JPH02260872A - Picture signal compressing coding device and expansion decoder - Google Patents

Abstract

PURPOSE:To reduce the capacity of a lookup table used for Huffman coding by devising the subject device such that a shift area data detection means allows a coding means to apply coding when a digital picture data is a data within a prescribed range. CONSTITUTION:A shift area data detection means 26 is provided, which detects it when a digital picture data split into blocks is a data with a prescribed range with low production probability and the shift area data detection means 26 uses a code coding means 24 to code a digital picture data when the digital picture data is a data within a prescribed range and uses a Huffman coding means 20 to code the data when the digital picture data is a data at the outside of the prescribed range. When the data production probability is low, since coding is applied by using a shift code in place of the Huffman coding, the capacity of the memory storing lookup table data used for the huffman coding is reduced.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an image signal compression encoding device and an expansion decoding device, and more particularly to an image signal compression encoding device that reduces the amount of data in a Huffman encoding table used for Huffman encoding, and an image signal compression encoding device that reduces the amount of data in a Huffman encoding table used for Huffman encoding. The present invention relates to a decompression decoding device that decodes data encoded in this manner.

11 and l When storing digital image data such as image data taken by an electronic still camera in memory, in order to reduce the amount of data and the storage capacity of the memory,
Various compression encoding methods are used. In particular, two-dimensional orthogonal transform encoding is widely used because it can perform encoding at a high compression rate and can also suppress image distortion caused by encoding.

In such two-dimensional orthogonal transform encoding, image data is divided into a predetermined number of blocks, and the image data within each block is subjected to two-dimensional orthogonal transform. The DC (DCI component) of the orthogonally transformed data is, for example, after calculating the difference from the DCI component data of the previous block, and this difference value is Huffman encoded.

On the other hand, the orthogonally transformed image data, that is, the AC+AC component of the transform coefficient, is subjected to truncation (coefficient truncation) of a portion below a predetermined threshold value and quantization, ie, normalization, using a predetermined quantization step width. This suppresses the value of the conversion coefficient, that is, the dynamic range of the amplitude.

The conversion coefficient data of the normalized AC component often becomes 0 as it becomes a high frequency component, so the length of the 0 value data, so-called O run length, and the value of the non-0 value data. , run-length encoding is performed to convert the data into so-called non-zero amplitude data, and then compression-encoded image data is obtained by performing two-dimensional Huffman encoding on the run-length encoded data.

By the way, in two-dimensional Huffman encoding.

For each of the difference value of the DC IDc) component data and the run-length encoded AC fAcl component data, encoded data output by Huffman encoding is prepared in advance as a lookup table, corresponding to all possible values. It is necessary to remember.

In two-dimensional Huffman encoding, the above-mentioned direct current (DC)
Difference values of component data and run-length encoded AC (Considering the frequency of occurrence of MCI component data, short encoded data should be obtained for these values with a high degree of occurrence, and values with a low occurrence frequency. The data for encoding is stored in the lookup table mentioned above so as to obtain long encoded data for values that occur infrequently. This method has the disadvantage that coded data is stored, which requires a large capacity of the lookup table.

[The present invention solves the drawbacks of the prior art and provides an image signal compression encoding device that can reduce the capacity of a lookup table used in Huffman encoding, and data encoded by this device. An object of the present invention is to provide a decompression/decoding device for decoding.

According to the present invention, an image signal compression encoding device that divides digital image data constituting one screen into a plurality of blocks and compresses and encodes the image data of each block is provided. Huffman encoding means for Huffman encoding digital image data divided into blocks; code encoding means for converting digital image data into encoded data including a predetermined code; and code encoding means for converting digital image data into encoded data including a predetermined code; Shift area data detection means detects when the digital image data is within a predetermined range, and the shift area data detection means detects a code when the digital image data is within a predetermined range. When the digital image data is encoded by the encoding means and the digital image data is outside a predetermined range,
It is encoded by Huffman encoding means.

