Digital Camera Patent Abstract
A color image data compression coding apparatus of the present invention
includes a signal processing section for processing RGB (Red, Green
and Blue) color image data input from an imaging device to thereby
output YC (luminance and chrominance) image data consisting of luminance
data and chrominance data. When a monochrome mode is set by a system
controller, the signal processing section fixes the values of chrominance
data to zero. The luminance data and fixed chrominance data are
written to a frame memory and then sequentially read out of the
memory in blocks component by component. A compression coding section
codes the image data red out of the memory by use of a JPEG (Joint
Photographic Coding Expert Group) system. As for the coded data
of the chrominance data, Huffman codes are assigned only to DC components
and EOB codes. As a result, an amount of codes produced by subtracting
the amount of codes allocated to the chrominance data from a desired
total amount of code is entirely allocated to the luminance data
as a target amount of code. After the luminance data have been coded
in the target amount, the chrominance data are sequentially coded.
A method for implementing the apparatus and a digital camera incorporating
the apparatus are also disclosed. Digital Camera Patent Claims
What is claimed is:
1. An apparatus for compressing and coding image data representative
of a color image and including a luminance component and chrominance
components, and outputting resulting coded image data, said apparatus
comprising:
a signal processing circuit for processing the image data to output
processed image data;
a storage for storing the processed image data and allowing the
processed image data to be read out in preselected blocks component
by component;
a compression coding circuit for compressing and coding the processed
image data read out of said storage component by component to output
coded data;
an outputting circuit for outputting the coded data; and
a system controller for controlling said compression coding circuit
in accordance with a mode for compressing the image data;
said system controller being operative in response to a monochrome
mode selected for compressing and coding the image data such that
the image data render a monochrome image to fix the chrominance
components of the image data to a single, fixed value, and control
said compression coding circuit to compress and code resulting fixed
chrominance components and the luminance component.
2. An apparatus in accordance with claim 1, wherein said signal
processing circuit transforms an RGB color image signal input thereto
to the image data including the luminance component and the chrominance
components, and fixes the chrominance components to the fixed value
when the monochrome mode is selected.
3. An apparatus in accordance with claim 2, further comprising
a signal generating circuit for generating the RGB color image signal.
4. An apparatus in accordance with claim 3, wherein said signal
generating circuit includes an imaging device for shooting a scene
and generating an image signal representative of the scene, said
imaging device comprising an RGB color filter arranged on a light-sensitive
surface thereof.
5. An apparatus in accordance with claim 1, wherein said outputting
circuit includes a recording device for recording the coded data
output from said compression coding circuit in a recording medium,
the coded data being recorded in the recording medium in a preselected
format.
6. An apparatus in accordance with claim 5, wherein the recording
medium comprises a memory card having a semiconductor memory therein,
said outputting circuit writing the coded data in the memory card.
7. An apparatus in accordance with claim 1, wherein said compression
coding circuit comprises:
an orthogonal transforming circuit for executing orthogonal transform
with the image data read out of said storage to thereby output corresponding
transform coefficients;
a quantizing circuit for normalizing the transform coefficients
to thereby output normalized transform coefficients; and
a coding circuit for assigning preselected codes to the normalized
transform coefficients.
8. An apparatus in accordance with claim 7, wherein said quantizing
circuit quantizes, among the transform coefficients read out of
said storage and subjected to orthogonal transform, AC components
of the chrominance components to all zero in each block.
9. An apparatus in accordance with claim 7, wherein said compression
coding circuit includes a code amount controller for limiting a
length of the data to be coded to less than a preselected length
inclusive.
10. An apparatus in accordance with claim 9, wherein said code
amount controller subtracts an amount of codes which the fixed chrominance
components will have when compressed and coded from a total amount
of codes having the preselected length, sets a resulting difference
as a target amount of codes to be allocated to the luminance component,
and limits the amount of codes of the luminance component to less
than the target amount of codes inclusive.
11. An apparatus in accordance with claim 1, wherein said compression
coding circuit executes compression coding based on a JPEG system.
12. A method of compressing and coding image data representative
of a color image and including a luminance component and chrominance
components, said method comprising the steps of:
(a) writing processed image data output from a signal processing
circuit in a storage, and reading the processed image data out of
said storage in preselected blocks component by component;
(b) causing a compression coding circuit to compress and code the
processed image data read out of said storage component by component;
(c) outputting coded data output from said compression coding circuit;
(c1) selecting a monochrome mode for compressing and coding the
image data such that the image data render a monochrome image;
(c2) fixing the chrominance components of said image data to a
single, fixed value; and
(c3) controlling said compression coding circuit to compress and
code resulting fixed chrominance components and the luminance component.
13. A method in accordance with claim 12, further comprising (d)
causing said signal processing circuit to fix, among the image data,
the chrominance components to the fixed value, step (a) comprising
(e) writing the luminance component processed by step (d) and the
chrominance components fixed to the fixed value in said storage,
and reading the luminance components and the chrominance components
out of the said storage in blocks component by component.
14. A method in accordance with claim 12, wherein step (c) comprises
(f) recording the coded data in a recording medium in a preselected
format.
15. A method in accordance with claim 12, wherein step (b) comprises:
(g) executing orthogonal transform with the image data read out
of said storage to thereby output transform coefficients;
(h) quantizing the transform coefficients to thereby output normalized
transform coefficients; and
(i) assigning preselected codes to the normalized transform coefficients.
16. A method in accordance with claim 15, wherein step (h) comprises
(j) quantizing, among the transform coefficients read out of said
storage and subjected to orthogonal transform, AC components of
the chrominance components to all zero in each block.
17. A method in accordance with claim 15, further comprising (k)
limiting the data to be compressed and coded to less than a preselected
length inclusive.
