A color recognition method and device for a dynamic vision sensor

Abstract

The embodiment of the present invention relates to a color recognition method and device for a dynamic visual sensor, the method comprising: obtaining histogram data of first color, second color and third color events within a preset time; processing the histogram data to obtain a first color recognition cardinality, a second color recognition cardinality and a third color recognition cardinality; and recognizing the color according to the first color recognition cardinality, the second color recognition cardinality and the third color recognition cardinality. The technical solution of the embodiment of the present invention adopts dynamic visual sensor technology, superimposes a color filter layer, and effectively generates event information related to color; and uses the histogram data of the generated event information for processing, so as to effectively recognize the color information in fast-moving scenes.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of data image processing, and in particular to a color recognition method and device for a dynamic vision sensor. Background Art

[0002] Existing dynamic vision sensors do not have or only have microlenses on the pixel surface. They absorb light of the entire band and therefore do not differentiate between the light intensity of any band. They are all processed as grayscale or intensity values. Therefore, they are unable to distinguish and identify application scenarios in which light intensity changes occur due to specific bands or colors.

[0003] Existing event cameras can obtain dynamically changing information, and generate and read asynchronous event information as long as the signal generated by the change in light intensity in the environment exceeds or falls below a certain threshold. Information generation depends on ambient light and light intensity changes, so it is very suitable for machine vision applications, but when it comes to color recognition, especially when it comes to single-camera applications, current technology cannot achieve it. In autonomous driving applications, traditional image sensors cannot identify red and green well due to issues with exposure mode, readout speed and dynamic range; and traditional dynamic vision sensors cannot recognize colors, which is greatly limited in many specific applications. Summary of the invention

[0004] Based on the above situation of the prior art, the purpose of the embodiments of the present invention is to provide a color recognition method and device for a dynamic vision sensor, which adopts dynamic vision sensor technology, superimposes a color filter layer, effectively generates event information related to color, and further uses the histogram data of the event information for efficient color recognition, thereby solving the problem of color recognition needs in fast-moving scenes.

[0005] To achieve the above object, according to one aspect of the present invention, a color recognition method of a dynamic vision sensor is provided, comprising:

[0006] Obtain histogram data of the first color, the second color, and the third color events within a preset time, wherein the data includes the maximum value N of the histogram of the first color, the second color, and the third color events Gmax 、N Ymax and N Rmax ;

[0007] Process the histogram data to obtain a first color recognition base N′ G , the second color recognition cardinality N′ Y and the third color recognition cardinality N′ R :

[0008]

[0009] Among them, α G, α Y , α R are the first color response coefficient, the second color response coefficient and the third color response coefficient respectively; θ G ,θ Y ,θ R are respectively a first color threshold, a second color threshold and a third color threshold;

[0010] The color is identified according to the first color identification base, the second color identification base and the third color identification base.

[0011] Furthermore, it also includes:

[0012] When N′ G >σN′ Y And N′ G >σN′ R When the first color is identified as the main response color, a first color interrupt signal is output;

[0013] When N′ Y >σN′ G And N′ Y >σN′ R When the second color is identified as the main response color, a second color interrupt signal is output;

[0014] When N′ R >σN′ Y And N′ R >σN′ G When the third color is identified as the main response color, a third color interrupt signal is output;

[0015] Wherein, σ is the first judgment threshold.

[0016] Furthermore, the first color is green, the second color is yellow, and the third color is red; and further comprising:

[0017] When N′ G / N′ Y , N′ G / N ′ R or N′ R / N′ Y When the second judgment threshold is met, white is identified as the main response color.

[0018] Furthermore, it also includes:

[0019] Disposing filters of a first color, a second color, and a third color in a pixel circuit of the dynamic vision sensor;

[0020] The histogram data is obtained according to the photometric signal output by the filter.

[0021] Furthermore, the photometric signal output by the filter is obtained according to the following steps:

[0022] R′=R G Or R′=R Y ;or,

[0023] R′=R G |R Y , R G and R Y When any of the events is valid, R′ is valid; or,

[0024] R′=R G +R Y ;

[0025] Among them, R G is the photometric signal output by the first color filter; R Y is the photometric signal output by the second color filter; R′ is the processed signal of the photometric signals output by the two third color filters.

[0026] Further, filters of the first color, the second color, and the third color are arranged in the pixel circuit of the dynamic vision sensor according to the following layout:

[0027] A first color filter row and a second color filter row are alternately arranged in each row of pixel circuits, and a third color filter is alternately arranged in the filters of the first color filter row and the second color filter row respectively; or,

[0028] The first and second color filter rows and the third color filter row are arranged alternately in each row of the pixel circuit. In the first and second color filter rows, the first color filter and the second color filter are arranged alternately.

[0029] Furthermore, it also includes:

[0030] The recognition method is triggered according to the photometric signal output by the filter;

[0031] The bandpass filtering method is used to filter out high-frequency and ultra-low-frequency photometric signals.

