JP2915922B2 - Image signal compression coding device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to an image signal compression encoding apparatus, and more particularly, to an image signal compression encoding apparatus that keeps a data amount of a compression-encoded image constant.

BACKGROUND ART When digital image data such as image data captured by an electronic still camera is stored in a memory,
In order to reduce the amount of data and the storage capacity of the memory, various types of compression encoding are performed. In particular, two-dimensional orthogonal transform coding is widely used because it can perform coding at a large compression rate and can suppress image distortion due to coding.

In such two-dimensional orthogonal transform coding, image data is divided into a predetermined number of blocks, and image data in each block is subjected to two-dimensional orthogonal transform. The orthogonally transformed image data, that is, the transform coefficient, is compared with a predetermined threshold, and a portion below the threshold is truncated (coefficient truncation).
Is performed. As a result, the conversion coefficient equal to or smaller than the threshold value is thereafter processed as 0 data. Next, the transform coefficient subjected to coefficient truncation is divided by a predetermined quantization step value, that is, a normalization coefficient, and quantization by a step width, that is, normalization is performed. Thereby, the value of the conversion coefficient, that is, the dynamic range of the amplitude can be suppressed.

The comparison with the threshold value and the normalization process are often performed simultaneously. That is, when the conversion coefficient is normalized by a predetermined normalization coefficient and the result is converted into an integer, the conversion coefficient having a value lower than the normalization coefficient becomes zero.

The normalized transform coefficients are then Huffman coded and stored in memory.

In such two-dimensional orthogonal transform encoding, when the above-described normalization coefficient is normalized as a constant value and encoded, the amount of encoded data differs depending on the image data, which is inconvenient for recording in a memory. Met.

Therefore, it is conceivable to calculate a degree to which each image contains a high frequency component, that is, a so-called activity, set a normalization coefficient for each image based on the activity, and thereby perform normalization. By setting the normalization coefficient in this manner, the data amount of each image can be made constant.

By the way, when the number of bits of the coded data to be allocated to one entire image is constant, since the activity of each block constituting the image is different, the number of bits to be allocated varies according to the activity for each block. It is desirable to make it.

When the number of bits allocated to each block is changed in this way, the output buffer that stores the data encoded for each block has a capacity of the maximum number of bits that can be allocated to each block, for example, about 5 Kbit. Requires the capacity of Therefore, an output buffer having a large capacity is required, and there is a drawback that the apparatus becomes large.

An object of the present invention is to solve the problems of the prior art and to provide an image signal compression encoding apparatus capable of reducing the capacity of an output buffer for accumulating data for each encoded block. I do.

DISCLOSURE OF THE INVENTION According to the present invention, an image signal compression encoding apparatus that divides digital image data forming one screen into a plurality of blocks and performs two-dimensional orthogonal transform encoding on the image data of each block includes a plurality of blocks. Orthogonal transform means for two-dimensional orthogonally transforming the digital image data divided into blocks, encoding means for encoding data orthogonally transformed by the orthogonal transform means, and activity of image data for each divided block is calculated Block activity calculating means, coded data amount allocating means for calculating the amount of coded data allocated to each block based on the activity for each block calculated by the block activity calculating means, and output from the coded data amount allocating means Limits the amount of encoded data for each block output from the encoding means according to Encoding data control means, and the encoded data amount distribution means restricts the upper limit value of the encoded data amount distributed based on the activity of each block to a predetermined value and outputs the upper limit to the encoded output control means. The encoded output control means limits the encoded data amount for each block output from the encoding means in accordance with the encoded data amount restricted by the upper limit value.

DESCRIPTION OF THE EMBODIMENTS Next, with reference to the accompanying drawings, embodiments of the image signal compression encoding apparatus according to the present invention will be described in detail.

FIG. 1 shows an embodiment of an image signal compression encoding apparatus according to the present invention.

