JP2006284971A - Burning phenomenon correction method, self-light emitting apparatus, burning phenomenon correction apparatus and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a burning phenomenon correction method of a self-light emitting apparatus in which a plurality of self-light emitting elements are arranged in a matrix form, and to suppress expansion of degradation difference among pixels by making effect of reducing gradation difference larger from a low gradation region to a high gradation region, as variations of correction amounts become larger and as a result, by reducing luminance difference. <P>SOLUTION: The burning phenomenon correction apparatus for correcting the burning phenomenon of the self-light emitting apparatus in which the plurality of self-light emitting elements are arranged on a base in the matrix form, comprises; (1) a correction amount determination section for determining a correction amount corresponding to each pixel based on a degradation amount which is calculated in each pixel; (2) a variation determining section for calculating information indicating a variation rate of correction amount distribution; and (3) a gradation converting section for converting an input gradation value to an output gradation value by referring to a gamma curve having larger effect of reducing the gradation difference from the low gradation region to the high gradation region, as the variations of the correction amount distribution become larger. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  One embodiment of the present invention relates to a method for correcting a burn-in phenomenon that occurs in a self-luminous device. One embodiment of the present invention relates to a burn-in phenomenon correcting device and a self-light-emitting device equipped with the same. One embodiment of the present invention relates to a program for causing a computer mounted on a self-luminous device to execute a burn-in correction function.

Flat panel displays are widely used in products such as computer displays, portable terminals, and televisions. Currently, liquid crystal display panels are mainly used, but the narrow viewing angle and slow response speed continue to be pointed out.
On the other hand, an organic EL display formed of a self-luminous element can overcome the above-mentioned problems of viewing angle and responsiveness, and can achieve a thin form, high brightness, and high contrast that do not require a backlight. Therefore, it is expected as a next-generation display device that replaces the liquid crystal display.

By the way, it is generally known that organic EL elements and other self-light-emitting elements have a property of deteriorating depending on the light emission amount and the light emission time.
On the other hand, the content of the image displayed on the display is not uniform. For this reason, the deterioration of the self-luminous element is likely to proceed partially. For example, the self-light-emitting element in the time display area (fixed display area) progresses more rapidly than the self-light-emitting elements in other display areas (moving image display areas).
The luminance of the self-luminous element that has deteriorated is relatively lowered as compared with the luminance of other display areas. In general, this phenomenon is called “burn-in”. Hereinafter, partial deterioration of the self-luminous element is referred to as “burn-in”.

At present, various methods are being studied for improving the “burn-in” phenomenon. Some of them are listed below.
In this document, input data for each pixel constituting the display panel is integrated for each pixel at a constant period, and the integrated value of each pixel is subtracted from the maximum value of each pixel. A method for setting the correction amount is disclosed. Further, a method is disclosed in which the display characteristics of each pixel are made uniform by emitting each pixel with a constant luminance for a time proportional to the amount of correction in a non-use state.

In this document, display data and display time are stored only when a still image is displayed, and the difference ΔY between the display data and the maximum luminance and the time T when the still image is displayed are integrated. A method of setting the amount ΔY · T as correction data is disclosed. Also, this document discloses a method for correcting a burn-in phenomenon by executing a display for correction only when the lid is closed or not used. This correction method also has the same problem as in Patent Document 1. Japanese Patent Laid-Open No. 2000-132139 discloses a method for integrating input data for each pixel and converting the integrated value into a correction value using a correction table. In addition, a method is disclosed in which input data of each pixel is corrected with the obtained correction value to make it difficult to visually recognize the burn-in phenomenon.

JP-A-2001-175221 discloses a method for determining a correction value so as to lower the luminance data of other pixels in accordance with the pixel having the lowest luminance among the pixels. In addition, a method is disclosed in which luminance data of each pixel is converted with the obtained correction value so that the burn-in phenomenon is difficult to visually recognize.

  However, the existing correction method needs to change the image greatly in order to realize the correction of the burn-in phenomenon. That is, as a result, there is a possibility that the image quality is significantly deteriorated. In addition, the existing method has a problem that requires a large amount of calculation processing.

