JP2844619B2 - Digital filter for image signal - Google Patents

Description

The present invention relates to a digital filter for an image signal, in particular,
The present invention relates to a digital filter applied to a loop filter or a post filter of an encoder of a hybrid code of DPCM and DCT.

[Conventional technology]

When image data is transmitted via a narrow-band transmission line in a video conference, a videophone, or the like, a hybrid code combining DPCM and DCT (cosine transform code) is used to compress the transmission data. FIG. 4 shows the configuration of the encoder of this hybrid code.

In FIG. 4, image data is supplied to an input terminal indicated by 1. The image data is supplied to the blocking circuit 2,
The data order is converted into an (8 × 8) block structure. The output signal of the blocking circuit 2 is supplied to the subtraction circuit 3. Data of the previous frame formed as described later is supplied to the subtraction circuit 3, and a frame difference is obtained from the subtraction circuit 3. This frame difference is supplied to the DCT transformer 4.

In the transformer 4, a two-dimensional cosine transform process is performed, and coefficient data is generated from the transformer 4. This coefficient data is supplied to the quantization circuit 5, and the coefficient data is quantized at a predetermined quantization step. The output signal of the quantization circuit 5 is supplied to a variable length coding circuit 6, where run-length coding and Huffman coding are performed. The output signal of the variable length coding circuit 6 is supplied to the multiplexer 7. The data of the quantization step is supplied from the quantization circuit 5 to the multiplexer 7, the motion vector is supplied from the motion compensation circuit 14, and the side information and the coefficient data are converted into transmission data in the multiplexer 7. The information of the quantization step is also supplied to the variable length coding circuit 6.

The output signal of the multiplexer 7 is supplied to the buffer memory 8. Transmission data is extracted from the buffer memory 8 to the output terminal 9. The buffer memory 8 is provided for controlling the data rate of the transmission data so as not to exceed the capacity of the transmission path. A control signal for controlling the quantization step is supplied from the buffer memory 8 to the quantization circuit 5. When the transmission data is too large, the quantization step is coarsened, and when the transmission data is small, the quantization step is performed. Control for controlling the density is performed.

The output signal of the quantization circuit 5 is supplied to the inverse transformer 10, and the output signal (frame difference) of the inverse transformer 10 is supplied to the addition circuit 11. The output signal of the adding circuit 11 is supplied to the frame memory 12. Frame memory
At 12, the restored image is reproduced, and the output signal of the frame memory 12 is supplied to the subtraction circuit 3 and the addition circuit 11 via the loop filter 13.

Further, a motion compensation circuit 14 is provided.
Then, the image data of the current frame (the output signal of the blocking circuit 2) and the image data of the previous frame (the output signal of the frame memory 12) are supplied. The motion compensation circuit 14 detects a motion vector indicating a motion between frames by block matching, and stores the motion vector in a frame memory.
12 and the multiplexer 7.

The above-described hybrid code performs a cosine transform on the frame difference obtained by the DPCM, so that a high compression rate can be realized.

FIG. 5 shows a configuration of a decoder used correspondingly to the encoder of FIG. The received data supplied from the input terminal 21 is stored in the buffer memory 22. The output signal of the buffer memory 22 is supplied to the demultiplexer 23, where the coefficient data and the side information are separated.

The coefficient data is decoded by the variable length decoding circuit 24 using the quantization step in the side information. Variable length decoding circuit 24
Are supplied to an inverse transformer 25 for inverse cosine transform. The output signal of the inverse transformer 25 is a frame difference, and the frame difference is supplied to the adding circuit 26. The addition circuit 26 includes a motion compensation circuit 27
Supplies the data of the previous frame, and restore data is obtained from the adder circuit 26. The received motion vector is supplied to the motion compensation circuit 27. The restored data from the adding circuit 26 is supplied to the block decomposing circuit 28,
The restored data returned to the original order by 28 is taken out to the output terminal 29.

The above-described loop filter 13 is provided to make random noise and block distortion generated by quantization of coefficient data in the quantization circuit 5 inconspicuous.

Conventionally, the filter coefficients shown in FIG. 6A are applied inside the edge of the block, and at the edge of the block,
The filter coefficients shown in FIG. 6B were applied, and the filter coefficients shown in FIG. 6C were applied at the corners of the block. In the case shown in FIG. 6A, the coefficient at the center corresponding to the pixel of interest is set to 4, and the coefficients for the eight surrounding pixels are set to 1 or 2.