This device has a memory controller 10. The memory controller 10 is connected to a memory 60, and reads out still image data for one frame captured by, for example, an electronic still camera from the memory 60. The output from the memory controller 10 is input to the blocking unit 12. Blocking unit 12
Is constituted by a frame buffer, image data input from the memory controller 10 and stored in the blocking unit 12 is divided into a plurality of blocks, read out for each block, and sent to the two-dimensional orthogonal transform unit 14. The two-dimensional orthogonal transform unit 14 performs two-dimensional orthogonal transform on the image data of each block. As the two-dimensional orthogonal transform, a known orthogonal transform such as a discrete cosine transform or a Hadamard transform is used.

The image data for each block subjected to the two-dimensional orthogonal transformation in the two-dimensional orthogonal transformation unit 14 is arranged vertically and horizontally, and low-order data of the AC component is arranged in order from the upper left part, and the higher-order data is arranged in the lower right direction. Data. The DC component data is arranged at the upper left corner. 2D orthogonal transform unit
The output of 14 is sent to the normalization unit 16.

The normalization unit 16 performs normalization after performing coefficient truncation on the image data that has been subjected to the two-dimensional orthogonal transformation in the two-dimensional orthogonal transformation unit 14, that is, the transform coefficients. The coefficient truncation is to compare the orthogonally transformed transform coefficient with a predetermined threshold value and to cut off a portion below the threshold value. Normalization is performed by dividing the truncated transform coefficient by a predetermined quantization step value, that is, a normalization coefficient α, and a normalization coefficient α
Quantization is performed. The normalization coefficient α is obtained from a look-up table based on a value obtained by summing the activities for each block, as described later.

The image data for each block output from the blocking unit 12 is also sent to the block activity calculation unit 20. The block activity calculation unit 20 calculates the activity of each block, that is, the degree to which the block contains image data of high frequency components. The activity of a block is obtained by adding the absolute value of the difference between each pixel data constituting the block and the average value of these pixel data.

The activity for each block output from the block activity calculation unit 20 is the total activity calculation unit.
26 and to the bit allocation calculation unit 40 of the encoding unit 28.

The total activity calculation unit 26 adds the activity for each block sent from the block activity calculation unit 20, and calculates the total value of the activities. The total activity calculating unit 26 outputs the total value of the activities to the normalization coefficient setting unit 22 and the bit allocation calculating unit 40.

The normalization coefficient setting unit 22 sets a normalization coefficient according to the total value of the activities input from the total activity calculation unit 26. The setting of the normalization coefficient is performed, for example, by a conversion as shown in FIGS. 3A and 3B using a look-up table stored in a storage unit (not shown).
According to the conversion shown in FIG. 3A, the normalization coefficient changes in proportion to the total value of the activities. The transformation shown in FIG. 3B is
Since the increase in the normalization coefficient is small with respect to the increase in the total value of the activities, high-precision encoding can be performed. The normalization coefficient setting unit 22 outputs the normalization coefficient thus set to the normalization unit 16.

The normalization unit 16 performs normalization using the normalization coefficient sent from the normalization coefficient setting unit 22. That is, the image data for each block is divided by the normalization coefficient. Since the normalization coefficient used for normalization is set based on the value obtained by summing the activities for each block as described above,
It is common for the entire image, that is, for all blocks.

The normalization is performed by dividing the transform coefficient after coefficient truncation by the value α of one selected normalization coefficient, and replacing the data stored in the weight table T as shown in FIG. And the normalization coefficient α may be used together. As for the conversion coefficient, since the low-frequency component is important as data and the high-frequency component is low in importance, the weight table T as shown in FIG. Is normalized by dividing the transform coefficient subjected to the coefficient truncation by a value α · T obtained by multiplying the data of the table T by the normalization coefficient α. It may be performed.

The normalized transform coefficients are arranged in blocks in the same manner as the pixel data shown in FIG. 2, and their AC components are scanned in a zigzag manner in order from the low-frequency components as shown in FIG. It is input to the 28 AC encoding unit 32.

