Image analysis – Image compression or coding – Interframe coding
Reexamination Certificate
2000-05-17
2003-10-14
Couso, Jose L. (Department: 2621)
Image analysis
Image compression or coding
Interframe coding
Reexamination Certificate
active
06633676
ABSTRACT:
The invention relates to a method of encoding a video signal comprising a series of pictures in a motion-compensated encoder, said method comprising the steps of receiving the series of pictures and estimating motion vectors in the series of pictures with sub-resolution accuracy.
The invention also relates to a motion-compensated encoder for encoding a video signal comprising a series of pictures, said encoder comprising receiving means for receiving the series of pictures, estimation means for estimating motion vectors in the series of pictures with sub-resolution accuracy, a picture memory for storing the pictures, and means for forming and supplying a motion-compensated encoded video signal by means of the picture memory.
The invention further relates to a camera system comprising an image sensor for picking up a video signal which comprises a series of pictures, and a motion-compensated encoder.
The article by Regis Saint Girons “The digital simulcast AD-HDTV coding system”, IEEE Trans. on Consumer Electr. Vol. 38, No. 4, November 1992, pp. 778-782 describes an Advanced Digital High Definition Television System which comprises an MPEG video encoder. In the MPEG encoder, a video signal is received after it has been processed in a preprocessor. The encoder comprises a device for estimating motion. The article describes a technique for estimating motion in macroblocks of 16×16 pixels with half-pixel accuracy. Motion vectors are generated which indicate the motion of a macroblock in a picture with respect to a previous picture. When a motion vector is known, pixel values can be predicted for a macroblock from a previous, reconstructed picture. For the purpose of compression, the predicted pixel values are subtracted from the actual pixel values of the relevant macroblock. The difference is a residual macroblock. The residual macroblock is further encoded and applied, together with the motion vectors, to an output of the encoder. In this way, use is made of time redundance in the video signal.
It is an object of the invention to enhance the resolution in a video signal.
To this end, the method according to the invention is characterized in that it comprises the step of making a choice between generating and supplying a motion-compensated encoded video signal and generating and supplying a higher-resolution picture having a higher resolution than the pictures of the series, if the motion-compensated encoded video signal is generated, forming the motion-compensated encoded video signal by means of the picture memory, and if the higher-resolution picture is generated, forming the higher-resolution picture from the series of pictures and the motion vectors by means of a picture memory. Forming a higher-resolution picture is based on the availability of sub-resolution motion information and on the presence of aliasing in the video signal. The information present in various pictures is thus used for forming a new picture with a higher resolution. The resolution is the extent of fineness or focus of the picture. In many cases, the resolution will correspond to the number of pixels from which the picture is built up. If a number of pixels is jointly used for displaying the picture information of one pixel, the number of pixels may be larger than the resolution. This occurs, for example, when using a color filter grating which samples color channels through a limited set of pixels, or when projecting a picture on a larger number of pixels without increasing the picture contents. Sub-resolution accuracy involves a greater accuracy than the fineness or focus of the picture and corresponds in many cases to sub-pixel accuracy.
It is to be noted that enhancement of resolution is known per se from European patent application EP 0 731 600. Another method of obtaining a higher-resolution picture is known from the article: Debin Chen et al “Extraction of High-Resolution Video Stills from MPEG Picture Sequences”, ICIP '98, October 1998, Chicago. Furthermore, such a method is described in the non-prepublished, international patent application PCT/IB 98/01966, our reference PHN 16.674. In the method described in this document, motion vectors present in an MPEG signal are used for forming the higher-resolution picture.
The known methods are used for displaying the video signal. The enhancement of resolution in accordance with the invention is, however, performed when encoding the video signal in a motion-compensated encoder. Although European patent application EP 0 731 600 describes that the video signal can be remote-displayed via a communication network, this patent application does not give any indication about encoding the video signal in a motion-compensated encoder. An advantage of enhancing the resolution in a motion-compensated encoder is that means for estimating motion are already present in such an encoder. Moreover, there is already a picture memory in which various pictures from the series can be stored. According to the invention, the motion estimator present in the motion-compensated encoder and the picture memory are used to advantage, both for generating the motion-compensated encoded video signal and for generating the higher-resolution picture.
By repeating the method according to the invention, a series of pictures with a higher resolution is supplied from an output of the encoder. In a preferred embodiment of the invention, the series of pictures is adapted to a predetermined standard output signal. This may be, for example, a standard number of pictures per second. Another, generally larger number of lower-resolution pictures per second can be applied to the input, which lower-resolution pictures are used for forming the standard number of pictures at the output with a higher resolution, dependent on the video signal. Two higher-resolution pictures can be combined to one interlaced picture.
Generally, the invention requires storage space for storing the various pictures from the series. This means that there should be more storage space in an encoder in which the resolution is enhanced than in an encoder in which the resolution is not enhanced, because more pictures are generally stored in the case of resolution enhancement. An embodiment complying with the quantity of required storage space is characterized in that the method comprises the steps of selecting a predetermined region of interest from the video signal and storing the series of pictures in the picture memory, the pictures relating to the region of interest. The region of interest forms a sub-signal of the video signal. The region of interest is a preselected sub-region of the display supplied by the video signal. The choice of the region of interest is passed on to the encoder, whereafter the region of interest is selected from the video signal. The sub-pictures relating to the region of interest occupy less space than the full pictures. When the region of interest is sufficiently small, several sub-pictures can be stored in a conventional picture memory without an extension of this memory being required. The conventional picture memory is the picture memory which is present in a comparable encoder in which no higher-resolution picture is formed. In this way, an inexpensive and practical method is possible for enhancing the resolution. A series of sub-pictures of a region of interest is combined to one picture having a higher resolution than the separate sub-pictures, which higher-resolution picture preferably has the same number of pixels as a normal picture.
The motion is preferably estimated in a hybrid encoder in which both motion compensation and other encoding operations are performed. An example is MPEG. An MPEG encoder already comprises standard means for estimating motion and generating motion vectors, and a picture memory for storing the various pictures. A higher-resolution picture may be supplied as an intracoded frame. An MPEG decoder can decode it without additional information. Another example of a compression technique is H.263.
The motion-compensated encoder according to the invention
Kleihorst Richard Petrus
Op De Beeck Marc Joseph Rita
Van Der Avoird André
Van Der Werf Albert
Couso Jose L.
Koninklijke Philips Electronics , N.V.
LandOfFree
Encoding a video signal does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Encoding a video signal, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Encoding a video signal will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3152642