18. A method in accordance with claim 17, wherein step (k) comprises
(1) subtracting an amount of codes which the fixed chrominance components
will have when compressed and coded from a total amount of codes
having the preselected length, setting a resulting difference as
a target amount of codes to be allocated to the luminance component,
and limiting the amount of codes of the luminance component to less
than the target amount of codes inclusive.
19. A digital camera for compressing and coding image data representative
of a color image and including a luminance component and chrominance
components, and outputting resulting coded image data, said camera
comprising:
a signal processing circuit for processing the image data to output
processed image data;
a storage for storing the processed image data and allowing the
processed image data to be read out in preselected blocks component
by component;
a compression coding circuit for compressing and coding the processed
image data read out of said storage component by component to output
coded data;
an outputting circuit for outputting the coded data; and
a system controller for controlling said compression coding circuit
in accordance with a mode for compressing the image data;
said system controller being operative in response to a monochrome
mode selected for compressing and coding the image data such that
the image data render a monochrome image to fix the chrominance
components of the image data to a single, fixed value, and control
said compression coding circuit to compress and code resulting fixed
chrominance components and the luminance component;
said system controller also being operative in response to a command
signal for causing the color image to be so compressed and coded
as to turn out a monochrome image when reproduced being input on
an operation panel to select the monochrome mode, and fix the chrominance
components of the image data to the fixed value.
20. A camera in accordance with claim 19, wherein said signal processing
circuit transforms an RGB color image signal input thereto to the
image data including the luminance component and the chrominance
components, and fixes the chrominance components to the fixed value
when the monochrome mode is selected.
21. A camera in accordance with claim 20, further comprising a
signal generating circuit for generating the RGB color image signal.
22. A camera in accordance with claim 21, wherein said signal generating
circuit includes an imaging device for shooting a scene and generating
an image signal representative of the scene, said imaging device
comprising an RGB color filter arranged on a light-sensitive surface
thereof.
23. A camera in accordance with claim 19, wherein said outputting
circuit includes a recording device for recording the coded data
output from said compression coding circuit in a recording medium,
the coded data being recorded in the recording medium in a preselected
format.
24. A camera in accordance with claim 23, wherein the recording
medium comprises a memory card having a semiconductor memory therein,
said outputting circuit writing the coded data in the memory card.
25. A camera in accordance with claim 19, wherein said compression
coding circuit comprises:
an orthogonal transforming circuit for executing orthogonal transform
with the image data read out of said storage to thereby output corresponding
transform coefficients;
a quantizing circuit for normalizing the transform coefficients
to thereby output normalized transform coefficients; and
a coding circuit for assigning preselected codes to the normalized
transform coefficients.
26. A camera in accordance with claim 25, wherein said quantizing
circuit quantizes, among the transform coefficients read out of
said storage and subjected to orthogonal transform, AC components
of the chrominance components to all zero in each block.
27. A camera in accordance with claim 25, wherein said compression
coding circuit includes a code amount controller for limiting a
length of the data to be coded to less than a preselected length
inclusive.
28. A camera in accordance with claim 27, wherein said code amount
controller subtracts an amount of codes which the fixed chrominance
components will have when compressed and coded from a total amount
of codes having the preselected length, sets a resulting difference
as a target amount of codes to be allocated to the luminance component,
and limits the amount of codes of the luminance component to less
than the target amount of codes inclusive.
29. A camera in accordance with claim 19, wherein said compression
coding circuit executes compression coding based on a JPEG system.
Digital Camera Patent Description
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a method and an apparatus for
the compressing and coding of image data representative of a color
image, and a digital camera including the same. More particularly,
the present invention relates to an image processing method for
compressing and coding image data derived from, e.g., a shot in
either one of a color mode and a monochrome mode and then recording
them in a recording medium, an apparatus therefor, and a digital
camera including the same.
2. Description of the Background Art
A digital still camera, for example, belongs to a family of image
processing apparatuses of the type using a CCD (Charge Coupled Device)
image sensor or similar solid imaging device and a semiconductor
memory, magnetic disk or similar image data recording medium. The
digital still camera transforms a color image signal representative
of a scene picked up by the image sensor to a digital signal, compresses
and codes the transformed image data by bidimensional orthogonal
transform or similar transform system, and records the coded image
data in the recording medium.
In the digital still camera, color image data produced by, e.g.,
the block interleaving of luminance data and chrominance data are
subjected to bidimensional orthogonal transform and coding. The
resulting coded data are compressed and coded to turn out, e.g.,
a preselected amount of codes and then written to a desired recording
medium. For compression coding, use is made of, e.g., a JPEG (Joint
Photographic Coding Experts Group) system based on "ISO/IEC
DIS 10918-1". The processed image data can not only be printed,
but be further processed by a general-purpose personal computer
for any desired application.
However, the conventional digital still camera has some problems
left unsolved, as follows. The camera has customarily been constructed
principally to shoot a natural image or similar subject and output
a corresponding color image. The camera therefore cannot output,
e.g., line images, characters or monotone or similar monochrome
images without resorting to special arrangements. It follows that
to record monochrome image data in a recording medium, the recording
format or even the construction of the entire camera must be substantially
changed.
To record monochrome image data in a recording medium, the CCD
imaging sensor or similar imaging device may be implemented by an
exclusive imaging device for monochrome. Alternatively, in a monochrome
mode, an exclusive color filter for color images may be removed
from the front of the imaging device. Further, an exclusive signal
processing circuit for monochrome may be additionally built in the
camera in order to generate only the luminance component of an image
signal and subject it to compression coding.