Further, according to the present invention, the image signal expansion and decoding device that receives and decodes compression-encoded digital image data of one screen includes Huffman decoding means that Huffman-decodes the input image data.

The code detecting means detects when the input image data includes a predetermined code, and the code decoding means decodes the image data including the predetermined code. If a code is detected from the input image data, the code is decoded by the code decoding means, and if no code is detected from the input image data, the Huffman decoding means is caused to perform Huffman decoding.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, embodiments of an image signal compression encoding device and an expansion decoding device according to the present invention will be described in detail with reference to the accompanying drawings.

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

This device has a blocking section 12. Blocking unit 12
is constituted by a frame buffer, and one frame worth of still image data captured by an electronic still camera is input through the image data input section 10 and stored. [stored in the blocking unit 12] frame image data is divided into a plurality of blocks, read out block by block, and sent to the average value calculation unit 148 and the two-dimensional orthogonal transformation unit 30.

The average value calculation unit 14 calculates the average value of the data for each block sent from the blocking unit 12, that is, the direct current (OC) component of the data of each block. The output of the average value calculation section 14 is sent to the difference value calculation section 16. Difference value calculation unit 16
calculates the difference between the average value of the block sent from the average value calculation unit 14 and the average value of the block immediately before that block, and outputs it to the address generation unit I8. The average value of the block output from the average value calculation section 14 is data (8 bits) in the range of 1, for example, 0 to 255, and the difference value output from the difference value calculation section 16 is data in the range of -255 to 255. The difference value (9 bits) output from the difference value calculation unit 16 is distributed with a probability as shown in FIG. As shown in the figure, the value "0" has the highest probability of occurrence, and the probability of occurrence decreases as the value moves away from 0. Values below -128 and above 128 have a significantly low probability of occurrence. The reason for such a distribution is that the average value (DC component) rarely changes significantly between adjacent blocks.

The address generation unit 18 generates an address for Huffman encoding based on the difference value sent from the difference value calculation unit I6. That is, the difference value data 255 to 255 sent from the difference value calculation unit 16 is converted into addresses I to 511. The output from the address generation section 1B is sent to the address identification circuit 26.

The address identification circuit 26 determines whether the difference value calculated by the difference value calculation part 16 is -255 to -128 or 128 to 2 in FIG. 6 depending on the address sent from the address generation part 18.
55 (shift area).

The output of the address identification circuit 26 is the lookup table 2
Sent to 8. The lookup table 28 includes a Huffman encoder 20 and a shift code generator 24.

If the difference value calculated by the difference value calculation unit 16 is within the range of the shift area, the address identification circuit 26 generates a shift code from the shift code generation unit 24 because the data has an extremely low probability of occurrence. let In this case, as shown in FIG. 3, data consisting of the shift code generated from the shift code generation section 24 and the difference value data sent from the difference value calculation section 16 is output to the DC code output section 22. Ru. The difference value data sent from the difference value calculation unit 16 is 9-bit data in the range of -255 to 255 as described above, but the difference value data in FIG. 3 when a shift code is generated is -255. Since it is in the range of ~-128 or 128-255, it can be represented by 8 bits.

In this way, when the difference value is data within the shift region with a low probability of occurrence, data consisting of the shift code and the difference value data is output to the DC code output section 22.

In the decoding device, the original difference value can be decoded from the shift code and the difference value data.

On the other hand, if the difference value calculated by the difference value calculation unit 16 is data outside the above shift area, that is, -
If the data is within the range of 127 to 127, the address identification circuit 26 detects this and outputs a control signal to the Huffman encoding unit 2o of the lookup table 28.

The Huffman encoding unit 2o performs Huffman encoding on the data from the address generation unit 18.

In this case, since the data has a relatively high probability of occurrence, the lookup table for Huffman encoding does not require a large amount of data;
Normal Huffman encoding is performed at DC code output section 2.
Output to 2.

