Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1999-01-05
2003-01-28
Lee, Young (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
Reexamination Certificate
active
06512792
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a moving image encoding apparatus, and more particularly to a moving image encoding apparatus for use with a visual telephone system or a video conference system.
2. Description of the Related Art
Conventionally, a moving picture encoding system which employs motion compensation and orthogonal transform in combination is used frequently for compression coding of a moving image signal. Meanwhile, an input image signal inputted to an encoding apparatus includes high frequency components of noise, and in order to prevent deterioration of the coding efficiency by such high frequency components, a pre-processing filter is used to eliminate noise.
As such pre-processing filters, a recursive interframe filter for suppressing temporal noise components and an intraframe filter for suppressing spatial noise components are known. For a recursive interframe filter, a first-order recursive filter for pixels which includes a frame memory for one frame is widely used, and for an intraframe filter, a two-dimensional filter of 3×3 elements is widely used. Moving image encoding apparatus which employ such pre-processing filters as described above are disclosed in Japanese Patent Laid-Open No. 154588/1990, Japanese Patent Laid-Open No. 84024/1997, and so forth.
FIG. 9
shows an example of a moving image encoding apparatus which employs a pre-processing filter mentioned above. Conventional moving image compression coding is described below with reference to FIG.
9
. The moving image encoding apparatus shown processes an input image
101
in units of a pixel block (for example, 16×16 pixels or 8×8 pixels).
In particular, a pre-processing filter circuit
21
filters a pixel block of the input image
101
based on a result of discrimination of a moving/still discrimination circuit
22
and outputs a result of the filtering to a subtractor
1
. The subtractor
1
subtracts a predicted image
102
outputted from a motion compensation circuit
10
from the pixel block filtered by the pre-processing filter circuit
21
.
A result of the subtraction of the subtractor
1
is inputted through a DCT circuit
23
and a quantization circuit
24
to an encoding circuit
4
, by which it is encoded, and an output of the encoding circuit
4
is stored into a buffer
5
.
The output of the quantization circuit
24
is inputted not only to the encoding circuit
4
but also to another adder
8
through a dequantization circuit
6
and an inverse DCT circuit
7
. The adder
8
adds the predicted image
102
to the output of the quantization circuit
24
, that is, an output of the inverse DCT circuit
7
, and an output of the adder
8
is stored into a frame memory
9
.
A quantization step size
103
which is used by the quantization circuit
24
is determined by an encoding control circuit
25
. The quantization step size
103
used by the quantization circuit
24
is inputted also to the encoding circuit
4
and the dequantization circuit
6
simultaneously.
The encoding circuit
4
encodes the quantization step size
103
into coded data while the dequantization circuit
6
performs dequantization with the quantization step size
103
.
Referring to
FIG. 10
, the quantization circuit
24
includes a multiplier
41
, a rounding circuit
42
, and a multiplication coefficient table
43
. A quantization step size
103
inputted to the quantization circuit
24
is varied in units of a pixel block, but is constant during processing of the same pixel block.
The multiplication coefficient table
43
includes a correspondence between the quantization step size
103
and a multiplication coefficient
401
which is used by the multiplier
41
. For example, where the dequantization circuit
6
should output a result of multiplication of an input thereto by the quantization step size
103
, the multiplication coefficient table
43
in the quantization circuit
24
is set so that the input is multiplied by an inverse number to the quantization step size
103
.
At the top of processing of a pixel block, a multiplication coefficient
401
corresponding to the input of the quantization step size
103
is set to the multiplier
41
. After the multiplication coefficient
401
is set, an input to the quantization circuit
24
is multiplied by the multiplication coefficient
401
by the multiplier
41
. Then, an output of the multiplier
41
is rounded by the rounding circuit
42
, and an output of the rounding circuit
42
is outputted as an output of the quantization circuit
24
.
In the conventional moving image encoding apparatus described above, noise elimination from an input image signal in moving image encoding processing, particularly in encoding processing at a low bit rate, is an important factor to assure a high encoding efficiency. However, the conventional moving image encoding apparatus is disadvantageous in that a large amount of calculation is required for implementation of a pre-processing filter and this requires a large apparatus scale.
SUMMARY OF THE INVENTION
It is an object of the present invention to provide a moving image encoding apparatus which has a similar function to noise elimination by a pre-processing filter without requiring a large amount of calculation and a large apparatus scale.
In order to attain the object described above, according to the present invention, there is provided a moving image encoding apparatus, comprising prediction means for predicting a pixel block in a current frame of an input moving image signal from a preceding frame of the input moving image signal using motion compensation to produce a predictive error frame, orthogonal transform means for orthogonally transforming the predictive error frame, quantization means for quantizing the orthogonally transformed frame, encoding means for encoding the quantized frame, and a transmission buffer for storing the frame encoded by the encoding means, the quantization means quantizing the orthogonally transformed frame from the orthogonal transform means with a quantization step size having a different characteristic from that of a quantization step size which is used in dequantization included in the prediction processing of the prediction means and is used for the encoding processing by the encoding means.
Preferably, the moving image encoding apparatus further comprises moving/still discrimination means for detecting a moving area and a still area of the input image signal, and control means for controlling the quantization characteristic of the quantization means based on an output of the moving/still discrimination means and a remaining capacity of the transmission buffer.
In the moving image encoding apparatus, the quantization means quantizes the orthogonally transformed frame from the orthogonal transform means with a quantization step size having a different characteristic from that of a quantization step size which is used in dequantization included in the prediction processing, particularly, in motion compensation, of the prediction means and is used for the encoding processing by the encoding means. Consequently, the moving image encoding apparatus is advantageous in that it can be provided with a similar function to noise elimination by a pre-processing filter without requiring a large amount of calculation and a large apparatus scale.
Preferably, the quantization means quantizes the orthogonally transformed frame from the orthogonal transform means with a step size with which the orthogonally transformed frame is attenuated different from that which is used for the encoding operation by the encoding means. With the moving image encoding apparatus, a temporal variation of the input image becomes less likely to be encoded, and a temporal noise component elimination effect similar to that achieved where a recursive interframe filter is used can be achieved without using a recursive interframe filter.
Preferably, the quantization means quantizes the orthogonally transformed frame from the orthogonal transform means while att
Lee Young
Scully Scott Murphy & Presser
LandOfFree
Moving image encoding apparatus with a quantization step... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Moving image encoding apparatus with a quantization step..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Moving image encoding apparatus with a quantization step... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3054019