However, adding an exclusive imaging unit or an exclusive signal
processing circuit for monochrome is not practicable without reconstructing
the entire camera. Also, image data output by a special compression
system dealing only with the luminance component of an image signal
cannot be reproduced by a conventional reproducing system. Moreover,
a special black-and-white mode must be prepared even for the format
for recording the image data. Therefore, the camera, like an exclusive
monochrome camera, cannot be compatible with conventional cameras
or image data reproducing apparatuses, i.e., it does not allow image
data output therefrom to be reproduced by the conventional apparatuses.
Japanese patent laid-open publication No. 107904/1990 assigned
to the same assignee as the present application teaches a digital
still camera capable of recording, when a black-and-white mode is
selected, only original image data relating to a luminance component
and subjected to compression coding in both of the luminance component
area and chrominance component area of a memory cartridge. This
camera, however, does not give full consideration to the compatibility
of recorded data. For example, it is difficult with this type of
camera to decode recorded data within the range of a standardized
basic system and reproduce an image.
As stated above, it is necessary to substantially change the construction
of a conventional digital still camera of the type outputting color
image data and recording them in a recording medium. This increases
the cost and makes it difficult to implement a simple and convenient
digital still camera capable of outputting not only a color image
but also a monochrome image, as desired. For example, it is difficult
to realize a simple and convenient digital still camera adaptive
to both of a color image and a monochrome image and outputting image
data while insuring compatibility therebetween.
SUMMARY OF THE INVENTION
It is therefore an object of the present invention to provide a
method and an apparatus for the compression coding of image data
and capable of outputting coded data representative of a monochrome
image with a simple configuration while being compatible with a
conventional image processing system of the type outputting color
image data, and a digital still camera including the same.
In accordance with the present invention, an apparatus for compressing
and coding image data representative of a color image and including
a luminance component and chrominance components and outputting
the resulting coded image data includes a signal processing circuit
for processing the image data to thereby output processed image
data. A storage stores the processed image data and allows them
to be read out in preselected blocks component by component. A compression
coding circuit compresses and codes the processed image data read
out of the storage component by component to thereby output coded
data. An outputting circuit outputs the coded data. A system controller
controls a mode for compressing the image data. When a monochrome
mode for compressing and coding the image data such that the image
data render a monochrome image is selected, the system controller
causes the chrominance components of the image data to be fixed
to a preselected value, and causes the resulting fixed chrominance
components and luminance component to be compressed and coded by
the compression coding circuit.
Also, in accordance with the present invention, a method of compressing
and coding image data representative of a color image and including
a luminance component and chrominance components begins with a step
of writing processed image data output from a signal processing
circuit in a storage, and reading the processed image data out of
the storage in preselected blocks component by component. A compression
coding circuit is caused to compress and code the processed image
data read out of the storage component by component. Coded data
output from the compression coding circuit are output. When a monochrome
mode for compressing and coding the image data such that the image
data render a monochrome image is selected, the chrominance components
of the image data are fixed to a preselected value, and the resulting
fixed chrominance components and the luminance component are compressed
and coded by the compression coding circuit.
Further, in accordance with the present invention, a digital camera
for compressing and coding image data representative of a color
image and including a luminance component and chrominance components,
and outputting the resulting coded image data, includes a signal
processing circuit for processing the image data to thereby output
processed image data. A storage stores the processed image data
and allows them to be read out in preselected blocks component by
component. A compression coding circuit compresses and codes the
processed image data read out of the storage component by component
to thereby output coded data. An outputting circuit outputs the
coded data. A system controller controls a mode for compressing
the image data. The system controller causes, when a monochrome
mode for compressing and coding the image data such that the image
data render a monochrome image is selected, the chrominance components
of the image data to be fixed to a preselected value, and causing
the resulting fixed chrominance components and luminance component
to be compressed and coded by the compression coding circuit. When
a command signal for causing the color image to be so compressed
and coded as to turn out a monochrome image when reproduced is input
on an operation panel, the system controller selects the monochrome
mode and fixes the chrominance components of the image data to the
preselected value.
BRIEF DESCRIPTION OF THE DRAWINGS
The objects and features of the present invention will become more
apparent from the consideration of the following detailed description
taken in conjunction with the accompanying drawings in which:
FIG. 1 is a block diagram schematically showing a digital still
camera embodying the present invention;
FIG. 2 is a block diagram schematically showing a compression coding
section included in the embodiment in detail;
FIG. 3 shows a specific format for recording coded data output
in the embodiment in a recording medium;
FIG. 4 shows a quantization table included in the format of FIG.
3 in detail;
FIG. 5 shows a Huffman table also included in the format of FIG.
3 in detail; and
FIG. 6 shows a frame header further included in the format of FIG.
3 in detail.
DESCRIPTION OF THE PREFERRED EMBODIMENT
Referring to FIG. 1 of the drawings, a digital still camera embodying
the present invention is shown and generally designated by the reference
numeral 10. As shown, the camera 10 includes an image pick-up 12
for picking up a desired scene and outputting an image signal representative
of the scene. The image signal is transformed to corresponding image
data, subjected to compression coding, and then written to a memory
card or similar recording medium, as will be described specifically
later. The camera 10 compresses and codes the image data by, e.g.,
the standardized JPEG system mentioned earlier and is selectively
operable in a color mode or a monochrome mode. The user of the camera
is capable of selecting either the color mode or the monochrome
mode on an operation panel 14. The portions of the camera 10 not
directly relevant to the understanding of the present invention
will not be shown or described. In the following description, signals
are represented by the reference numerals designating connection
lines on which they appear.