FIG. 4 shows an example of data output from the DC code output section 22. In the example shown in the figure, the first difference value [25
J and the next difference value "7" are output as Huffman codes,
For the next difference value r-131J, data including a shift code is output. For the subsequent difference value "16", the Huffman code is output again.

In addition, in the case of the data within the above shift area, the address generation unit 18 generates the data from =255 to -12 as shown in FIG.
8 or 1 for data in the range 128-255! Since 28 addresses are output, the probability of occurrence of -128 is high.

Therefore, the length of the data encoded by the Huffman encoding section 20 is as shown in FIG.

The output from the blocking unit 12 is also: The image data is sent to the two-dimensional orthogonal transform unit 30, and the two-dimensional orthogonal transform unit 30 performs two-dimensional orthogonal transform on the image data for each block. As the two-dimensional orthogonal transformation, well-known orthogonal transformations such as discrete cosine transformation and Hadamard transformation are used.

The image data for each block that has been subjected to two-dimensional orthogonal transformation in the two-dimensional orthogonal transformation unit 30 is arranged vertically and horizontally as shown in FIG. The following data will be obtained. The output of the two-dimensional orthogonal transform section 30 is sent to the normalization section 32.

The normalization unit 32 performs normalization after truncating coefficients on the image data that has been subjected to two-dimensional orthogonal transformation in the two-dimensional orthogonal transformation unit 30, that is, the transformation coefficients. is compared with a predetermined threshold, and the portion below the threshold is discarded. In normalization, the transform coefficients whose coefficients have been truncated are divided by a predetermined quantization step value, that is, the normalization coefficient α, and quantization is performed using the normalization coefficient α.

By the way, this normalization is performed by dividing the truncated transform coefficient by the value α of one normalization coefficient, and by normalizing it with the data stored in the weight table T as shown in Figure 16. The coefficient a may also be used in combination with the coefficient a, that is, the value α·T obtained by multiplying the data in table T in FIG. Divide. Normalization may be performed by When the transformation coefficient before normalization is represented by X, the transformation coefficient X° after normalization is expressed as: X'=X/α·T.

The normalized transform coefficients are arranged in blocks in the same way as the data before normalization shown in FIG. 14, and are scanned in a zigzag pattern in order from the low frequency component in the zigzag scanning section 34 as shown in FIG. and output.

The output of the zigzag scanning section 34 is detected by a non-zero detection section (not shown) composed of, for example, n inverters, and is then sent to the run length counting section 3.
Sent to 6. The run length counting unit 36 is constituted by, for example, an inverter and a counter, and counts the run length of data of zero rQJ. The counted zero run length and the amplitude of non-zero data that appears thereafter are output to the address generator 38.

In this embodiment, since the size of the block is 8x8=64, the maximum number of consecutive "0" data is 64, so the run length of zero is represented by 6-bit data as shown in FIG. 8A. be done. As shown in the figure, the non-zero amplitude is 3
Represented by bit data. The output of the run length count section 36 is output to the address generation section 38.

The address generating section 38 generates 9-bit data as an address, which is composed of 6-bit data indicating the above-mentioned zero run length and 3-bit data indicating the non-zero amplitude. In other words, 0 is created using 9 bits with 6 bits of zero run length as the upper address and 3 bits of non-zero amplitude as the lower address.
512 addresses of ~511 are generated.

The frequency of occurrence of this 512i address is shown in FIG. As shown in the figure, the frequency of occurrence of a specific address is zero. This is because the lower three bits constituted by non-zero amplitude never become O.

The output from the address generation section 38 is the address identification circuit 42.
sent to. The address identification circuit 42 is the address generation section 3
It is detected whether the 9-bit address sent from 8 exceeds 255 or not. If the address does not exceed 255, ie, if the zero run length does not exceed 32, the address identification circuit 42 further determines whether there is a predetermined overflow in the non-zero data. The output from address identification circuit 42 is sent to lookup table 44. The lookup table 44 includes a Huffman encoder 46 and a shift code generator 40.

