JPH04180355A - Density correction device in image recorder - Google Patents

Abstract

PURPOSE:To obtain a conversion table fitted in desired characteristic by changing the discrete data of displacement distance for a coordinate point to be corrected on a reference conversion table virtually by applying interpolation, and displacing it by prescribed distance in a normal direction. CONSTITUTION:A reference input image signal Xi is inputted from a reference image signal generation circuit 28 to a signal conversion circuit 22, and the reference input image signal Xi is converted to a light modulation signal based on a reference conversion table. A correction coordinate point decided from density di' obtained when image recording is performed actually and an image signal Xi' to be conformed to the density at every reference input image signal Xi, and an interpolation arithmetic operation for an input image signal Xi'' is applied to the discrete data of the displacement distance ri, and the data of the displacement distance for each input image signal X can be obtained as the continuous data virtually. Thereby, it is possible to set the conversion table capable of satisfying the desired characteristic of the recording density for the input image signal X.

Description

[Detailed Description of the Invention] <Industrial Application Field> The present invention relates to a density correction device in an image recording apparatus, and in particular, to prevent variations in the density of a recorded image due to variations in various factors related to image recording. The present invention relates to a density correction device for the purpose of.

<Prior art> Conventionally, as an image recording device, a light beam is modulated based on an image signal, the modulated light beam is scanned over a photosensitive material as a recording medium, and the exposed photosensitive material is developed. An optical scanning recording device is known (Japanese Unexamined Patent Publication Nos. 54-42243 and 59-8).
(See Publication No. 3150, etc.).

An example of the optical scanning recording device will be explained based on FIG. 7. A light beam (laser beam) 2 emitted from a semiconductor laser 1 is made into a parallel beam by a collimator lens 3 and then reflected by a mirror 4. After that, the light enters an optical deflector 5 such as a rotating polygon mirror. The light beam 2 is reflected and deflected by the optical deflector 5 and is passed through a focusing lens 6 onto a photosensitive material 7.
The top is scanned (main scan) in a direction substantially perpendicular to the direction of arrow Y.

At the same time, the photosensitive material 7 is conveyed in the direction of the arrow Y by a conveying device (not shown) for sub-scanning, whereby the photosensitive material 7 is irradiated with the light beam 2 two-dimensionally.

The light beam 2 is modulated based on an image signal output from an image signal output device, and the photosensitive material 7 exposed to the modulated light beam is coated with the image signal or a photographic latent image. recorded as an image.

The above-mentioned image signal is a digital signal, and as shown in FIG. 8, it is passed through an image memory 8 and a signal conversion circuit 9, and is converted into an optical modulation signal based on a predetermined conversion table, and this optical modulation signal is After passing through the latch circuit 10, the D/A converter 1
] is converted to an analog signal. The analog optical modulation signal is input to the optical modulation circuit 12, and the optical modulation circuit 12 modulates the laser beam emitted from the semiconductor laser 1, and converts the modulated laser beam into a photosensitive material (film). 7, the image is recorded as a photographic latent image.

Furthermore, the exposed photosensitive material 7 is then sent to a known developing machine 13, where it is subjected to development, fixing, washing, and drying processes. Through this development process, the photographic latent image is developed, and an image carrying the image signal is recorded as a visible image on the photosensitive material 7.

Incidentally, in the above optical modulation circuit 12, the semiconductor laser l
However, when a gas laser is used as a light source, an intensity modulator such as an acousto-optic device (AOM) is usually used.

<Problems to be Solved by the Invention> By the way, FIG. 9 shows an example of an image signal and density characteristics recorded on a photosensitive material, that is, gradation characteristics. As is clear from this figure, the gradation characteristics are generally not linear with respect to the image signal, and it is impossible to obtain a good gradation image as is.

The reasons why the recording density does not become linear with respect to the image signal as described above include the following.

(1) The characteristics of a recording medium such as a photosensitive material are not linear.

(The relationship between the logarithm of the amount of light and the concentration is approximately linear.) (2)
The characteristics of an intensity modulator such as an acousto-optic device (AOM) are not linear.

(3) When a semiconductor laser is used, the relationship between the current supplied to the semiconductor laser and the amount of emitted light is not linear.

For this reason, a signal conversion circuit 9 as shown in FIG. 8 described above is usually provided to convert the image signal into an inverse function of the function representing the relationship between the image signal and the recording density, as shown in FIG. 10, for example. A table is provided in the signal conversion circuit 9, and the input image signal is converted based on the conversion table, so that the total characteristic (image signal vs. n density) is made as close to a linear characteristic as possible. There is.