[Problems to be solved by the invention]

Conventional loop filters, like preserving edges,
As shown in FIG. 6, the filter coefficients are switched.
However, in actuality, as a result of the filtering process, there is a problem that edges are blurred and image quality of a restored image is deteriorated.

SUMMARY OF THE INVENTION It is therefore an object of the present invention to provide a digital filter for an image signal which can suppress block distortion and noise and has a good edge preserving effect.

[Means for solving the problem]

The present invention provides an image signal digital filter in which data of a target pixel and data of a plurality of pixels around the target pixel are each multiplied by a filter coefficient, and a multiplied output is added. Means for detecting a maximum value and a minimum value with respect to data of a plurality of pixels around the pixel of interest, and detecting a dynamic range that is a difference between the maximum value and the minimum value; Means for generating a first filter coefficient that gradually increases so as to become larger as the range becomes larger and becomes a predetermined value when the threshold is exceeded, and a filter coefficient for peripheral pixels, where the detected dynamic range is larger. Means for generating a gradually decreasing second filter coefficient so that the second filter coefficient becomes smaller when the threshold value is exceeded and becomes 0 when the threshold value is exceeded. Means for multiplying the filter coefficient by the data of the pixel of interest and the second filter coefficient by the data of the peripheral pixels, and adding the respective multiplied outputs. .

[Action]

In the region including the edge of the image, there is a sharp change in the level, and the dynamic range DR is large. On the other hand, in a flat region of the image, the dynamic range DR is small. When the filtering process is performed, there is a problem that an edge is blurred. However, according to the present invention, in a region where the dynamic range DR is large, that is, when the threshold value TH is exceeded, no filtering is applied, so that the edges can be well preserved. Also,
In a flat area of an image, noise and block distortion can be suppressed.

〔Example〕

Hereinafter, an embodiment of the present invention will be described with reference to the drawings. The present invention provides a loop filter 13 in the encoder of the hybrid code shown in FIG. 4 (the loop filter may be provided on the output side of the adder circuit 11).
Alternatively, it is applied to a post filter connected to the output terminal 29 of the decoder shown in FIG. In this embodiment, the present invention is applied to a post filter.

In FIG. 1, an input terminal indicated by 31 is supplied with restored data restored on the receiving side and converted into the original order. Line delay circuits 32 and 33 indicated by LD are cascade-connected to the input terminal 31, and a sample delay circuit 34 indicated by SD is provided at each of the input terminals 31 and the output terminals of the line delay circuits 32 and 33.
35, 36, 37, 38 and 39 are respectively connected. The data generated on the output side of the sample delay circuit 36 is the data of the target pixel. At the output side of the line delay circuit 32 and the output side of the sample delay circuit 37, data of pixels before and after the delay pixel are obtained. In addition, from the input terminal 31 and the output sides of the sample delay circuits 34 and 35, data of three pixels on the line below the line of the pixel of interest is obtained. Further, from the output side of the line delay circuit 33 and the output sides of the sample delay circuits 38 and 39, data of three pixels on the line above the line of the pixel of interest is obtained.

Therefore, the target pixel and the eight surrounding pixels are obtained at the same time. A memory may be used instead of the line delay circuit and the sample delay circuit. These attention pixels and 8
The peripheral pixels are supplied to the multiplication circuits 40 to 48, respectively. The output signals of the multiplication circuits 40 to 48 are added by the addition circuit 49, and the added output is supplied to the division circuit 50, where it is divided by (1/16). The division circuit 50 is configured by a shift register.

The data of the target pixel and the eight pixels surrounding it are supplied to the maximum and minimum value detection circuit 52, and the maximum value MAX and the minimum value MIN are detected. The maximum value MAX and the minimum value MIN are supplied to the subtraction circuit 53, and the dynamic range DR represented by (MAX−MIN) is obtained from the subtraction circuit 53. The filter coefficients k and 2k corresponding to the dynamic range DR are read from the ROM 54. The threshold value TH from the threshold value generation circuit 55 is also supplied to the ROM 54. The filter coefficient k from the ROM 54 is supplied to a 1 generation circuit 56 for generating a filter coefficient 1 and is also supplied to multiplication circuits 40, 42, 46 and 48. The filter coefficient 2k is supplied to the multiplication circuits 41, 43, 45, and 47. The filter coefficient 1 is supplied to the multiplication circuit 44.

FIG. 2 shows an example of the filter coefficient. The filter coefficient for the target pixel is set to 1, the filter coefficient for the upper and lower and left and right pixels is set to 2k, and the filter coefficient for the pixel in the oblique position of the target pixel is set to k. The filter coefficients k and l are as follows.