The image pick-up 12 is a unit for picking up a desired scene and
outputting an image signal representative of the scene. In the illustrative
embodiment, the image pick-up 12 includes an imaging device and
a driver for driving it. When a scene incident via a lens, not shown,
is focused on the light-sensitive surface of the imaging device,
the imaging device outputs a corresponding electric signal. The
image pick-up 12 further includes an iris mechanism, an optical
zoom mechanism, and an autofocus (AF) mechanism, although not shown
specifically. The lens has a macro-function for implementing, e.g.,
close shots and allows the user to shoot even a color or monochrome
printing on which a fine text pattern or graphic pattern is recorded.
The AF mechanism is implemented as a range finding and focus adjusting
mechanism based on a passive system using visible rays or an active
system using infrared rays. Alternatively, use may be made of a
contrast detection system for executing control in accordance with
the contrast of the image signal output from the imaging device.
With such an AF mechanism, the illustrative embodiment is capable
of outputting a more clear-cut image.
The imaging device included in the image pick-up 12 is advantageously
implemented by a CCD image sensor or similar solid imaging device.
The imaging device has several hundred thousands of pixels to one
million and several hundred thousands of pixels. An RGB (red, green
and blue) color filter is positioned on the light-sensitive surface
of the imaging device. In the illustrative embodiment, the imaging
device has, e.g., 640 pixels in the horizontal direction (H) and
480 pixels in the vertical direction (V), i.e., about 300,000 valid
pixels in total. The pixels are primary color square pixels. The
driver, not shown, drives the imaging device with a drive signal
so as to cause it to generate an ROB color image signal at the speed
of an electronic shutter corresponding to the drive. The ROB color
image signal is fed from the image pick-up 12 to a signal processing
16 via a connection line 100.
The signal processing 16 executes various kinds of processing including
tonality correction and color correction with the input ROB color
image signal and transforms the processed image signal to eight-bit
digital values, i.e., ROB color image data. Further, the signal
processing 16 converts the RGB color image data to YC (luminance
and chrominance) image data constituted by luminance data Y, chrominance
data Cr an chrominance data Cb in a ratio of 4:2:2. The chrominance
data Cr and Cb are produced by subsampling the RGB color image data
in the horizontal direction. With a 3.times.3 matrix circuit, for
example, the signal processing 16 multiplies each of the R, G and
B data components by a particular coefficient and adds the resulting
products for thereby producing luminance data Y. As for the chrominance
data Cr and Cb, the signal processing 16 multiplies each of the
R, G and B data components by a particular coefficient and then
performs addition and subtraction with the resulting products. If
desired, the signal processing 16 may be provided with a digital
image processing function for executing analog image signal processing
with the input RGB color image signal, digitizing the analog image
signal to produce digital RGB image data, and correcting the digital
RGB image data.
In the monochrome mode particular to the illustrative embodiment,
the signal processing 16 fixes, in response to a control signal
102 fed from a system controller 18, the values of the chrominance
data Cr and Cb to the center value "128 (0.times.80h)"
substantially representative of zero with hardware. As a result,
luminance data Y corresponding to the RGB color image signal 100
and chrominance data Cr and Cb having the above fixed value appear
on the output 104 of the signal processing 16. In the color mode
which is usually set up, the signal processing 16 outputs luminance
data Y and chrominance data Cr and Cb corresponding to the RGB color
image signal 100. In the monochrome mode, the chrominance signals
Cr and Cb may, of course, be fixed to any suitable value other than
"128". With a value other than "128", it is
possible to render a sepia image or similar monochrome image different
from a black-and-white image, as desired. The system controller
18 is implemented by a CPU (Central Processing Unit).
Assume that the control signal 102 input to the signal processing
16 is indicative of a high contrast shot in the monochrome mode.
Then, the signal processing 16 may compare the value of the luminance
data Y with a preselected threshold, fix the value either to a maximum
or a minimum luminance level on the basis of the result of comparison,
and output the fixed value. The signal processing 16 is capable
of executing contour enhancement with the input color image signal,
i.e., generating digital image signal having enhanced contours in
response to a control signal fed from the controller 18. Further,
the signal processing 16 has a digital zooming function for interpolating
the pixels of the digital image data to thereby output image data
representative of only a desired portion of the image in an enlarged
scale.
The output 104 of the signal processing 16 is connected to a memory
controller 20. A frame memory or storage 22 is connected to the
memory controller 20. The memory controller 20 controls the access
to the frame memory 22 for writing or reading image data in or out
of the memory 22. The frame memory 22 has storage areas respectively
assigned to the luminance data Y, chrominance data Cr and chrominance
data Cb constituting at least one frame of image. The frame memory
22 may have a capacity great enough to store image data representative
of a plurality of pictures, e.g., a plurality of frames continuously
picked up by the image pick-up 12. In such a case, the storage areas
of the frame memory 22 should advantageously be extendable.
Further, in response to a control signal fed from the controller
18 via a connection line 108, the memory controller 20 reads each
of the luminance data Y and chrominance data Cr and Cb out of the
frame memory 22 while dividing them into a plurality of blocks each
having 8 (H).times.8 (V) pixels. The luminance data Y and chrominance
data Cr and Cb so read out by the memory controller 20 component
by component are delivered to a compression coding 24 via a connection
line 110 component by component.
The compression coding 24 compresses and codes the block-by-block
image data 10 read out of the frame memory 22, using a baseline
DCT (Discrete Cosine Transform) coding scheme by way of example.
The crux is that the compression coding scheme be capable of compressing
and coding the luminance data Y and chrominance data Cr and Cb separately
and storing each of the coded data in a particular area. In the
illustrative embodiment, for a coded data frame particular to the
baseline DCT coding scheme, use is made of a noninterleave system
in which the coded luminance data Y, chrominance data Cr and chrominance
data Cb each is transferred independently of the others, so that
a single image is coded component by component.