However, even if the same image recording device and developing machine are used and a certain type of photosensitive material is used,
Due to differences in machines, differences in adjustment conditions, differences in environmental conditions, changes in the exposure light source over time, changes in photosensitive material lots, etc.
The recording density may change, and for example, in the recording of medical images that require particularly high gradations, the above-mentioned problem may impair the diagnostic performance of the image.

For this reason, there are conventional methods in which the conversion characteristics (conversion table) in the signal conversion circuit are changed in response to the above-mentioned changes in conditions (Japanese Patent Laid-Open No. 58-190
950, JP-A-59-83150, etc.).

For example, according to a conversion table setting method disclosed in Japanese Patent Application Laid-Open No. 58-190950, image recording and development are performed without using a conversion table based on a known image signal, and the density of the recorded and developed image is measured. A characteristic function of the recording density for the image signal is obtained based on a plurality of sets of the image signal and the obtained image density, and a conversion table is set as an inverse function of this function. Performs interpolation calculations on discrete data related to concentration,
Furthermore, a completely new conversion table for converting the input image signal into a light amount control signal (light modulation signal) by smoothing processing is set.

As mentioned above, data with discrete conversion characteristics consisting of density data obtained by actually performing image recording and development based on known image signals is converted into continuous data by performing interpolation calculations and smoothing processing. When setting a new conversion table as data, there is a problem that if the number of sample data is small, the interpolation error for the desired characteristics between sample data becomes large, making it impossible to accurately obtain the desired conversion characteristics. This may be due to the desired conversion characteristic, or it is generally a curve as shown in Figure 10, and the rate of change (first derivative) of the curve is locally significantly large in the area corresponding to the image signal region R in Figure 10. It is caused by

On the other hand, in the method disclosed in Japanese Patent Application Laid-Open No. 59-83150, three reference recording densities are set, and a reference signal that should be the reference recording density in the reference gradation characteristic is determined, and based on this reference signal, Image recording and development are actually performed, and the density of the obtained recorded image is measured. Then, based on the obtained density data, the sensitivity (level of the gradation curve) and gradation degree (the slope of the gradation curve) in the standard gradation characteristic are determined.
The reference gradation characteristics are corrected so as to satisfy the characteristics.

However, according to the above method, if the density characteristics change in a small range, it is possible to sufficiently deal with it by correcting the standard gradation curve, or if the density characteristics change significantly, the change will occur. Since there are only a few parameters that indicate the state of There are problems, such as using it as a reference gradation characteristic is cumbersome and cumbersome.

The present invention was developed in view of the above-mentioned problems, and it is possible to accurately update and set an image signal conversion table even if the number of sample data is small or the characteristics change greatly, and it is possible to update and set an image signal conversion table with high accuracy. An object of the present invention is to provide a density correction device for an image recording apparatus that can obtain a conversion table corresponding to a change in conditions by using an arbitrary table as a standard conversion table.

<Means for Solving the Problems> Therefore, in the present invention, as shown in FIG. In an image recording apparatus configured to record a gradation image by scanning a recording medium with light modulated by a light modulation means based on a signal, a plurality of reference input image signals are converted based on a reference conversion table. input each to the optical modulation means,
density data detection means for obtaining density data of a recorded image actually obtained corresponding to each of a plurality of reference input image signals; and said density data detection means based on a preset relationship between the desired recorded image density and the input image signal. density-corresponding signal setting means for determining a desired input image signal corresponding to each of the plurality of density data detected in the two-dimensional coordinates consisting of the input image signal axis and the converted image signal axis in the standard conversion table; , a minimum distance to each correction coordinate point determined as a signal obtained by converting the desired input image signal set by the density corresponding signal setting means and the corresponding reference input image signal using the reference conversion table. a conversion table displacement calculation means for calculating the corrected coordinate point on the reference conversion table and its minimum distance;
interpolation calculation means for interpolating discrete data of the minimum distance to the coordinate point to be corrected on the standard conversion table obtained by the conversion table displacement calculation means; a conversion table correction setting means for correcting the reference conversion table based on minimum distance data and setting the corrected conversion table as a predetermined conversion table in the signal conversion means; I decided to do so.