The inventors propose a method of executing the following processing as a correction processing method for a burn-in phenomenon that occurs in a self-light-emitting device in which a plurality of self-light-emitting elements are arranged in a matrix.
(1) Processing for determining a correction amount corresponding to each pixel based on the deterioration amount calculated for each pixel (2) Processing for calculating information indicating the degree of variation in the correction amount distribution (3) Distribution of correction amount Refers to a gamma curve that has a greater effect of compressing the gradation difference from the low gradation area to the high gradation area as the variance of the image becomes larger, and converts the input gradation value to the output gradation value

These technical methods can be applied not only to the self-luminous device itself, but also to various electronic devices that process image signals output to the output device. Further, these technical methods can be realized not only as hardware but also as software. Of course, part of the processing can be executed as hardware, and the remaining processing can be executed as software.
Incidentally, the self-luminous device includes an organic EL (electroluminescence) panel, a PDP (plasma display panel), a CRT (cathode ray tube), an FED (field emission display) panel, an LED panel, and a projector.

  By applying this technique, the effect of compressing the gradation difference from the low gradation area to the high gradation area can be increased as the variation in the correction amount increases. As a result, it is possible to suppress the brightness difference from being reduced and the deterioration amount difference between pixels from being increased.

Hereinafter, an embodiment example of a burn-in phenomenon correction technique that employs the technical technique according to the invention will be described. Hereinafter, the burn-in phenomenon correction device is referred to as a “correction device”.
In addition, the well-known or well-known technique of the said technical field is applied to the part which is not illustrated or described in particular in this specification.
The embodiment described below is one embodiment of the present invention and is not limited thereto.

(A) Form example of burn-in phenomenon correction apparatus (A-1) Form example 1
(A) Device Configuration FIG. 1 shows an example of a correction device. In the case of this embodiment, the correction device 1 is arranged for each pixel that emits light of the same color. For color image reproduction, R (red), G (green), and B (blue), which are the three primary colors of light, are used, and complementary colors are used as necessary. Note that the correction apparatus 1 can be arranged in common for each color signal.
The correction device 1 includes a deterioration amount calculation unit 3, a deterioration amount storage memory 5, a correction amount determination unit 7, a variation determination unit 9, and a gradation conversion unit 11 as main components.

The deterioration amount calculation unit 3 is a processing device that calculates the deterioration amount of each pixel based on an input signal before correction processing. In the case of this embodiment, the input signal is gradation data.
The deterioration amount may be an amount that can estimate the luminance deterioration of each pixel, and any calculation method can be applied. Of course, existing calculation methods can be applied.
For example, a method of accumulating gradation data corresponding to each pixel for each frame can be applied.
In addition, for example, when a proportional relationship is not established between the gradation data and the degree of progress of deterioration, a method of calculating the deterioration amount using a conversion coefficient reflecting an actual measurement result can be applied. The inventors define this conversion coefficient by the concept of “deterioration rate”. The deterioration rate is defined as the rate of decrease in light emission luminance when the self-light-emitting element continuously emits light with a certain gradation data.

FIG. 2 shows an example of a conversion table showing the correspondence between the gradation data and the deterioration rate. In this case, the deterioration amount is calculated as a value obtained by multiplying the deterioration rate R corresponding to each gradation data by the light emission period T.
However, a method of directly reading out the deterioration amount shown in FIG. 2 from the gradation data corresponding to each pixel may be applied. In the case of this embodiment, the deterioration amount corresponding to each pixel is treated by accumulating the deterioration amount for each frame read out in this manner for each frame.
In addition, a method of giving the difference of the deterioration amount of each pixel with respect to the deterioration amount of the reference pixel can be applied.

The deterioration amount storage memory 5 is a memory for storing a cumulative deterioration amount that is updated every frame.
The correction amount determination unit 7 is a processing device that determines a correction amount corresponding to each pixel based on the deterioration amount data read from the deterioration amount storage memory 5. The correction amount determination unit 7 determines the correction amount in a direction that eliminates the difference in deterioration amount existing between pixels.
However, any method can be applied to the correction amount determination method. This embodiment is characterized in that the variation of the determined correction amount is obtained and used, and it is not related to the method for determining the correction amount.
Therefore, the correction amount may be determined by, for example, applying a method for determining the correction amount so that the deterioration of other pixels catches up with the most deteriorated reference pixel. In this case, the correction amount is determined so as to increase the luminance of pixels other than the reference pixel. In this case, the correction amount is given as a “positive value”.