Particularly, in the monochrome mode, the compression coding 24
makes the amount of AC codes to be allocated to the chrominance
data Cr and Cb zero. Specifically, as shown in FIG. 2, the compression
coding 24 includes a DCT/IDCT (Inverse Discrete Cosine Transform)
operation 30 for executing bidimensional orthogonal transform with
the block-by-block image data 10. In the illustrative embodiment,
bidimensional orthogonal transform is implemented by DCT. The DCTI/DCT
operation 30 arranges vertically and horizontally the DC components
and AC components of DCT transform coefficients generated by bidimensional
orthogonal transform. Subsequently, the operation 30 rearranges
the AC components and DC components such that the DC components
occupy the top left portion and are sequentially replaced by the
high frequency data of the AC components toward the bottom right.
The output 200 of the operation 30 is connected to a quantization/dequantization
32.
The quantization/dequantization 32 normalizes the DC components
and AC components output from the DCT/IDCT operation 30, using a
quantization table 34 listing quantizing steps respectively assigned
to the DC components and AC components. Specifically, a code amount
controller 36 feeds a control signal 202 to the quantization/dequantization
32 so as to cause it to select an adequate quantization table in
accordance with, among others, the characteristic of an input image.
The normalized DCT transform coefficients are arranged in blocks
and then sequentially scanned zigzag, low frequency AC components
being first. The transform coefficients so scanned are sequentially
fed to a Huffman coding/decoding 38 via a connection line 204.
The Huffman coding/decoding 38 executes entropy coding with the
block-by-block quantized data applied to its input 204. In the illustrative
embodiment, the coding/decoding 38 codes the input data by assigning
Huffman codes listed in a Huffman table 40 to the input data, as
follows. First, the coding/decoding 38 codes an estimated value
of the input DC component, i.e., a difference between the DC component
value of an input block and that of the immediately preceding block.
Subsequently, the coding/decoding 38 codes the AC components scanned
zigzag from the low frequency side to the high frequency side. Thereafter,
the coding/decoding 38 inserts an EOB (End Of Block) code after
a code corresponding to the last valid component, thereby indicating
the end of the block. The coding/decoding 38 sequentially performs
the above processing with the luminance data Y and chrominance data
Cr and Cb. The block-by-block coded data are sequentially fed from
the coding/decoding 38 to an I/O (Input/Output) buffer 42 via a
connection line 206. Further, in response to a control signal 208
fed from the code amount controller 36, the coding/decoding 38 determines
the final valid components by cutting off the data to be coded,
e.g., limits the coded data to less than a preselected length inclusive.
The code amount controller 36 controls the amount of the compressed
and coded data to a total amount of codes matching with the compression
coding accuracy. Specifically, the controller 36 controls the quantization/dequantization
32, Huffman coding/decoding 38 and I/O buffer 42 such that the data
output from the coding/decoding 38 have less than a preselected
length inclusive, e.g., a fixed length. Further, the controller
36 controls the amount of codes of the decoded data by a two- path
control scheme, i.e., by calculating an activity representative
of the characteristic of an image, selecting an adequate quantizing
step based on the calculated activity, compressing and coding image
data with the quantizing step, and again compressing the image data
on the basis of the amount of resulting codes. Alternatively, the
controller 36 may use a single path control scheme or any other
suitable control scheme, if desired.
Assume that the system controller (CPU) 18 sends mode data indicative
of the monochrome mode to the code amount controller 36. Then, the
controller 36 controls the Huffman coding/decoding 38 by setting
the amount of AC codes which is zero, as mentioned earlier. Specifically,
the controller 36 assigns a target code amount of zero to the AC
codes of the chrominance data Cr and Cb. As for the luminance data
Y, the controller 36 selects as a target amount of codes a value
produced by subtracting the amount of DC codes and the amount of
EOB codes relating to the chrominance data Cr and Cb from the total
amount of codes. The controller 36 sets such two target amounts
allocated to the chrominance data and luminance data, respectively,
in the Huffman coding/decoding 38.
For example, assume that the number of DC components is zero. Then,
on a JPEG default Huffman table, the chrominance data Cr and Cb
need two bits for each block and two bits for an EOB code, i.e.,
four bits multiplied by the number of blocks for all the blocks.
Therefore, in the case of the 4:2:2 compression scheme using the
640 (H).times.480 (V) pixel matrix, the target amount of codes to
be allocated to the chrominance data Cr and Cb is 4,800 bits (2
bits.times.2,400 blocks) for each of the DC components and EOB codes,
i.e., 9,600 bits in total. The illustrative embodiment adds extra
sixteen bits to 9,600 bits in consideration of a difference between
the actual DC difference and the initial value. As a result, the
controller 36 assigns 9,616 bits in total to the entire chrominance
data Cr and Cb. Assuming a compression accuracy of 2 bpp (bits per
pel or pixel), then the total amount of codes corresponding to image
data is 614,400 bits. Therefore, as for the luminance data Y, the
controller 36 selects 604,784 bits produced by subtracting 9,626
bits allocated to the chrominance data Cr and Cb from the 614,400-bit
total amount of codes as a target amount of codes.
The I/O buffer 42 is a buffer memory for temporarily storing the
coded data output from the Huffman coding/decoding 38 under the
control of the code amount controller 36 and then delivering them
to a record controller 50, FIG. 1, connected to the output of the
compression coding 24.