<Operation> According to this configuration, a plurality of reference input image signals are each converted based on the reference conversion table and an image is recorded. If the relationship is satisfied, the density of the image recorded by converting the reference input image signal should match the desired recorded image density.

Here, for example, if a signal B is obtained by converting a reference input image signal based on a reference conversion table, and the recording density C obtained based on this signal B does not satisfy the desired characteristics and the image Assuming that there is an input image signal D to obtain density C, if the standard conversion table is corrected and set so that the input image signal is converted to signal B by the conversion table, the conversion table that satisfies the desired characteristics can be corrected. I know that it can be configured.

Here, when making a correction by displacing a point on the characteristic curve of the standard conversion table by a predetermined amount in the normal direction on the two-dimensional coordinates consisting of the input image signal axis and the converted image signal axis in the conversion table, , coordinate point after correction (corrected coordinate point)
Since it has been confirmed that it is sufficient to be a coordinate point consisting of the input image signal R and the converted image signal B, conversely, as mentioned above, the corrected coordinate point (D, B
) Once you know the coordinate point (corrected coordinate point) on the reference conversion table that is the minimum distance to ) and the minimum distance, you can determine how much the point on the conversion table should be displaced in its normal direction. It can be seen whether correction can be made to the characteristic point.

Therefore, by actually recording images based on each of a plurality of reference input image signals, the coordinate points (D, B)
Find multiple correction coordinate points (points that should pass through the conversion table after correction) that match the desired characteristics, and for each correction coordinate point (D, B), select the point on the reference conversion table that will become the coordinate point to be corrected. point, a displacement distance for moving the corrected coordinate point in the normal direction to the corrected coordinate point (D, B) (minimum distance between the converted coordinate point (D, B) and the reference conversion table), and seek.

Furthermore, if the discrete data of the displacement distance (minimum distance) with respect to the corrected coordinate point on the reference conversion table obtained as described above is interpolated to become virtually continuous data,
A conversion table matching desired characteristics can be obtained by displacing each point on the reference conversion table by a predetermined distance in the normal direction.

Here, even if the reference conversion table has linear characteristics, the displacement distance data function obtained morphometrically as described above does not have any part where the rate of change of the function is significantly large. Even if the number of data, in other words the number of reference input image signals, is small, the correction coordinate point (D).

Compared to the case of performing interpolation connecting B), it is possible to perform interpolation calculations with higher accuracy using a known interpolation calculation, and thus it is possible to correct and set the reference conversion table with higher accuracy.

<Example> The present invention will be explained in detail below.

In FIG. 2 showing one embodiment, the image signal is inputted to the image recording device as a digital signal, and this digital image signal is stored in the -H image memory 21.

The input image signals stored in the image memory 21 are sequentially input to a signal conversion circuit 22 as a signal conversion means, and are converted into optical modulation signals in the signal conversion circuit 22 according to a predetermined conversion table. The predetermined conversion table is configured to obtain desired characteristics of a recorded image with respect to an input image signal.
This is a conversion table for converting an input image signal into a light modulation signal.

The optical modulation signal obtained by conversion by the signal conversion circuit 22 passes through a latch circuit 23 and then is converted into an analog signal by a D/A converter 24.

The analog signal is then inputted to an optical modulation circuit 25 serving as an optical modulation means, and modulates a light beam emitted from the optical modulation circuit 25 or a light source such as a semiconductor laser (not shown). Note that when a semiconductor laser is used as a light source, the semiconductor laser is directly modulated and driven by the optical modulation circuit 25, and when a gas laser is used as a light source, an intensity modulator such as an acousto-optic device (AOM) or the like is used. Used as modulation means.

The modulated light beam is passed through a collimator lens.

A film (photosensitive material) 26 as a recording medium is scanned through a scanning means consisting of an optical device consisting of a mirror, a light deflector, a condensing lens, etc. and a conveying device, and an image or a photographic latent image is recorded on the film 26. be done.

Furthermore, the exposed film 26 is then sent to a known developing machine 27, where it is subjected to development, fixing, washing, and drying processes. Through this developing process, the photographic latent image is developed and recorded on the film 26 as a visible image, either the image signal or the carried image.

Further, the developing machine 27 may be provided integrally with the image recording device, and the film 26 exposed by the image recording device may be automatically sent to the developing machine 27 to be developed.
The image recording device and the developing machine 27 may be separate bodies, and the exposed film 26 may be manually transported and set in the developing machine 27.