If there is no overflow in the non-zero data, the address identification circuit 42 outputs a control signal to the Huffman encoder 46 of the lookup table 44, and the Huffman encoder 4
The Huffman encoding unit 46 performs normal two-dimensional Huffman encoding in step 6. The Huffman encoding unit 46 performs two-dimensional Huffman encoding on the address sent from the address generation unit 38, that is, data consisting of a zero run length and a non-zero amplitude. In this case, data consisting of a zero run length and non-zero amplitude sent from the address generation unit 38 has a somewhat high probability of occurrence, and the amount of data in the lookup table for Huffman encoding is too large. Therefore, ordinary two-dimensional eight-man encoding is performed. The two-dimensional Huffman encoded data is output to the AC code output section 48.

On the other hand, if the address exceeds 255, that is, if the zero run length exceeds 32 and the address falls within the shift area in FIG. A control signal is output to the converter 30. Even when the address does not exceed 255, if there is an overflow in the non-zero data, the address identification circuit 42 outputs a control signal to the shift code generation section 40 and the two-dimensional orthogonal transformation section 30.

When the shift code generating section 40 receives the control signal.

Using the transform coefficients sent from the two-dimensional orthogonal transform section 30, a shift code which is the output of the LUT corresponding to the 8-bit input as shown in FIG. 8B is generated. As shown in the figure, the shift code is 8-bit data, where the most significant bit is "1" and the next 4 bits are the two-dimensional orthogonal transform unit 30.
The upper 4 bits of the conversion coefficient sent from
The bit is roooJ. Such a shift code is a code indicating that the address generated by the address generation unit 38, that is, the data to be two-dimensionally Huffman encoded, deviates from a predetermined range.

Figure 13 shows the code lengths of Huffman encoded data and shift codes. Data is allocated, and longer encoded data are allocated to those that occur less frequently. The shift code is the output of the LIT corresponding to the 8-bit input, and the lower 3 bits are 0, so as shown in Figure 11, it corresponds to the part of the occurrence frequency that was originally 0, and the first
As shown in FIG. 3, the Huffman encoded data is allocated to a long part that was originally not used.

When a shift code is generated from the shift code generation section 40, as shown in FIG. The lower 12 bits of the transform coefficients sent from the AC code output section 48 are outputted to the AC code output section 48.

An example of data output from the AC code output unit 48 is shown in the 10th example.
As shown in the figure. In the example shown in the figure, since the first data has an amplitude range of 2 and a run length of 7, a normal Huffman code is output. The following data has an amplitude range of 7,
Since the run length is 33 and exceeds 32, data including the shift code is output.

The subsequent data has an amplitude range of 5. Since the run length is l, a normal Huffman code is output.

As described above, according to the present device, for the DC component, when the difference value data sent from the difference value calculation unit 16 is data in the shift area, the shift code generation unit 24 generates a shift code, and this and output the difference value data. On the other hand, if the difference value data is data outside the shift area, normal Huffman encoding is performed. Since it is not necessary to store the Kur-Nobu table data, the capacity of the lookup table can be reduced. Moreover,
Since the frequency of occurrence of shift regions that are not Huffman encoded is extremely small, the encoding efficiency hardly decreases.

On the other hand, regarding the encoding of AC component data, when the address generated from the address generation unit 38, that is, the data to be encoded, is outside the frequently occurring shift region and there is no overflow even in non-zero amplitudes, performs normal Huffman encoding in the Huffman encoder 46 using the lookup table from the Huffman LIT 44. If the data to be encoded is in the infrequently occurring shift region or if there is an overflow in the non-zero amplitude,
The shift code generator 40 generates the shift code as described above without performing normal Huffman encoding, and outputs a run length and a transform coefficient.

Since the shift code is assumed to be data corresponding to a part that does not exist in the Huffman encoded data, there is no confusion with the Huffman encoded data.Also, if the shift code is output, it can be used to convert the run length. Since the coefficients are output as they are, the decoding device can obtain pre-encoded data without Huffman decoding. Furthermore, since the shift code includes the upper 4 bits of the transform coefficient, only the lower 12 bits may be included in the data output together with the shift code.