The record controller 50 shown in FIG. 1 controls the recording
of the coded data received from the compression coding 24 in a recording
medium 52 removably fitted in the camera 10. In the illustrative
embodiment, the recording medium 52 may advantageously be implemented
by a memory card having a semiconductor memory thereinside. Of course,
the memory card may be replaced with an optical disk, magnetic disk
or similar rotary recording medium or a magnetic tape, optical or
magnetic card or similar sequential recording medium. Further, the
recording medium 52 may even be implemented as a device capable
of sending the coded data to another information equipment, e.g.,
an information recording apparatus connectable to the camera 1 via
analog and digital paths.
FIG. 3 shows a format in which data are written to the recording
medium 52 in the illustrative embodiment. As shown, the format corresponds
to the JPEG baseline DCT format. In the illustrative embodiment,
the camera 10 is so constructed as to write the coded data output
from the Huffman coding/decoding 38 in the format shown in FIG.
3. The format allows both of the coded data derived from the color
mode and the coded data derived from the monochrome mode to be written
to the medium 52. The coded data stored in the medium 52 are reproduced
by a conventional reproducing apparatus adaptive to the camera 10
and JPEG system.
As shown in FIG. 3, the format basically conforms to "Requirements
and Guideline; ISO/IEC DIS 10918-1" of a color still image
compression standard (JPEG). Compression coded data constituting
a single image are made up of a compressed data start code SOI (Start
of Image) heading the format, information APP1 (Reserved for Application
Segments) attached to an application, a quantization table DQT (Define
Quantization Table), a Huffman coding table DHT (Define Huffman
Table), a frame header SOF (Start of Frame), a scan header SOS (Start
of Scan), entropy coded data which are actual compressed data, and
an entropy coded data end code EOI (End of Image).
The compressed data start code SOI is a marker heading the entire
image and showing the beginning of compressed data. The information
APP1 attached to an application is a marker code to be freely used
by an application. In the illustrative embodiment, the information
APP1 may be used to record data not used for recording an image
in the monochrome mode, but used for another purpose, e.g., data
representative of a reduced version or thumbnail image of the coded
data recorded in a frame. This kind of information can be used by
an application, as needed.
The quantization table DQT is representative of table numbers,
accuracy and step sizes for quantization. As shown in FIG. 4 in
detail, the quantization table DQT lists a marker prefix, DQT representative
of the definition of the table, a field length representative of
a size following DQT, accuracy (Pq=0 (eight bits)) and table numbers
(Nq=0, 1, 2), and quantization tables (Y, B-Y and R-Y) for luminance
data (Y) and chrominance data B-Y(Cb) and R-Y(Cr).
The Huffman coding table DHT lists information relating to Huffman
tables, i.e., table numbers, a code length, and code symbols. As
shown in FIG. 5 in detail, the table DHT shows a marker prefix,
DHT representative of the Huffman table, a field length representative
of a size following DHT, table numbers assigned to the DC and AC
components of the luminance data and chrominance data, and DHT parameters
representative of Huffman code symbols.
The frame header SOF is a marker designating a coding algorithm
and an entropy coding system and representative of the beginning
of a frame. In the illustrative embodiment, the frame header SOF
shows SOF.sub.0 representative of a baseline DCT frame. As shown
in FIG. 6 in detail, the frame header SOF indicates a marker prefix,
SOF representative of the frame header, a field length showing a
size following the header, the accuracy of data representative of
the number of bits for a pixel, the number of vertical lines and
the number of horizontal lines representative of the size of an
image in combination, and the number of components (three) representative
of the number of components constituting a frame. Each component
(1 (Y), 2 (B-Y) or 3 (R-Y)) consists of a component number, a horizontal
and vertical sampling ratio (H0 and V0, H1 and V1 or H2 and V2),
and a quantization table number to be assigned to the component.
The scan header SOS is information representative of a scan head
marker. In the illustrative embodiment, the frame segment has three
kinds of scanning respectively assigned to luminance and chrominance
components; one scan header is assigned to each kind of scanning.
The entropy coded data end code EOI is a marker representative
of the end of entropy coded data, i.e., the end of the entire image.
Referring again to FIG. 1, the system controller 18 controls the
entire camera 10 and is implemented by a microcomputer or similar
processing system. The system controller 18 feeds necessary reference
signals to the various constituents of the camera 10. For example,
when the system controller 18 is triggered by a release signal received
from the operation panel 14, it controls the image pick-up 12, signal
processing 16, memory controller 20, and compression coding 24.
Further, the system controller 18 controls the writing or reading
of data in or out of the recording medium 52 removably mounted to
the record controller 50. Particularly, in the illustrative embodiment,
the system controller 18 selects either one of the color mode and
monochrome mode in accordance with a command input on the operation
panel 14, and generates control signals for setting the mode selected
in the various sections of the camera 10.
The display 54 outputs various kinds of information meant for the
user and including the operation statuses of the camera 10. When
the display 54 is implemented by a color LCD (Liquid Crystal Display)
panel or similar display device, it may display an image corresponding
to image data picked up, if desired.
The operation of the camera 10 having the above construction will
be described hereinafter. First, assume that the user of the camera
10 selects the color mode on the operation panel 14. Then, in response
to a release command also input on the operation panel 14, the system
controller 18 feeds a control signal to the image pick-up 12, commanding
it to shoot a scene. On the receipt of the control signal, the image
pick-up 12 controls the focus and iris and then controls the electronic
shutter so as to generate an image signal representative of the
scene. The image signal is subjected to tonality and color correction
and then converted to RGB color image data by the signal processing
16. Further, the RGB color data are transformed to YC image data
consisting of luminance data Y and chrominance data Cr and Cb and
then written to the frame memory 22 under the control of the memory
controller 20.