Here, in the present invention, in addition to the above configuration, the signal conversion circuit 22
In order to quickly correct the signal conversion table in response to changes in recording density characteristics due to environmental changes or changes over time, the following configuration is provided.

That is, instead of the normal input image signal, a reference image signal generation circuit 28 is provided for checking the characteristics of the recording density actually obtained for the input image signal, and this reference image signal generation circuit 28 In response to a command from the conversion table correction circuit 29, a plurality of types of reference input image signals Xi for test patterns are set in advance for the signal conversion circuit 22.
(i=0.l.

2.3. ... n) is output.

Specifically, signals are generated so that images corresponding to each of the plurality of types of reference input image signals Xi are recorded in a predetermined size, for example, in the form of a fold.

The conversion table correction circuit 29 corrects the conversion table in the signal conversion circuit 22 to a characteristic suitable for the current recording conditions based on, for example, an external trigger signal, and is used when setting the conversion table correction. controls the switch SW to forcibly input the reference input image signal Xi from the reference image signal generation circuit 28 to the signal conversion circuit 22 instead of the image signal recorded in the image memory 21.

At this time, the signal conversion circuit 22 and subsequent parts perform the same processing as in normal image recording to record an image on the film 26 based on the reference input image signal Xi.

If such film 27 is developed, the reference input image signal X
The recorded image is visualized based on i, where:
A density measuring device 30 is provided, and a reference input image signal X
The density of the image recording (test pattern) based on i can be measured for each of the plurality of reference input image signals Xi.

Incidentally, in the case where the image recording device and the developing machine 27 are provided integrally and development is performed automatically, the reference input image signal
The density measurement by the density measuring device 30 can be automatically performed only when the recording based on i (test pattern recording) is performed.

Furthermore, if the developing device 27 is provided separately from the image recording device and optional development is performed, the density measurement of the test pattern may be performed manually.

The measurement results obtained by the concentration measuring device 30 are input to the conversion table correction circuit 29 online or by manual keyboard operation.

The conversion table correction circuit 29 updates the conversion table in the signal conversion circuit 22 based on the density characteristics as a result of actually recording an image based on the reference input image signal Xi as described above. Note that the reference input image signal may be set in advance, or may be arbitrarily set by a keyboard operation or the like.

In this embodiment, the function as concentration data detection means is performed by the concentration measuring device 30. Reference image signal generation circuit 28. The conversion table correction circuit 29 is realized by a conversion table correction circuit 29, and the conversion table correction circuit 29 further includes a concentration corresponding signal setting means and a conversion table displacement amount calculation means.

It also serves as interpolation calculation means and conversion table correction setting means.

Next, the correction setting of the conversion table performed by the above configuration will be explained in detail based on the flowchart of FIG. 3 and the diagrams of FIGS. 4 to 6.

In the flowchart of FIG. 3, first, in Sl, the reference input image signal Xi is inputted from the reference image signal generation circuit 28 to the signal conversion circuit 22, and in S2, the reference input image signal Xi is converted into an optical signal based on the reference conversion table. Convert to modulated signal.

In this embodiment, the conversion table is updated in accordance with changes in recording conditions by correcting the existing conversion table, and the reference conversion table is used by the signal conversion circuit 22 for conversion during actual image recording. It may be a conversion table that has a simple linear relationship between the input image signal and the optical modulation signal, or a conversion table that can approximately satisfy the desired characteristics of recording density for the image signal in advance. These may be stored and used, or of course, the configuration may be such that the conversion table updated and set last time is corrected again.

Also, the purpose of conversion table correction is, for example, the purpose of correcting slight variations caused by environmental conditions while using the same image recording device, developing machine, and film, or correction when changing the machine or film type. Depending on the
Any one of the plurality of reference conversion tables may be selected.

In S3, the reference human input image signal Xi is converted based on the reference conversion table, and light modulation is performed based on the optical modulation signal obtained, and the film 26 is scanned to perform a test corresponding to the reference input image signal Xi. A photographic latent image of the pattern is recorded.

Then, in S4, the recorded result is visualized by being developed with the developing device 27, and the recorded image density d obtained by the development is
In the next step S5, i° is measured by the density measuring device 30 for each of a plurality of reference input image signals Xi.

The following processes from 86 onwards are performed in an arithmetic unit (not shown) built in the conversion table correction circuit 29.

In S6, by referring to the desired recording density characteristic g(X) (see FIG. 4) for the image signal that has been set and stored in advance, it is possible to determine what is actually obtained as a result of image recording based on the reference input image signal Xi. The desired characteristic g (
X) and the image signal Xi゛(2g−I(d i')
).