Further, for example, a method of determining the correction amount so that the deterioration of other pixels catches up with the reference pixel that is most delayed in deterioration may be applied. In this case, the correction amount is determined so as to lower the luminance of pixels other than the reference pixel. In this case, the correction amount is given as a “negative value”.
In the case of this embodiment, the correction amount is given by the difference in the accumulated deterioration amount existing between the reference pixel and each pixel.
The determined correction amount is given to the variation degree determination unit 9.
The variation determination unit 9 is a processing device that calculates information indicating the degree of variation of the correction amount distribution. Here, information indicating the degree of variation (hereinafter referred to as “variation value”) is given as the maximum value of the correction amount. When the correction amount is only a negative value, the absolute value of the correction amount is given as the largest value. When the correction amount is given as both a positive value and a negative value, the correction value is given as the largest absolute value.

The gradation conversion unit 11 refers to a gamma curve that has a larger effect of compressing the gradation difference between the low gradation region and the high gradation region of the output gradation value as the variation of the correction amount distribution increases. A processing device that converts values into output tone values.
FIG. 3 shows an internal configuration example of the gradation converting unit 11. The gradation conversion unit 11 includes a variation value memory 11A, a gamma curve generation unit 11B, and a gradation conversion execution unit 11C.
The variation value memory 11A is a memory used for storing variation values.
The gamma curve generation unit 11B is a processing device that generates a gamma curve having input / output characteristics according to the variation value.

The gamma curve is a curve that defines the correspondence between the input gradation value and the output gradation value, and in the case of this embodiment, its shape changes according to the variation value.
FIG. 4 shows an example of a gamma curve corresponding to the variation value. FIG. 4 is a diagram expressed in terms of gray scale. The horizontal axis represents the input gradation value, and the vertical axis represents the output gradation value.
When the variation value is zero, the gamma curve is a straight line as shown by a broken line. At this time, output gradation value = input gradation value.
When the variation value is non-zero, the gamma curve is an inverted S-shaped curve. This curve translates dark areas to lighter and bright areas to darker.

When this gamma curve is used, the gradation difference from the low gradation area to the high gradation area of the input display signal is compressed and output. That is, the input image can be converted into an output image with a small contrast.
Note that the gamma curve changes to a shape in which the intermediate gradation area is almost horizontal as the variation value increases. In FIG. 4, the change in shape is indicated by an arrow. It can be seen that the gamma curve changes from 0 → 1 → 2 as the variation value increases.

As long as such input / output characteristics are obtained, the gamma curve may be generated by any method.
For example, a method is used in which a gamma curve is generated for each of the low gradation region and the high gradation region with the intermediate gradation value (127) as a boundary. In this case, the gamma value b1 (<1) is applied to the low gradation range, and the output gradation value y is calculated by the input gradation value x to the b1th power. Further, the gamma value b2 (> 1) is applied to the high gradation region, and the output gradation value y is also calculated by the square of the input gradation value x.
It is assumed that the relationship between the gamma value to be used and the magnitude of the variation value is set in advance. Note that a gamma value of 1 is associated with a variation value of 0 (zero).
The gradation conversion execution unit 11C is a processing device that reads an output gradation value corresponding to an input gradation value with reference to the generated gamma curve.

(B) Correction Processing Operation FIG. 5 shows a processing procedure example executed by the correction device 1.
First, the deterioration amount calculation unit 3 calculates the deterioration amount for each pixel (S1). This process is executed based on gradation data that is an input signal.
Next, the correction amount determination unit 5 calculates a correction amount corresponding to each pixel based on the calculated deterioration amount (S2).
Next, the variation determination unit 9 compares the absolute values of the correction amounts calculated for all the pixels, and obtains the maximum value (S3).

When the maximum value is determined, the gradation converting unit 11 (gamma curve generating unit 11B) generates a gamma curve having a compression effect corresponding to the maximum value (S4). The generated gamma curve is stored in the memory in a table format in which the output gradation value is associated with the input gradation value.
The gradation conversion unit 11 (gradation conversion execution unit 11C) reads the output gradation value corresponding to the input gradation value with reference to the generated gamma curve (S5). The read output gradation value is output to the subsequent circuit as an output display signal.
This conversion process is repeatedly executed for each pixel until the data defining the input / output characteristics of the gamma curve is updated. This data update is executed, for example, every frame or every scene change.