After image data representative of one frame have been written
to the frame memory 22, luminance data Y included in the image data
are sequentially read out in 8.times.8 pixel blocks and fed to the
compression coding 24. Subsequently, chrominance data Cr and Cb
also included in the above image data are sequentially read out
of the frame memory 22 in blocks and also fed to the compression
and coding 24. In the compressing coding 24, the DCT/IDCT operation
30 executes DCT with the luminance data Y block by block, rearranges
the resulting DC components and AC components, and outputs the rearranged
two kinds of components. Thereafter, the DCT/IDCT operation 30 executes
DCT with the chrominance data Cr and Cb block by block, rearranges
the resulting DC components and AC components, and outputs the rearranged
components.
The quantization/dequantization 32 receives the DC components and
AC components from the DCT/IDCT operation 30 and normalizes each
of them by using a particular quantizing step stored in the quantization
table. The transform coefficients of each normalized component are
arranged in blocks and then scanned zigzag, the DC components being
first, while being sequentially output. After the DC components,
the AC components are scanned in the ascending order with respect
to frequency.
The Huffman coding/decoding 38 codes each transform coefficient
output from the quantization/dequantization 32 by assigning a particular
Huffman code listed in the Huffman table thereto. Specifically,
as for the first block of DC components belonging to each of the
luminance data Y and chrominance data Cr and Cb, the coding/decoding
38 directly codes their values. As for the subsequent blocks of
DC components, the coding/decoding 38 codes a difference between
the value of each DC component and that of the DC component of the
immediately preceding block to thereby output an estimated value.
The coding/decoding 38 sequentially scans the AC components of each
component zigzag in the ascending order with respect to frequency.
At this instant, the code amount controller 36 causes the coding/decoding
38 to end coding as soon as it codes the last valid component. The
coding/decoding 38 outputs the sequence of coded data of each component
after adding the EOB code to the end of the sequence.
The coded data output from the Huffman coding/decoding 38 are once
stored in the I/O buffer 42 and then fed to the record controller
50 connected to the compression coding 24. The record controller
50 arranges the input coded data in a format matching with the recording
medium 52, and then writes them in the medium 52.
As stated above, in the color mode which is usually selected, a
full-color image identical with a desired scene is recorded in the
recording medium 52. The coded data stored in the medium 52 by the
above procedure may be read out and reproduced by the camera 10.
For example, the compression coding 24 receives via the I/O buffer
42 the coded data read out of the medium 52 by the record controller
50. In the compression coding 24, the Huffman coding/decoding 38
decodes the received coded data, and the quantization/dequantization
32 dequantizes the resulting decoded data. Subsequently, the DCT/IDCT
operation 30 executes IDCT with the data output from the dequantization
32. The decoded image data output from the decoding 38 are once
written to the frame memory 22 via the memory controller 20 and
then fed to the display 54, so that the user can see an image represented
by the above image data on the display 54.
Now, assume that the user selects the monochrome mode on the operation
panel 14. Then, in response to a release command also input on the
operation panel 14, an image signal output from the image pick-up
12 is fed to the signal processing 16, as during the color mode
operation. The signal processing 16 executes tonality and color
correction with the input image signal and coverts the corrected
image signal to digital values. When the controller 18 delivers
the control signal 102 commanding contour enhancement to the signal
processing 16, the signal processing 16 enhances the contours of
the image signal before digitizing the signal.
Further, the signal processing 16 produces luminance data Y and
chrominance data Cr and Cb from the digital RGB data by calculation.
The luminance data Y are directly written to the frame memory 22
via the memory controller 20. The chrominance data Cr and Cb are
fixed to "128" without exception and then written to the
frame memory 22 via the memory controller 20.
The luminance data Y and chrominance data Cr and Cb stored in the
frame memory 22 by the above procedure are sequentially read out
component by component, as during the color mode operation. A control
signal indicative of the monochrome mode is fed from the system
controller 18 to the compression coding 24. In response, the code
amount controller 36 calculates an amount of codes to be allocated
to the luminance data Y. For example, in the illustrative embodiment,
the controller 36 subtracts the amount of DC codes and that of EOB
codes allocated to the chrominance data Cr and Cb from the total
amount of codes determined by the accuracy of compression coding,
as stated earlier. The controller 36 determines the resulting difference
to be the target amount of codes to be allocated to the luminance
data Y, and sets it in the Huffman coding/decoding 38.
The DCT/IDCT operation 30 executes DCT with the luminance data
Y block by block, rearranges the DC components and AC components
of the resulting transform coefficients, and delivers the rearranged
DC components and DC components to the quantization/dequantization
32. The quantization/dequantization 32 normalizes the transform
coefficients by using a quantizing step assigned to the luminance
data Y. The normalized transform coefficients are arranged in blocks
and then scanned zigzag, the DC components being first, while being
sequentially output. After the DC components, the AC components
are scanned in the ascending order with respect to frequency.
The Huffman coding/decoding 38 codes each transform coefficient
output from the quantization/dequantization 32 by assigning a particular
Huffman code listed in the Huffman table thereto. Specifically,
as for the first block of DC components, the coding/decoding 38
codes the values of the DC components. As for the subsequent blocks
of DC components, the coding/decoding 38 codes a difference between
the value of each DC component and that of the DC component of the
immediately preceding block to thereby output an estimated value.
The coding/decoding 38 sequentially scans the AC components of each
component zigzag in the ascending order with respect to frequency.
At this instant, the code amount controller 36 causes the coding/decoding
38 to end coding on the basis of the target amount of codes set
by the code amount controller 36. The coding/decoding 38 outputs
the sequence of coded data after adding the EOB code to the end
of the sequence.
Subsequently, the chrominance data Cr are read out of the frame
memory 22 block by block while being input to the compression coding
24. The chrominance data Cr with the fixed value "128"
are subjected to DCT by the DCT/IDCT operation 30 and then normalized
by the quantization/dequantization 32. The resulting transform coefficients
output from the quantization/dequantization 32 are input to the
Huffman coding/decoding 38.
