Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1999-09-15
2002-11-12
Diep, Nhon (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
Reexamination Certificate
active
06480540
ABSTRACT:
FIELD OF THE INVENTION
The invention relates to a method of image compression in which the images are coded according to groups of variable lengths.
It relates more particularly to a method of the MPEG type, particularly of MPEG2 type. Although the invention is not limited to this standard, it will be referred to primarily in the remainder of the description.
BACKGROUND OF THE INVENTION
The principle of such compression is reiterated below.
In the video MPEG2 standard, compression of the digital video signals is obtained by exploiting the spatial redundancy and the temporal redundancy of the coded images.
The spatial redundancy is evaluated principally by virtue of a succession of three operations: an operation commonly called discrete cosine transform and denoted DCT (“Discrete Cosine Transform”), an operation of quantisation of the coefficients arising from the DCT and an operation of variable-length coding to describe the quantified coefficients arising from the DCT.
The temporal redundancy is analysed by a movement-compensation operation which consists, by translation of each block of the current image, in searching for the most similar block situated in the reference image. The analysis of the temporal redundancy leads to a field of translation vectors being determined, commonly called movement vectors, as well as a prediction error which is the difference between the signal of the current image and the signal of the image predicted by movement compensation. The prediction error is then analysed according to the principle of spatial redundancy.
MPEG coding is of predictive type. It follows that the decoding which is associated with it should be regularly reinitialised so as to protect the signal against any transmission error or any break in signal due to the decoder being switched over from one programme to another.
To this end, the MPEG2 standard provides that, periodically, the images should be coded in spatial mode, that is to say according to a mode exploiting only spatial redundancy. The images coded in spatial mode are called INTRA images or I images.
The images coded by exploiting temporal redundancy are of two types: on the one hand, the images constructed by reference to a temporally previous image on the basis of a front prediction and, on the other hand, the images constructed by reference to two temporally previous and subsequent images on the basis of a front prediction and of a back prediction.
The coded images constructed on the basis of a front prediction are called predicted images or P images and the coded images constructed on the basis of a front and of a back prediction are called bidirectional images or B images.
An I image is decoded without reference being made to images other than itself. A P image is decoded by referring to the P or I image which precedes it. A B image is decoded by relying on the I or B image which precedes it and on the I or P image which follows it.
The periodicity of the I images defines a group of images widely denoted GOP (“Group Of Pictures”).
Within a single GOP, the quantity of data contained in an I image is generally greater than the quantity of data contained in a P image and the quantity of data contained in the P image is generally greater than the quantity of data contained in a P image.
At 50 Hertz, the GOP is presented as an I image followed by a sequence of B and P images which, most of the time, exhibits the following sequence
I, B, B, P, B, B, P, B, B, P, B, B.
However, the standard does not demand N=12 images being provided in a GOP, as is the general case, nor that the distances M between two P images should always be equal to 3. More precisely, the distance M is the number n of B images preceding or following a P image, increased by one unit, i.e. M=N+1.
The number N represents the size or length of the GOP, while the number M represents its structure.
SUMMARY OF THE INVENTION
The invention results from the observation that it is possible to act on the M and N parameters to enhance the level of compression and/or enhance the quality of the coding.
The method of coding according to the invention is characterised in that at least one parameter is determined characterising the source images which are to be coded according to a group and in that the length and the structure of the group is made to depend on this parameter or these parameters.
In one embodiment, the parameter(s) characterising the source images is or are determined with the aid of a test coding in the course of which defined values are allocated to N, M and to the quantisation interval Q.
The test coding is carried out, for example, in open loop.
In one particularly simple embodiment, a parameter (Pcost) characterising the P images obtained during the test coding and a parameter (Bcost) characterising the B images obtained during the test coding are determined separately, these parameters characterising the P and B images being, preferably, the average costs of coding of the P and B images. The cost of coding an image is the number of bits (headers included) which is necessary for the coding.
In this case, the number N can be made to depend on the parameter characterising the P images and the number M on the parameter characterising the B images.
During trials carried out in the context of the invention, on sequences of images of various types, it was noted that, for each type of sequence, an optimal number N existed providing a minimum coding cost (or throughput) for the P images and an optimal number M providing a minimum coding cost (or throughput) for the B images, these costs being obtained during the test coding. These sequences are distinguished from another by movement of variable amplitudes, different objects, different spatial definitions and different contents.
It was noted, moreover, that a practically linear relationship exists between the optimal number N and the throughput of the P images. Likewise, a practically linear relationship exists between the number M and the throughput of the B images. Hence, knowing the throughputs of the P and B images, it is easy to calculate the numbers N and M providing the best results.
In an example corresponding to the MPEG2 standard, 50 Hz, the test coding is carried out with N=12, M=3 and Q=15, the relationship between N and the throughput of the P images is approximately as follows:
N
=
INT
[
389000
-
P
cos
t
10000
]
+
1
,
with
12
≤
N
≤
30
(
1
)
and the relationship between M and the throughput, or cost, Bcost of the B images is as follows:
N
=
INT
[
179000
-
B
cos
t
20000
]
+
1
,
with
1
≤
M
≤
7.
(
2
)
It is also possible to limit M to 5.
In these formulae, INT signifies the integer part.
The limitation on N between 12 and 30 and the limitation on M to a maximum value of 7 makes it possible to have a simple embodiment of the coders and to limit the programme-changing time. With the same aim, it is also possible to impose other limitations or constraints, particularly that M be constant in the GOP and/or that it be a sub-multiple of N.
In one embodiment, if the values of N and of M taken individually and together are not compatible with the constraints, the values of M and of N closest to the calculated values and which satisfy the stipulated compatibility will be chosen. In this case, the value of M will be favoured, that is to say that if a choice has to be made between several M, N pairs, the pair will be chosen for which the M value is closest to that which results from the calculation.
The formula (2) above applies only if Bcost does not exceed 179800. In the opposite case, that is to say if Bcost>179000, experiment has shown that it was necessary, in this example, for M to be chosen in the following way:
M
=
5.
INT
[
P
cos
t
B
cos
t
-
1
]
,
with
1
≤
M
≤
7.
(
3
)
If the cost of a B image is higher than the cost of a
Diep Nhon
Duffy Vincent E.
Kurdyla Ronald H.
Tripoli Joseph S.
LandOfFree
Method of image compression and device for implementing this... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method of image compression and device for implementing this..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of image compression and device for implementing this... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2924855