As described above, the shift code, run length, and transform coefficient are output without performing normal Huffman encoding in shift regions where occurrence frequency is low. Therefore, no lookup table for Huffman encoding is required for this region.

Conventionally, this shift area also required a lookup table. Furthermore, this region occurs less frequently (,
Despite the fact that it is unlikely to be used, encoded data with a long data length must be stored in the lookup table, so the storage capacity of the lookup table had to be increased. .

On the other hand, according to this device, there is no need to store a lookup table data in the shift area.
The storage capacity of the lookup table 44 can be reduced. Furthermore, since the frequency of occurrence of shift regions that are not subjected to Huffman encoding is extremely small, the 2-encoding efficiency hardly decreases.

FIG. 2 shows an embodiment of an image signal expansion/decoding device according to the present invention. This device decompresses and decodes image data encoded by the device shown in FIG.

This device includes a DC code input section 60 and an AC code input section 74.
has. The DC code input section 60 receives the DC component of the image data compressed and encoded by the compression encoding apparatus shown in FIG. 1, and the AC code input section 74 receives the AC component.

From the DC code input section 60, data containing Huffman encoded data in the Huffman encoding section 20 of the apparatus shown in FIG. 1 and a shift code generated from the shift code generation section 24 is input. The input DC component data is sent to the shift code detection section 64. The shift code detection unit 64 detects the shift code added by the device shown in FIG. 1 from the input DC component data. The output of the shift code detection section 64 is sent to the Huffman decoding section 62 and the amplitude value detection section 66. The Huffman decoding unit 62 performs Huffman decoding on the data input from the DC code input unit 60 when the shift code detection unit 64 does not detect a shift code. The output of the Huffman decoding section 62 is sent to the pre-oC value addition section 68.

When the shift code detection section 64 detects the shift code, it sends the input data to the amplitude value detection section 66 . The amplitude value detection section 66 extracts an 8-bit DC difference value from the data including the shift code as shown in FIG. calculate. That is, since the difference value in this case is in the range of -255 to -128 or 128 to 255 in Figure 6, the 8-bit D extracted from the input data
Original difference value data can be created using the C difference value.

Both the output of the Huffman decoding section 62 and the output of the amplitude value detection section 66 are the above-mentioned difference value data, and these are
In the C-value addition unit 68, this is added to the already calculated average value (DC component value) of the previous block to obtain the average value (DC component value) of that block. Front DC value addition section 6
The output of 8 is sent to the block synthesis section 70.

On the other hand, from the AC code input section 74, data containing Huffman encoded data in the Huffman encoding section 46 of the apparatus shown in FIG. 1 and a shift code generated from the shift code generation section 40 are input. The input AC component data is sent to the shift code detection section 84. The shift code detection section 84 detects the shift code added by the device shown in FIG. 1 from the input AC component data. The output from the shift codo detection section 84 is sent to a Huffman decoding run length decoding section 76 and a DCT coefficient reading running length decoding section 86.

The Huffman decoding run length decoding unit 76 detects the AC
Data input from the code input unit 74 is Huffman decoded. The output of the Huffman decoder 76 is sent to a raster scan converter 78.

The raster scan conversion unit 78 performs scanning according to the zero run length and non-zero amplitude values represented by the data input from the Huffman decoding run length decoding unit 76, and restores the data arranged vertically and horizontally. Raster scan converter 78
The output of is sent to the denormalization section 8o, and denormalization is performed in the denormalization section 80. The output of the denormalization unit 80 is sent to the block synthesis unit 7o, where it is combined into blocks, and then sent to the orthogonal inverse transformation unit 82, where it is subjected to orthogonal inverse transformation.