In other words, if there is a standard conversion table that satisfies the desired characteristic g (X) as shown in FIG. Since the density di° was actually obtained instead of the density di, it can be seen that the reference conversion table is not a characteristic that accurately satisfies the desired characteristic g (X).

Therefore, the density di° actually obtained based on the reference input image signal Currently, it has been confirmed that the recording density obtained by converting the standard input image signal Xi using the standard conversion table is di゛. If recorded using the optical modulation signal f(Xi), the density di°
I know what I'm getting.

Therefore, in order to satisfy the desired characteristic g (X), the concentration di
In order to make .degree. correspond to the input image signal Xi'', the optical modulation signal f(Xi) obtained by converting the image signal Xi' using a conversion table is sufficient.

Here, as shown in FIG. 5, which shows the conversion table on two-dimensional coordinates with the input image signal on the horizontal axis and the light modulation signal (converted image signal) on the vertical axis, a characteristic curve f showing the reference conversion table is shown.
When correcting and setting a conversion table that matches the desired characteristics by moving a predetermined point on (X) in its normal direction, the coordinate point (X i' .

The corrected coordinate point (X i”,
It is good to displace f (X i'')) by Δri in its normal direction.

Therefore, in S7, as described above, the reference input image signal
The optical modulation signal f(X
i) and the image signal Xi' that should correspond to this density.
The corrected coordinate point (Xi', f (Xi)
) is obtained for each reference input image signal Xi, while this correction coordinate point (Xi".

The coordinate point to be corrected (X i”, f (X i”)) on the separation reference conversion table to be displaced to Find it as a coordinate point on the conversion table, and
The coordinate point to be corrected (X i”, f (X i”))
Data on the displacement distance Δri is obtained at each time.

In this embodiment, the corrected coordinate point (Xi', f(Xi))
A distinction was made by setting the sign of Δri to 10 when it exists above the characteristic curve f (X) of the reference conversion table, and by setting the sign of Δri to - when it exists below the characteristic curve f (X).

By the above processing, at least the reference input image signal XiO
The number of corrected coordinate points that match the number will be found,
In this embodiment, instead of directly interpolating the correction coordinate points and setting a conversion table, we check whether the characteristic curve connecting these correction coordinate points is a characteristic curve that matches the desired characteristics. Modify the reference conversion table to obtain a new conversion table.

That is, in the next step 88, a plurality of coordinate points to be corrected (X1°', f
The discrete data of the displacement distance △ri obtained for each input image signal X is interpolated with respect to the input image signal make sure you get it.

The above Δri converts the reference conversion table into the desired gradation characteristic g (
This is data for correcting the characteristics corresponding to g( (see Figure 6), even if interpolation is performed based on a relatively small number of Δri data, the interpolation accuracy will not decrease significantly, and the value to which each point of the reference conversion table should be moved can be changed between the sample data. can be estimated with high accuracy,
In addition, even if the required characteristics change greatly from the standard conversion table and the absolute value of Δri increases, the rate of change of Δri with respect to the image signal level will not become significantly large, so even if the required characteristics change greatly, the interpolation accuracy will increase. can respond by maintaining sufficient

In this embodiment, the amount by which the reference conversion table (see Figure 5) is moved in the image signal direction is determined, so the reference conversion table may be set to linear characteristics or may be set to non-linear characteristics. It may be. However, when setting non-linear characteristics, it is particularly desirable to use a reference conversion table that substantially matches the desired gradation characteristics.

In S9, the displacement distance Δr with respect to the above human-powered image signal
By displacing each corrected coordinate point on the reference conversion table based on the continuous data of
A conversion table that can satisfy the desired characteristics of recording density is set.

Specifically, the point (X
i”, f (X i”)), the distance Δr is
By plotting the point (X, F (X) ) separated by (X i”), the corrected conversion table F(X)
Can you get it?

In this way, the conversion table used in the signal conversion circuit 22 or the correction setting based on the relationship between the input image signal and the actually obtained recording density may be affected by changes in the environmental temperature or changes over time. Even if the exposure characteristics of the medium or the light intensity characteristics of the exposure light source change, or the development characteristics of the recording medium change due to development temperature, development speed, etc., it is not possible to stably obtain the desired recording density for the image signal. , especially when recording medical images (radiation images) that require high gradation.
This can improve diagnostic performance.