The AC encoding unit 32 encodes the AC component of the normalized transform coefficient that is scanned and input in a zigzag manner as described above. Since the AC component of the conversion coefficient is often continuous with zeros, a continuous amount of data of zero value, that is, a zero run length is detected, and a zero run length and a non-zero amplitude are obtained. Encode. The output from the AC encoding unit 32 is sent to the fixed-length unit 36.

On the other hand, the DC component of the image data for each block output from the blocking unit 12 is sent to the DC coding unit 30 of the coding unit 28, and is subjected to Huffman coding in the DC coding unit 30. The DC encoded data output from the DC encoding unit 30 is sent to the output buffer 38.

The DC coded data is coded data of a fixed length, as described later. The bit allocation calculation unit 40 includes an activity for each block transmitted from the block activity calculation unit 20 and an activity transmitted from the total activity calculation unit 26. Are used to calculate the coded bits of the AC component allocated to each block. The coded bits allocated to each block are two-dimensional Huffman coded for each block, and determine how far the data output from the low-frequency component should be cut off, that is, how many bits of the coded data are output. This is the specified number of bits.

The coded bit bit_B1 allocated to each block is given by the following equation.

bit_B1 = {(act_b / act_t) xbit [flag]} + {Carry Over} (1) In the above equation, bit [flag] represents the total number of bits of the entire image with respect to the AC component. This is the number of bits allocated to the entire image composed of a plurality of blocks. That is, when the two-dimensional Huffman-encoded data for each block is output in order from the low-frequency component of the AC component, when the output of the data is terminated with a predetermined number of bits, the AC code of the AC component as the entire image is output. Is the number of bits indicating how many bits are allocated to the coded data. The amount of data that is compression-encoded and output is determined by the number of bits.

act_b is an activity for each block. act
_t represents the total activity, and is the total value of the activities of the block. Therefore, ｛(act_b / act_
t) xbit [flag]｝ is the AC component A for the entire image
The number of bits allocated to the entire C-encoded data is allocated by the ratio of the activity of each block to the total value of the activity of the block.

{Carry Over} represents carry-over bits from the previous block, and as will be described later, coded bits bit_
This is the number of bits carried over from the block immediately before the block for which B1 is to be obtained, that is, the number of unused bits that have been allocated to the previous block.

The coded bits bit_B1 output from the bit allocation calculator 40 and allocated to each block are sent to the limiter 34. The limiter 34 outputs a coded bit bit_B2 obtained by limiting the coded bit bit_B1 distributed to each block sent from the bit distribution calculation unit 40 to a predetermined size as in the following equation.

bit_B2 = limit (bit_B1) (2) For example, the maximum value of the coded bit bit_B1 is 5 kbit, but the limiter 34 limits the maximum value to 1 kbit, and when the coded bit bit_B1 exceeds 1 kbit, , Limiter 34
Outputs 1 kbit.

The output from the limiter 34 is sent to the fixed length unit 36. The fixed length unit 36 limits the output so that the AC encoded data transmitted from the AC encoding 32 does not exceed the encoded bit bit_B2 transmitted from the limiter 34. Therefore, the AC encoded data is output to the output buffer 38 within the range of the encoded bit bit_B2.

The fixed-length unit 36 also determines the difference between the coded bits bit_B1 allocated to each of the blocks and the total number of bits of the AC coded data actually transmitted to the output buffer 38 as total carry-over bits ｛Carry Over｝. It is obtained by the following equation and sent to the bit allocation calculation unit 40.

Carry Over = (bit_B1)-(total_block_bits) ...
(3) The carry-over bit {Carry Over} is used for calculating the coded bit bit_B1 allocated to each block according to Expression (1) as described above.

AC coded data whose number of bits is limited by the fixed length unit 36 is input to the output buffer 38. Output buffer
38 stores DC encoded data of a fixed code length and AC encoded data with a limited number of bits, and these data are recorded on the memory card 50 through a connector.

Note that a recording medium such as a magnetic disk or an optical disk may be used as the recording medium in addition to the memory card 50.