The coding/decoding 38 outputs the block-by-block transform coefficients
while assigning Huffman codes to the DC components included in the
transform coefficients. Specifically, the coding/decoding 38 assigns
Huffman codes to the DC components of the first block. As for the
subsequent blocks, the coding/decoding 38 assigns a Huffman code
to a difference between each DC component and the DC component of
the first block. Because the above difference is the same throughout
all the blocks following the first block, i.e., the DC different
is zero, the coding/decoding 38 assigns a two-bit Huffman code to
such blocks in accordance with the Huffman table. On the other hand,
the coding/decoding 38 does not assign any Huffman code to the AC
components of the chrominance data Cr because the amount of codes
to be allocated to the AC components is zero, as determined by the
code amount controller 36. Instead, the coding/decoding 38 adds
a two-bit EOB code to the end of each block of DC components provided
with Huffman codes and thereby codes the chrominance data Cr. The
coded data representative of the chrominance data Cr are written
to the preselected area of the recording medium 52 via the I/O buffer
42 and record controller 50. Thereafter, the chrominance data Cb,
like the chrominance data Cr, are read out of the frame memory 22,
compressed and coded, and then written to the recording medium 52.
As stated above, in the monochrome mode, coded image data representative
of a monochrome image corresponding to a desired scene are written
to the recording medium 52 in the format of a color image. It follows
that the coded image data can be read out of the medium 52 and reproduced
on the camera 10 in the same manner as the coded image data generated
in the color mode. The difference is that the image to appear on,
e.g., the color LCD panel or similar display 54 is monochrome because
the chrominance data Cr and Cb are fixed to zero.
In the above embodiment, the signal processing 16 is assumed to
fix the chrominance data Cr and Cb to "128" substantially
representative of zero. Alternatively, the quantization/dequantization
32 may be so constructed as to quantize the AC components of the
chrominance data to zero and normalize them. This alternative scheme
is particularly desirable when the signal processing 16 is not capable
of fixing the chrominance data to any particular value. As for the
control over the amount of codes, a target amount of codes may be
set, as in the illustrative embodiment.
In the illustrative embodiment, even in the monochrome mode, coded
data are written to the recording medium 52 in the form of a color
image. However, only DC components and EOB codes are actually assigned
to chrominance data, i.e., only the minimum necessary codes to implement
a color image are recorded in the medium 52. Such data can therefore
be reproduced in the same manner as a color image, and in addition
even a black-and-white or similar monochrome image can be reproduced.
The embodiment is therefore capable of outputting a monochrome
image with a compression coding system customarily installed in
an electronic still camera only if its parameters are changed. Particularly,
in the monochrome mode, the amount of codes to be allocated to the
chrominance data can be allocated to the luminance component, enhancing
image quality for a given total amount of codes.
The illustrative embodiment has concentrated on a digital still
camera selectively operable in a color mode or a monochrome mode.
The present invention is similarly applicable to, e.g., a compression
coding device having a color and a monochrome compression coding
mode, or a recording device having a color mode and a monochrome
mode and capable of recording image data after compressing and coding
them.
While the compression of a still image has been shown and described
as being implemented by the JPEG system, the present invention is
applicable even to, e.g., a moving picture compression system which
codes a plurality of consecutive frames at a time by compression.
Further, the illustrative embodiment is advantageously applicable
to a full-color image data format, e.g., Exif (Exchangable image
file format).
The present invention is practicable without regard to the control
over the amount of codes. For example, the present invention is
applicable to a two-path code amount control system having a first
and a second step. In the first step, the system calculates an activity
representative of the characteristic of an image, determines a quantizing
step by feeding the calculated activity forward, and compresses
image data with the quantizing step. In the second step, the system
corrects and updates the quantizing step on the basis of the amount
of codes by feedback processing and again compresses the image data
with the corrected quantizing step. In this manner, the present
invention is applicable to both of the feed forward system and feedback
system.
The embodiment shown and described directly compresses and codes
luminance data Y. Alternatively, the signal processing 16 may be
so constructed as to transform the luminance data Y to the bilevel
data taking either of the maximum and minimum levels with respect
to a preselected threshold, and compress the bilevel luminance data
and the fixed chrominance data. With this alternative configuration,
it is possible to generate coded data representative of a more clear-cut
bilevel image despite the color image outputting and recording system.
In summary, in accordance with the present invention, coded data
having a color image format, but which will turn out a monochrome
image when reproduced, are achievable with a simple construction
which changes the values of chrominance components of color image
data, fixes them to a preselected value, and then executes compression
coding. Because a minimum amount of codes should be allocated to
the chrominance components, the remaining amount of codes can be
fully allocated to a luminance component. Therefore, in a monochrome
mode, high resolution coded data clearly rendering a high definition
image with monotone can be output for the same amount of codes as
in a color mode as an entire image. The coded data so output can
be reproduced in the same manner as data representative of a usual
color image, so that compatibility and general-purpose application
of coded data are guaranteed. Moreover, control over the amount
of codes is practicable without resorting to sophisticated processing.
That is, the total amount of codes of coded data can be controlled
below desired one in the same manner as when a usual color image
is compressed and coded.
The entire disclosure of Japanese patent application No. 69788/1997
filed on Mar. 24, 1997 including the specification, claims, accompanying
drawings and abstract of the disclosure is incorporated herein by
reference in its entirety.
While the present invention has been described with reference to
the particular illustrative embodiment, it is not to be restricted
by those embodiment. It is to be appreciated that those skilled
in the art can change or modify the embodiment without departing
from the scope and spirit of the present invention.
|