(C) Effects of Embodiment As described above, using this correction device, the gradation difference appearing in the input display signal can be compressed and output as the correction amount increases. That is, the luminance difference between pixels constituting the output image can be made smaller than before gradation conversion, and the increase in the deterioration amount difference can be suppressed accordingly.
FIG. 6 shows a correspondence relationship between the light emission luminance values according to the input gradation values. In the figure, a gamma curve corresponding to a variation value of 0 (zero) is indicated by a broken line. The gamma curve when the variation value is small and large is indicated by a solid line.
As indicated by the arrows, the gamma curve changes from 0 → 1 → 2 as the variation value increases.
That is, the larger the variation value, the closer the gamma curve becomes horizontal. This means that the luminance difference is reduced.

In this gradation conversion, although the brightness difference is small, all display gradations of the original image can be stored in the output image. Therefore, as long as the degree of deformation of the gamma curve does not become considerably large, it is possible to add a process for correcting the burn-in phenomenon without visually recognizing a decrease in image quality.
Thus, the method according to the embodiment is advantageous for mounting on an actual product.
The gamma correction data can be generated by a simple calculation method. Therefore, the circuit scale of the burn-in correction circuit can be reduced.
In addition, the use of a relatively simple method called gamma conversion is advantageous in reducing the circuit scale.

(A-2) Embodiment 2
(A) Device Configuration FIG. 7 shows another example of the correction device. In FIG. 7, parts corresponding to those in FIG.
The correction device 21 includes a deterioration amount calculation unit 3, a deterioration amount storage memory 5, a correction amount determination unit 7, a variation determination unit 9, and a gradation conversion unit 23 as main components. The only new component is the gradation converter 23.
FIG. 8 shows an internal configuration example of the gradation conversion unit 23.

The gradation conversion unit 23 includes a variation value memory 21A, a gamma curve selection unit 21B, and a gradation conversion execution unit 21C. Among these, the variation value memory 21A is a memory used for storing variation values.
The gamma curve selection unit 21B is a processing device that selectively outputs gamma correction data corresponding to the variation value. It is assumed that the gamma curve selection unit 21B stores gamma curves 21B1, 21B2,... 21BN prepared according to each variation value.
The gradation conversion execution unit 21C is a processing device that actually converts an input gradation value into an output gradation value with reference to the generated gamma curve.

(B) Correction Processing Operation FIG. 9 shows a processing procedure example executed by the correction device 21.
Also in this case, the deterioration amount calculation unit 3 calculates the deterioration amount for each pixel (S11). This process is executed based on gradation data that is an input signal.
Next, the correction amount determination unit 5 calculates a correction amount corresponding to each pixel based on the calculated deterioration amount (S12).
Next, the variation determination unit 9 compares the absolute values of the correction amounts calculated for all the pixels and obtains the maximum value (S13).

When the maximum value is determined, the gradation conversion unit 23 (gamma curve selection unit 21B) selects a gamma curve having a compression effect corresponding to the maximum value (S14). The selected gamma curve is written in a table format in a reference memory.
The gradation conversion unit 23 (gradation conversion execution unit 21C) refers to the selected gamma curve and reads the output gradation value corresponding to the input gradation value (S15). The read output gradation value is output to the subsequent circuit as an output display signal.
This conversion process is repeatedly executed for each pixel until the data defining the input / output characteristics of the gamma curve is updated. This data update is also executed, for example, every frame or every scene change.

(C) Effect of Embodiment As described above, the same effect as Embodiment 1 can be realized even when this correction apparatus is used.
In the case of this correction apparatus, it is possible to reduce the processing load required for generating gamma correction data and reduce the processing time.

(A-3) Embodiment 3
FIG. 10 shows another example of the correction apparatus. In FIG. 10, the same reference numerals are assigned to the corresponding parts to those in FIG. The basic configuration and processing operation of the correction device 31 are the same as those in the first embodiment.
However, the correction device 31 is different from the correction device 1 in that the deterioration amount corresponding to each pixel is calculated based on the output display signal. Due to the difference in the configuration, the correction device 31 can directly reflect the actual light emission state in the calculation of the deterioration amount.
In the case of the first embodiment, a processing function for reflecting the correction effect in the gradation converting unit 11 on the deterioration amount is necessary to improve the accuracy of the burn-in correction. However, in the case of the correction device 31, such a processing function is not necessary, and a reduction in system scale and cost reduction can be realized. This correction device can be applied to other embodiments as well.