According to this embodiment, the AC of each block is based on the ratio of the activity of each block to the total activity.
The coded bits to be allocated to the coded output of the component data are determined. Therefore, regarding the AC component data, the number of bits allocated to the entire image composed of a plurality of blocks is fixed, and the number of bits of the AC encoded output allocated to each block is distributed according to the activity of the block. The encoded output data can be a valid data amount depending on the activity of the block. In other words, a block containing many high frequency components has a large activity, so it allocates many bits,
Since a block containing a large amount of low frequency components has a small activity, a small number of bits can be allocated to output AC coded data corresponding to each block.

Moreover, since the number of bits allocated as described above is limited by the predetermined maximum value,
Since this can be limited when the number of bits of the AC component data is particularly large, the capacity of the output buffer can be reduced.

If the number of bits allocated to the AC component encoded data of a specific block is particularly large, high-quality image data is often used. The effect is small, and the image quality of the entire image is not significantly reduced.

Further, the difference between the number of allocated bits and the number of bits of actually output data is allocated to the next block as a carry-over bit, so that there is almost no reduction in bit use efficiency.

Advantageous Effects According to the present invention, since the compression encoding apparatus limits the data amount of the encoded output of each block according to the activity of each block with respect to the total value of the activity, each block outputs the data amount according to the frequency component. can do.

In addition, when the amount of data allocated to a specific block is large, the output data amount is limited by a predetermined upper limit value, so that the output buffer capacity can be reduced.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of an image signal compression encoding apparatus according to the present invention. FIG. 2 is a diagram showing an example of pixel data constituting a block. FIGS. 3A and 3B are diagrams showing activities. FIG. 4 is a diagram showing an example of a look-up table for converting the total value of the data into a normalization coefficient, FIG. 4 is a diagram showing an example of weight table data, and FIG. 5 performs encoding of run-length and non-zero amplitude. It is a figure showing an order. Description of codes of main parts 14 two-dimensional orthogonal transform unit 20 block activity calculating unit 26 total activity calculating unit 28 coding unit 30 DC coding unit 32 AC coding unit 34 … Limiter 36 …… Fixed length section 38 …… Output buffer 40 …… Bit allocation calculation section 50 …… Memory card

Continuation of front page (72) Inventor Mikio Watanabe 2-26-30 Nishiazabu, Minato-ku, Tokyo Fuji Photo Film Co., Ltd. (56) References 1989 IEICE Spring National Convention Lecture Paper [Separate Volume 7] Information System "D-159 Rate-Adaptive DCT Coding for Solid-State Electronic Still Cameras" Watanabe et al., (March 1989)

Claims (2)

(57) [Claims] 1. An image signal compression encoding apparatus which divides digital image data constituting one screen into a plurality of blocks and performs two-dimensional orthogonal transform encoding on the image data of each block. Orthogonal transform means for two-dimensional orthogonal transform of digital image data divided into blocks, encoding means for encoding data orthogonally transformed by the orthogonal transform means, activity of image data for each of the divided blocks Block activity calculating means for calculating the amount of coded data allocated to each block based on the activity for each block calculated by the block activity calculating means; and the coded data The block output from the encoding means in response to the output from the quantity distribution means. Encoding output control means for limiting the encoded data amount for each block, wherein the encoded data amount distribution means sets an upper limit value of the encoded data amount distributed based on the activity for each block to a predetermined value. Output to the encoded output control means with the value being limited to a value, and the encoded output control means performs encoding for each of the blocks output from the encoding means according to the encoded data amount limited by the upper limit value. An image signal compression encoding device characterized by limiting a data amount. 2. The apparatus according to claim 1, wherein said coded data amount allocating means actually calculates said coded data amount distributed to a previous block and said coded data amount for said previous block. The amount of coded data allocated to each block is added by adding the difference from the data amount output from the means to the amount of coded data allocated to each block based on the activity of each block as a carry-over data amount. Is calculated.