Motion video signal processing for recording or reproducing – Local trick play processing – With randomly accessible medium
Reexamination Certificate
1999-02-02
2004-09-28
Tran, Thai (Department: 2615)
Motion video signal processing for recording or reproducing
Local trick play processing
With randomly accessible medium
C386S349000, C375S240100
Reexamination Certificate
active
06798977
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to an image processing apparatus and method of efficiently encoding image data and decoding the encoded data.
2. Related Background Art
H.261, MPEG-1, and MPEG-2 are conventionally known as moving image coding schemes and internationally standardized by ITU (International Telecommunication Union) or ISO (International Organization for Standardization). These H.261, MPEG-1, and MPEG-2 are put in writing as H.264 recommendations, ISO11172, and ISO13818, respectively. Motion JPEG (Joint Photographic Coding Experts Group) coding is also known which encodes each frame by adapting still image coding (e.g., JPEG coding) to the frame.
An encoding system which encodes a video signal by MPEG-1 will be described below with reference to FIG.
1
.
Referring to
FIG. 1
, a video signal supplied from a TV camera
1001
is input from an input terminal
1003
to a moving image encoding apparatus
1002
.
An A/D converter
1004
converts the input video signal from the input terminal
1003
into a digital signal and inputs the signal to a block forming unit
1005
.
The block forming unit
1005
forms a macro block composed of 16×16 pixels sequentially from the upper left pixel to the lower right pixel.
MPEG-1 can encode image data by three encoding modes: an I-frame mode (to be referred to as an I-frame hereinafter) for performing intra-frame encoding, a P-frame mode (to be referred to as a P-frame hereinafter) for performing inter-frame encoding from past frames, and a B-frame mode (to be referred to as a B-frame hereinafter) for performing inter-frame encoding from past and future frames.
A frame mode unit
1017
selects one of these three frame modes. A frame mode is determined by taking account of the bit rate of encoding, prevention of image quality deterioration caused by accumulation of operation errors in DCT (Discrete Cosine Transform), image editing, and scene changes.
A process of encoding an I-frame will be described first.
For an I-frame, a motion compensator
1006
does not operate and outputs “0”. A subtracter
1007
subtracts the output from the motion compensator
1006
from the output from the block forming unit
1005
and supplies the difference to a DCT unit
1008
.
The DCT unit
1008
performs DCT for the difference data supplied from the subtracter
1007
in units of blocks of 8×8 pixels and supplies the transformed data to a quantizer
1009
.
The quantizer
1009
quantizes the transformed data from the DCT unit
1008
and supplies the quantized data to an encoder
1010
.
The encoder
1010
one-dimensionally rearranges the quantized data from the quantizer
1009
, determines codes by the 0-run length and value, and supplies the encoded data to an output terminal
1011
.
The quantized data from the quantizer
1009
is also supplied to an inverse quantizer
1012
. The inverse quantizer
1012
inversely quantizes the supplied quantized data and supplies the inversely quantized data to an inverse DCT unit
1013
. The inverse DCT unit
1013
performs inverse DCT for the inversely quantized data and supplies the inversely transformed data to an adder
1014
. The adder
1014
adds the output “0” from the motion compensator
1006
and the output from the inverse DCT unit
1013
and stores the sum in a frame memory
1015
or
1016
.
A process of encoding a P-frame will be described next.
For a P-frame, the motion compensator
1006
operates, and an output from the block forming unit
1005
is input to the motion compensator
1006
. An image of a temporally immediately preceding frame is also input to the motion compensator
1006
from the frame memory
1015
or
1016
. The motion compensator
1006
performs motion compensation by using the input image data and outputs a motion vector and a predictive macro block.
The subtracter
1007
calculates the difference between the output from the block forming unit
1005
and the predictive macro block. This difference is subjected to DCT and quantization. The encoder
1010
determines codes on the basis of the quantized data and the motion vector and outputs the codes from the terminal
1011
.
The quantized data from the quantizer
1009
is also supplied to the inverse quantizer
1012
. The inverse quantizer
1012
inversely quantizes the supplied quantized data and supplies the inversely quantized data to the inverse DCT unit
1013
. The inverse DCT unit
1013
performs inverse DCT for the inversely quantized data and supplies the inversely transformed data to the adder
1014
. The adder
1014
adds the output from the inverse DCT unit
1013
and the output predictive macro block data from the motion compensator and stores the sum in a frame memory
1015
or
1016
.
A process of encoding a B-frame is as follows.
Although motion compensation is performed for this B-frame as for a P-frame, the motion compensator
1006
performs this motion compensation by using data from both the frame memories
1015
and
1016
, and forms and encodes a predictive macro block.
In the methods by which an entire image is encoded as described above, however, a background image with no motion must be repeatedly transmitted, and this wastes the code length. For example, in images in a video telephone system or video conference, only objects actually moving are persons, and the background remains stationary. In an I-frame which is transmitted for each fixed time, a background image with no motion is also transmitted to produce useless codes (code data of the background image).
FIG. 2
shows an image in a video conference or the like.
Referring to
FIG. 2
, a person
1050
faces a television camera in a video conference room. This person
1050
and a background
1051
are encoded in the same frame by the same encoding method.
Since the background
1051
remains still, almost no codes are produced if motion compensation is performed, but a large number of codes are produced in an I-frame.
Consequently, even for a portion with no motion, large encoded data is repeatedly and uselessly transmitted. Also, if the motion of the person
1050
is large and a large number of codes are generated by encoding, no enough code amount can be obtained by an I-frame encoding process performed after that. If this is the case, a quantization coefficient must be set for coarse quantization, and this undesirably deteriorates even the image quality of the background with no motion. Note that a moving object like the person
1050
described above will be called a subject hereinafter.
SUMMARY OF THE INVENTION
The present invention has been made in consideration of the above situation, and has as its object to provide an image processing apparatus and method of efficiently encoding input image data and decoding the encoded data.
To achieve the above object, according to one preferred aspect of the present invention, in an image processing apparatus and method, a plurality of objects are separated from input moving image data, a separated first object is encoded by a first encoding method, a separated second object is encoded by a second encoding method, and the encoding process for the second object is controlled in accordance with encoded data of the first object.
According to another preferred aspect of the present invention, in an image processing apparatus and method, a plurality of objects are separated from input moving image data, a separated first object is encoded by a first encoding method, a separated second object is encoded by a second encoding method, and the encoding process for the second object is controlled in accordance with a recording capacity of a recording medium for recording encoded data of the first object and encoded data of the second object.
According to still another preferred aspect of the present invention, in an image processing apparatus and method, a plurality of objects are separated from input moving image data, a separated first object is encoded by a first encoding method, a separated second object is encoded by a second encoding method, and the encoding pr
Canon Kabushiki Kaisha
Chieu Po-lin
Fitzpatrick ,Cella, Harper & Scinto
Tran Thai
LandOfFree
Image data encoding and decoding using plural different... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Image data encoding and decoding using plural different..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image data encoding and decoding using plural different... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3269388