In addition, in the interpolation calculation of △ri mentioned above, spline interpolation ("Spline function and its application J model") is used.

(Reference: Yoshimoto, Kyoiku Publishing, 1979, etc.), linear interpolation.

Any known interpolation calculation method, such as a combination of linear interpolation and smoothing, or n-th order function approximation, may be used.

Also, the reference input image signal Xi and the corresponding density d
The number of data for i should be about 2 to 50 points, preferably about 4 to 30 points. In the configuration of the present invention, when the configuration includes a concentration measurement operation by the concentration data detection means or manual operation, the number of data of Xi and dl is as follows:
It is preferable to use a relatively small number of points, about 2 to 10 points, and as the interpolation calculation, spline interpolation, n-th function approximation (
It is particularly preferable to use a high-order interpolation calculation where n≧3), but this is not the case if the concentration data detection means or automatic concentration measuring device is used.

Further, in this embodiment, as shown in FIG. 4, the desired gradation characteristic of the recording density for the image signal is a linear characteristic, or the desired characteristic may be non-linear depending on individual requirements. The tonal characteristics are not limited to linear characteristics.

In addition, in this example, by using the conversion table related to the present invention alone, fluctuations in the exposure light source, fluctuations in the exposure characteristics of the film,
For example, a conversion table for the purpose of correcting fluctuations in the output characteristics of the exposure light source and a conversion table correction setting means dedicated to the conversion table may be provided independently. A conversion table for correcting the output characteristics of the exposure light source is set while appropriately correcting based on the data obtained by photoelectric conversion of the scanning light, and when recording an image, the conversion table for correcting the light source characteristics and the conversion table according to the present invention are provided. It is also possible to combine and use a conversion table set based on the recorded image density.

Furthermore, a new setting of the conversion table by correcting the reference conversion table as described above may be performed automatically, for example, each time the image recording device is powered on, or a trigger switch may be set arbitrarily. You may also let them do it.

<Effects of the Invention> As explained above, according to the present invention, when correcting an image signal conversion table that is set to satisfy the desired characteristics of recording density for an image signal in accordance with a change in conditions, it is possible to Even if the number of sample data to investigate the situation is small, or the characteristics change significantly, the image signal conversion table can be updated and set with high precision.
Furthermore, since the conversion table to be corrected can be simply set to a linear characteristic, it is possible to update the conversion table to stabilize or equalize the relationship between the image signal and the recording density, or to update the conversion table easily and with high precision. It has the effect of becoming

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of the present invention. FIG. 2 is a system block diagram showing an embodiment of the present invention. FIG. 3 is a flowchart for explaining how to set a new conversion table in the same embodiment. , FIGS. 4 to 6 are diagrams for explaining how the conversion table is updated in the above embodiment, respectively, and FIGS. 7 and 8 are system schematic diagrams showing an example of a conventional image recording apparatus, respectively. FIG. 9 is a diagram showing an example of recording density characteristics for image signals, and FIG. 10 is a diagram showing an example of an image signal conversion table for obtaining desired gradation characteristics.

Claims (1)

[Scope of Claims] A signal converting means for converting an input image signal based on a predetermined conversion table, and a light modulating means modulates light onto a recording medium based on the image signal converted by the signal converting means. In an image recording apparatus configured to record a gradation image by scanning, a plurality of reference input image signals are converted based on a reference conversion table and input to the light modulation means, and the plurality of reference input image signals a density data detection means for obtaining density data of a recorded image actually obtained corresponding to each; and a plurality of density data detection means detected by the density data detection means based on a relationship between a desired recorded image density and a preset input image signal. density-corresponding signal setting means for determining a desired input image signal corresponding to each density data; and density-corresponding signal setting means on a two-dimensional coordinate consisting of an input image signal axis and a converted image signal axis in the reference conversion table. a reference conversion table that provides a minimum distance to each correction coordinate point determined by a desired input image signal set by the means and a signal obtained by converting the corresponding reference input image signal with the reference conversion table; a conversion table displacement calculation means for calculating the coordinate points to be corrected and their minimum distances on the reference conversion table; an interpolation calculation means for correcting the reference conversion table based on the continuous minimum distance data to the coordinate point to be corrected interpolated by the interpolation calculation means, and applying the corrected conversion table to a predetermined value in the signal conversion means. A density correction device for an image recording apparatus, comprising: conversion table correction setting means for setting a conversion table;