As shown in FIG. 3, the filter coefficient k gradually decreases according to the dynamic range DR of the target pixel and the peripheral pixels, and becomes 0 when the threshold value TH is exceeded. When the dynamic range DR is 1/2 TH, (k = 0.5)
When the dynamic range DR is TH, (k = 0).

The above-described filter coefficient is such that the coefficient for the pixel of interest is the maximum, the coefficient for the upper and lower pixels and the left and right pixels is 2k, and the coefficient for the obliquely slightly longer pixel is k. Also, the reason that the filter coefficient changes in accordance with the dynamic range DR is that when an edge is included, the dynamic range DR is increased, so that the degree of filtering is reduced, and the edge is not included.
This is because when the DR is small, the degree of filtering is increased. Therefore, the edge can be preserved, and noise or block distortion is effectively removed in a relatively flat region other than the edge.

Further, the threshold value TH generated from the threshold value generation circuit 55 may be changed according to the capacity of the transmission path. The capacity of the transmission line
In the case of 64 kbps and the case of 2 Mbps, the latter can reduce the quantization step and reduce noise. Therefore, the latter can make the threshold value TH smaller than the former.

Although the above-described embodiment is an example of a two-dimensional filter, a one-dimensional filter may be used, and the change of the filter coefficient with respect to the dynamic range DR is not limited to a cosine change, but monotonically decreases. But it is good.

As described above, the present invention can be applied not only to a post filter but also to a loop filter. Even in the case of a loop filter, the present invention is applied not to a block structure but to data in a state similar to that obtained by scan conversion. In the case of a loop filter, data with a motion vector of 0 and a frame difference of 0 is not filtered in order to prevent the data from being deteriorated by the filtering process. In this respect, it differs from a post-filter which is always filtered. Further, it goes without saying that the present invention can be applied to processing of image data other than the image transmission devices shown in FIGS. 4 and 5.

〔The invention's effect〕

According to the present invention, the filter coefficient is reduced according to the dynamic range DR of the target pixel and its surrounding pixels, and when the dynamic range DR exceeds the threshold, the filter coefficient is set to 0.
Therefore, edges can be well preserved and noise can be effectively suppressed.

Further, according to the present invention, the filter coefficient for the pixel of interest is gradually increased according to the dynamic range DR, and the filter coefficient for the peripheral pixels is gradually decreased according to the dynamic range DR. As a result, the function of the filter can be changed, and the effects of edge preservation and noise removal can be exhibited more favorably.

[Brief description of the drawings]

FIG. 1 is a block diagram of one embodiment of the present invention, FIG. 2 is a schematic diagram showing an example of a filter coefficient in this embodiment,
FIG. 3 is a schematic diagram showing changes in filter coefficients, FIG. 4 is a block diagram of an encoder of an image transmission apparatus to which the present invention can be applied, and FIG. 5 is a block of a decoder of an image transmission apparatus to which the present invention can be applied. FIG. 6 is a schematic diagram showing filter coefficients of a conventional loop filter. Explanation of main reference numerals in the drawings 31: input terminal, 40 to 48: multiplication circuit, 49: addition circuit, 51: output terminal, 52: maximum value and minimum value detection circuit, 53: subtraction circuit, 54: filter coefficient generation ROM.

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁶ , DB name) H04N 7/24-7/68 H04N 1/41-1/419

Claims (1)

(57) [Claims] 1. A digital filter for an image signal in which data of a pixel of interest and data of a plurality of pixels around the pixel of interest are respectively multiplied by filter coefficients and a multiplied output is added. Means for detecting a maximum value and a minimum value between data and data of a plurality of pixels around the pixel of interest, and detecting a dynamic range that is a difference between the maximum value and the minimum value; and a filter coefficient for the pixel of interest. The first dynamic range, which increases as the detected dynamic range increases, gradually increases to a predetermined value when the detected dynamic range exceeds a threshold.
Means for generating a filter coefficient of: a second filter, which is a filter coefficient for the peripheral pixels, wherein the filter coefficient decreases as the detected dynamic range increases, and gradually decreases to 0 when the detected dynamic range exceeds a threshold value. Means for generating a coefficient; means for multiplying the first filter coefficient by the data of the pixel of interest, multiplying the second filter coefficient by the data of the peripheral pixel, and adding respective multiplied outputs. A digital filter for an image signal.