When the shift code detection section 84 detects the shift code, it sends the input data to the length decoding section 86 which reads the CT coefficients. OCT coefficient reading running decoding unit 86
is 6 from data including a shift code as shown in Figure 9.
The bit run length value and the lower 12 bit data of the OCT coefficient are extracted. Furthermore, data of the upper 4 bits of the OCT coefficient as shown in FIG. 8B is extracted from the shift code data.

This results in a zero run length and a non-zero OCT coefficient of 1
The output of the 6-bit OCT coefficient read running decoder 86 is sent to the raster scan converter 78. The raster scan conversion section 78 performs scanning according to the zero run length and non-zero amplitude values represented by the data input from the OCT coefficient reading runs decoding section 86, and restores the data arranged vertically and horizontally. After the output of the raster scan conversion section 78 is sent to the block composition section 7o.

The 6 block synthesis section 7o, which is sent to the orthogonal inverse transformation section 82 and undergoes orthogonal inverse transformation, synthesizes the data of a plurality of blocks to obtain the original full-screen image data. Block synthesis section 7
The output of 0 is sent to the image data output section 72 through the orthogonal inverse transform section 82, and is output from the 1 image data output section 72.

According to the decoding device as described above, it is possible to decompress and decode image data compressed and encoded by the device shown in FIG.

For both DC and AC components, it is detected whether or not a shift code is included. If a shift code is included, the difference value is extracted from the input data for the DC component, and the difference value is extracted from the input data for the AC component. extracts the run length and OCT coefficients, and if a shift code is not included, decoding is performed by normal Huffman decoding.

Therefore, image data encoded by Huffman encoding can be decoded and reproduced, and data encoded using a shift code during compression encoding can also be decoded.

肱-1 According to the present invention, a compression encoding device uses a shift code instead of Huffman encoding when the probability of data occurrence is low in compression encoding of data that has been block-formed and orthogonally transformed. Therefore, the memory capacity for storing lookup table data used for Huffman encoding can be reduced.

Further, the decoding device can decode data encoded in this manner.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of an image signal compression encoding device according to the present invention, FIG. 2 is a block diagram showing an embodiment of an image signal expansion decoding device according to the present invention, and FIG. FIG. 4 is a diagram showing an example of DC component data including a shift code outputted from the device shown in FIG. 1, FIG. 5 is a diagram showing an example of DC component data outputted from the device shown in FIG. FIG. 2 is a diagram showing the probability of occurrence of data output from the difference value calculation unit in FIG. 1; FIG. 6 is a diagram showing the probability of occurrence of data output from the difference value calculation section in FIG. 1 using a shift code, and FIG. 7 is a diagram showing the probability of occurrence of data output from the difference value calculation section in FIG. FIG. 8A is a diagram showing output data from the address generation section of FIG. 1, and FIG. 8B is an example of a shift code generated by the shift code generation section of FIG. 1. FIG. 9 is a diagram showing an example of AC component data including a shift code output from the apparatus of FIG. 1. FIG. 1O is a diagram showing an example of AC component data output from the device of FIG. 1. FIG. 11 is a diagram showing the probability of occurrence of data output from the address generation section in FIG. 1, and FIG. 12 is a diagram showing the probability of occurrence of data output from the address generation section in FIG. FIG. 13 is a diagram showing the data length encoded by the Huffman encoding unit in FIG. 1. FIG. Figure 14 is a diagram showing an example of data that has been subjected to two-dimensional orthogonal transformation. Figure 15 is a diagram showing the order in which run lengths and non-zero amplitudes are encoded. Figure 16 is an example of weight table data. FIG. main . Average value calculation part Difference value calculation part Huffman encoding part Shift code generation part Address identification circuit Lookup table Two-dimensional orthogonal transformation part Normalization part Shift code generation part Address identification circuit Lookup table Huffman encoding part Huffman decoding unit Shift code detection unit Huffman decoding Run length decoding unit, inverse normalization unit Two-dimensional orthogonal inverse transformation unit Shift code detection unit Rev. Takao, add 4 to the shift code in Figure 3. Ice 4 Figure D (-p, "4 output zero run length and 6t--/t) 46Σ! Please see the page with 5 illustrations. rotom μm shift - 4 clothes race shift tro Aoi 1 trip yū 88 figure shift code tail q figure shift ']-ni' to feed the dragon milk/Q figure pcF+ number ratio ga calculation, N concave tail! Figure 2 Dorsibu with 4 proboscis mounds - ← Mathematics, 73 concave 16th garden

