. As the price of a receiver is sensitive to the chip count and memory, such receiver-enabled techniques are in practice limited to using only the neighboring frames relative to the subject frame for interpolation and, thus, do not have the benefit of using other frames which are prior and subsequent in time to the subject frame.
SUMMARY OF THE INVENTION
In accordance with the present invention, a process is disclosed for increasing the resolution of an existing video signal having a frame rate of 24 or 30 frames/second to a higher rate, such as 60 or 72 frames/second, for use with HDTV. The process interpolates, from the existing source sequence of frames, missing frames to generate the higher frequency signal by utilizing existing frames which are prior in time to a subject (interpolated) frame and subsequent in time to the subject frame.
The inventive process includes sampling the existing video frame sequence for a select number of times and increasing the sampling resolution by performing linear spatial interpolation of the sampled video signal. Shifts between picture elements (pels) in the frames of the video signal are then estimated from the increased resolution video signal and displacement vectors are constructed and plotted from the picture element shifts. Once the displacement vectors are plotted, a polynomial curve fit is performed and pel shifts are estimated therefrom. Utilizing the pel shifts and estimated pel shifts, the frames of the existing video signal are divided into groups based on the sharing of common sampling grids. In other words, frames having common sampling grids are grouped together. The required frames for the high resolution signal are then interpolated using frames within a corresponding group of frames.
For an existing video signal containing motion of objects, the process is performed by first dividing each frame into blocks of pixels and performing a hierarchical block matching technique on the blocks of each frame with respect to other frames in the sequence of existing frames to produce integer displacement vectors for each block in each frame. The integer displacement vectors are then calculated at ½ pel accuracy and a polynomial curve fit is performed for providing estimated displacements (vertical and horizontal) of each block at missing values of time, i.e. at time values for which a new frame is to be interpolated for use in the higher resolution HDTV signal. The blocks from each group are then divided into groups based on a common sampling grid criteria, and the missing or needed frames are interpolated piecemeal, one block at a time, using blocks within corresponding groups relative to the subject block for the interpolated frame.
Other objects and features of the present invention will become apparent from the following detailed description considered in conjunction with the accompanying drawings. It is to be understood, however, that the drawings are designed solely for purposes of illustration and not as a definition of the limits of the invention, for which reference should be made to the appended claims.
REFERENCES:
patent: 5473381 (1995-12-01), Lee
patent: 5508747 (1996-04-01), Lee
patent: 5600377 (1997-02-01), David et al.
patent: 5666164 (1997-09-01), Kondo et al.
patent: 5742348 (1998-04-01), Kuwahara et al.
patent: 5796437 (1998-08-01), Muraji et al.
patent: 5835150 (1998-11-01), Choi
patent: 5844616 (1998-12-01), Collet et al.
patent: 5886745 (1999-03-01), Muraji et al.
patent: 5940132 (1999-08-01), Kondo et al.
patent: 6069664 (2000-05-01), Zhu et al.
patent: 6128047 (2000-10-01), Chang et al.
Bugwadia Kobad
Petajan Eric D.
Puri Narindra N.
Desir Jean W.
Eisenzopf Reinhard
Lucent Technologies - Inc.
LandOfFree
Motion compensation image interpolation—frame rate... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Motion compensation image interpolation—frame rate..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Motion compensation image interpolation—frame rate... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2548079