(B) Example of mounting on self-luminous device FIG. 11 shows an example of mounting the burn-in phenomenon correcting device on the self-luminous device.
The self-light-emitting device 41 includes a burn-in phenomenon correction device 45 and a display device 47 mounted on a housing 43.
Here, the burn-in phenomenon correcting device 45 corresponds to one of the above-described embodiments. The burn-in phenomenon correction device 45 inputs an image signal generated at an external terminal or inside, and executes an input signal correction operation so that a deterioration amount difference does not occur between the correction target pixel and the reference pixel.

The display device 47 is assumed to be composed of a display device and its drive circuit. As the display device, an organic EL (electroluminescence) panel, a PDP (plasma display panel), an FED (field emission display) panel, an LED panel, or a CRT is used.
In the case of FIG. 11, the self-light-emitting device 41 is illustrated as having a burn-in phenomenon correction device 45 that is a processing device dedicated to correction of the burn-in phenomenon. However, when all the functions are executed in software. These functions are realized by a computer mounted on the self-luminous device.

(C) Mounting Example on Image Processing Device FIG. 12 shows a system example of the image processing device 53 on which the burn-in correction device 51 is mounted. The image processing device 53 is connected to the self-luminous display device 55 via a wired path or a wireless path.
In the case of this system example, the burn-in correction process is executed in the housing of the image processing apparatus 53. That is, the image signal output to the display device 55 is input to the burn-in correction circuit 51 disposed between the output interface and the burn-in correction process described above is executed.

  The image processing device 53 of this type includes, for example, a video camera, a digital camera, and other imaging devices (including not only a camera unit but also a device integrated with a recording device), a computer (including a server). Various information processing terminals (portable computers, mobile phones, portable game machines, electronic notebooks, etc.), various image playback devices (including home servers), image editing devices, and game machines can be applied. is there.

(D) Other Embodiments (a) In the embodiment described above, the correction amount is given by the difference in the accumulated deterioration amount existing between the reference pixel and each pixel. However, it is possible to use a correction amount calculated by any method including a known technique.
(B) In the above-described embodiment, the case where the variation value is given as the maximum value of the correction amount has been described. However, the variation value may be given as a variance value or a standard deviation value.
(C) In the second embodiment, a case has been described in which data defining the input / output characteristics of the gamma curve is prepared in advance according to individual variation values.
However, this type of data may not be prepared for all variation values, but may be prepared so as to thin out the number thereof. For example, the variation value may be divided into a plurality of sections, and data defining input / output characteristics of one gamma curve may be prepared for each section.

(D) In the above-described embodiment, the case where the gradation value is given by 8 bits (that is, the case where the gradation value is given by 0 to 255 has been described. However, the number of bits giving the gradation value may be other than 8 bits. The present invention can be applied to other cases such as 10 bits, 12 bits, and the like.
(E) In the above-described embodiment, the functional configuration of the burn-in phenomenon correction apparatus has been described. Needless to say, an equivalent function can be realized as hardware or software.
Further, not only all the functions constituting the burn-in phenomenon correcting apparatus are realized by hardware or software, but some of the functions can also be realized by using hardware or software. That is, a combination of hardware and software may be used.
(F) Various modifications can be considered for the above-described embodiments within the scope of the gist of the invention. Various modifications and application examples created based on the description of the present specification are also conceivable.

It is a figure which shows the example of a form of the burn-in phenomenon correction apparatus. It is a figure which shows the example of a conversion table holding the correspondence of a gradation value and a deterioration rate. It is a figure which shows the internal structural example of a gradation conversion part. It is a figure which shows the change of the shape of the gamma curve according to the variation value. It is a flowchart which shows the example of correction | amendment operation | movement of a burn-in phenomenon. It is a figure which shows the change according to the variation value of the gamma curve which gives the relationship between an input gradation value and light emission luminance. It is a figure which shows the other example of a burn-in phenomenon correction apparatus. It is a figure which shows the internal structural example of a gradation conversion part. It is a flowchart which shows the example of correction | amendment operation | movement of a burn-in phenomenon. It is a figure which shows the other example of a burn-in phenomenon correction apparatus. It is a figure which shows the system example which mounts the burn-in phenomenon correction apparatus in the self-light-emitting device. 1 is a diagram illustrating a system example in which a burn-in phenomenon correction apparatus is mounted on an image processing apparatus.