Claims (1)

[Claims] 1. An image signal compression encoding device that divides digital image data constituting one screen into a plurality of blocks and compresses and encodes the image data of each block, the device comprising: Huffman encoding means for Huffman encoding the digital image data divided into blocks; code encoding means for converting the digital image data into encoded data including a predetermined code; and the digital image divided into the blocks. shift area data detection means for detecting when the data is data within a predetermined range with a low probability of occurrence, and the shift area data detection means detects when the digital image data is data within the predetermined range. If the digital image data is outside the predetermined range, the digital image data is encoded by the code encoding means, and if the digital image data is outside the predetermined range, the Huffman encoding means encodes the digital image data. Features: Image signal compression encoding device. 2. The apparatus according to claim 1, wherein the Huffman encoding means performs Huffman encoding on a difference value between an average value of the digital image data divided into the plurality of blocks and an average value of the previous block. An image signal compression encoding device characterized in that: 3. In an image signal compression encoding device that 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, the device comprises: orthogonal transformation means for two-dimensional orthogonal transformation of the divided digital image data; normalization means for normalizing the data orthogonally transformed by the orthogonal transformation means; and Huffman encoding for the data normalized by the normalization means. Huffman encoding means for converting data normalized by the normalization means into encoded data including a predetermined code; and code encoding means for converting the data normalized by the normalization means into coded data including a predetermined code; shift area data detection means for detecting when the data is within a low predetermined range, and the shift area data detection means detects when the normalized data is within the predetermined range. The code encoding means encodes the normalized data, and when the normalized data is outside the predetermined range, the Huffman encoding means encodes the normalized data. An image signal compression encoding device characterized in that the image signal compression encoding device encodes the data. 4. The apparatus according to claim 3, wherein the Huffman encoding means encodes a zero run length and a non-zero amplitude value of the normalized data, and the code is configured to encode a zero run length and a non-zero amplitude value of the normalized data. An image signal compression encoding device characterized in that a value corresponding to a non-zero amplitude value of the data is set to 0, and this is used as an address of a Huffman lookup table. 5. An image signal expansion/decoding device that receives and decodes one screen of compression-encoded digital image data, the device comprising: Huffman decoding means for Huffman decoding the input image data; code detection means for detecting when the input image data includes a predetermined code; and code decoding means for decoding the image data including the predetermined code; The image is characterized in that when the code is detected from the image data, the code decoding means decodes it, and when the code is not detected from the input image data, the Huffman decoding means causes the Huffman decoding. Signal expansion decoding device. 6. The apparatus according to claim 5, wherein the Huffman decoding means and the code decoding means obtain a difference value between an average value of the digital image data divided into the blocks and an average value of the previous block. An image signal expansion/decoding device characterized by: 7. An image signal expansion decoding device that receives one screen of compression-encoded digital image data and performs two-dimensional orthogonal inverse transform decoding, the device comprising Huffman decoding means for Huffman decoding the input image data. and denormalization means for denormalizing the data decoded by the Huffman decoding means; code detection means for detecting when the input image data includes a predetermined code; code decoding means for decoding the image data including; and orthogonal inverse transform means for performing two-dimensional orthogonal inverse transform on data obtained by either the denormalization means or the code decoding means, the code detection means When the code is detected from the input image data, the code decoding means causes the code to be decoded, and when the code is not detected from the input image data, the Huffman decoding means causes the Huffman decoding means to decode the code. An image signal expansion and reproduction device characterized by: 8. The apparatus according to claim 7, wherein the Huffman decoding means decodes the input data to obtain a zero run length and a non-zero amplitude value. Device.