Explanation of symbols

1, 21, 31, 45, 51 Burn-in phenomenon correction device 3 Deterioration amount calculation unit 5 Deterioration amount storage memory 7 Correction amount determination unit 9 Variation determination unit 11, 23 Gradation conversion unit 11A, 21A Variation value memory 11B Gamma curve generation unit 11C, 21C gradation conversion execution unit 21B gamma curve selection unit

Claims (11)

A method of correcting a burn-in phenomenon of a self-light-emitting device in which a plurality of self-light-emitting elements are arranged in a matrix,
A process of determining a correction amount corresponding to each pixel based on the deterioration amount calculated for each pixel;
A process of calculating information indicating the degree of variation in the correction amount distribution;
A process of converting an input gradation value into an output gradation value by referring to a gamma curve that has a larger effect of compressing a gradation difference from a low gradation area to a high gradation area as the variation of the correction amount distribution increases. A burn-in phenomenon correction method comprising: In the burn-in phenomenon correction method according to claim 1,
The burn-in phenomenon correction method, wherein the correction amount is given as a difference between a deterioration amount of each pixel and a deterioration amount of a comparison target pixel. In the burn-in phenomenon correction method according to claim 1,
The burn-in phenomenon correction method, wherein the information indicating the degree of variation of the correction amount is given as an absolute value of the correction amount. In the burn-in phenomenon correction method according to claim 1,
The burn-in phenomenon correction method, wherein the information indicating the degree of variation in the correction amount is given as a dispersion value of the correction amount. In the burn-in phenomenon correction method according to claim 1,
The burn-in phenomenon correction method, wherein the information indicating the degree of variation in the correction amount is given as a standard deviation value of the correction amount. In the burn-in phenomenon correction method according to claim 1,
The burn-in phenomenon correction method, wherein the gamma curve has different compression effects in a low gradation region and a high gradation region. In the burn-in phenomenon correction method according to claim 1,
The burn-in phenomenon correction method, wherein the gamma curve is calculated based on information indicating the degree of variation. In the burn-in phenomenon correction method according to claim 1,
The burn-in phenomenon correction method, wherein the gamma curve corresponding to the degree of variation is selected from a plurality of types of gamma curves. A self-light-emitting device in which a plurality of self-light-emitting elements are arranged in a matrix,
A correction amount determination unit that determines a correction amount corresponding to each pixel based on the deterioration amount calculated for each pixel;
A variation determination unit that calculates information indicating a variation degree of the distribution of the correction amount;
A gradation that converts an input gradation value into an output gradation value by referring to a gamma curve that has a larger effect of compressing a gradation difference from a low gradation area to a high gradation area as the variation of the correction amount distribution increases. A self-luminous device comprising: a conversion unit. A burn-in phenomenon correction device that corrects a burn-in phenomenon of a self-light-emitting device in which a plurality of self-light-emitting elements are arranged in a matrix,
A correction amount determination unit that determines a correction amount corresponding to each pixel based on the deterioration amount calculated for each pixel;
A variation determination unit that calculates information indicating a variation degree of the distribution of the correction amount;
A gradation that converts an input gradation value into an output gradation value by referring to a gamma curve that has a larger effect of compressing a gradation difference from a low gradation area to a high gradation area as the variation of the correction amount distribution increases. A burn-in phenomenon correction apparatus comprising: a conversion unit. A process of determining a correction amount corresponding to each pixel based on the deterioration amount calculated for each pixel;
A process of calculating information indicating the degree of variation in the correction amount distribution;
A process of converting an input gradation value into an output gradation value by referring to a gamma curve that has a larger effect of compressing a gradation difference from a low gradation area to a high gradation area as the variation of the correction amount distribution increases. A program for correcting a burn-in phenomenon of a self-light-emitting device in which a plurality of self-light-emitting elements are arranged in a matrix by causing a computer to execute the above.

Images