[0032] According to a second aspect of the present invention, a color recognition device for a dynamic vision sensor is provided, comprising a histogram data acquisition module, a color recognition cardinality calculation module, and a color recognition module; wherein,

[0033] The histogram data acquisition module is used to acquire the histogram data of the first color, the second color and the third color events within a preset time, and the data includes the maximum value N of the histogram of the first color, the second color and the third color events. Gmax 、N Ymaxand N Rmax ;

[0034] The color recognition cardinality calculation module is used to process the histogram data to obtain a first color recognition cardinality N′ G , the second color recognition cardinality N′ Y and the third color recognition cardinality N′ R :

[0035]

[0036] Among them, α G , α Y , α R are the first color response coefficient, the second color response coefficient and the third color response coefficient respectively; θ G ,θ Y ,θ R are respectively a first color threshold, a second color threshold and a third color threshold;

[0037] The color recognition module is used to recognize colors according to the first color recognition base, the second color recognition base and the third color recognition base.

[0038] Furthermore, the color recognition module recognizes the color, including:

[0039] When N′ G >σN′ Y And N′ G >σN′ R When the first color is identified as the main response color, a first color interrupt signal is output;

[0040] When N′ Y >σN′ G And N′ Y >σN′ R When the second color is identified as the main response color, a second color interrupt signal is output;

[0041] When N′ R >σN′ Y And N′ R >σN′ G When the third color is identified as the main response color, a third color interrupt signal is output;

[0042] Wherein, σ is the first judgment threshold.

[0043] Furthermore, the color recognition module recognizes the color and further includes:

[0044] When N′ G / N′ Y , N′ G / N′ R or N′ R / N′ Y When the second judgment threshold is met, white is identified as the main response color.

[0045] In summary, the embodiment of the present invention provides a color recognition method and device for a dynamic visual sensor, the method comprising: obtaining histogram data of a first color, a second color, and a third color event within a preset time; processing the histogram data to obtain a first color recognition cardinality, a second color recognition cardinality, and a third color recognition cardinality; and identifying the color according to the first color recognition cardinality, the second color recognition cardinality, and the third color recognition cardinality. The technical solution of the embodiment of the present invention adopts dynamic visual sensor technology, superimposes a color filter layer, and effectively generates event information related to color; the histogram data of the generated event information is used for processing, so that the color event information in a fast-moving scene can be effectively identified, and the following beneficial technical effects can be achieved: the color event information in a fast-moving scene can be quickly identified according to the color-related event information, which can effectively assist the application of RGB sensors, so that the event information generated by different colors can be quickly distinguished, and effective scene and background information can be provided for identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 is a schematic diagram of the dynamic vision sensor system architecture;

[0047] Figure 2 It is a schematic diagram of the pixel array circuit of the dynamic vision sensor;

[0048] Figure 3 It is a schematic diagram of the output mechanism of loudness and brightness of dynamic vision sensor;

[0049] Figure 4 is a flow chart of a color recognition method of a dynamic vision sensor provided by an embodiment of the present invention;

[0050] Figure 5 It is a schematic diagram of the arrangement of color filters on the photodiode of the dynamic vision sensor;

[0051] Figure 6 is a schematic diagram of a layout of a color filter on a pixel array according to an embodiment of the present invention;

[0052] Figure 7 is another schematic diagram of the layout of the color filter on the pixel array according to an embodiment of the present invention;

[0053] Figure 8 is another schematic diagram of the layout of the color filter on the pixel array according to an embodiment of the present invention;

[0054] Fig. 9is another schematic diagram of the layout of the color filter on the pixel array according to an embodiment of the present invention;

[0055] Fig.10 is another schematic diagram of the layout of the color filter on the pixel array according to an embodiment of the present invention;

[0056] Fig.11 FIG. 4 is another schematic diagram of the layout of the color filter on the pixel array according to an embodiment of the present invention. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings. It should be understood that these descriptions are only exemplary and are not intended to limit the scope of the present invention. In addition, in the following description, the description of well-known structures and technologies is omitted to avoid unnecessary confusion of the concept of the present invention.

[0058] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in one or more embodiments of the present invention should be understood by people with ordinary skills in the field to which the present disclosure belongs. The "first", "second" and similar words used in one or more embodiments of the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0059] The following is a detailed description of the technical solution of the embodiment of the present invention in conjunction with the accompanying drawings. The system architecture of the dynamic vision sensor is as follows: Figure 1 As shown, it includes peripheral processing circuits such as pixel array and event information processing circuit. After the pixel array generates asynchronous event information, it is read out at the fastest speed through the peripheral synchronization circuit. With the assistance of the overall system timing control and storage module, it is output through the adaptive interface after certain algorithms and signal processing. Each pixel in the pixel array responds independently to changes in light intensity. The schematic diagram of the pixel array circuit is shown in Figure 2As shown, it includes a signal acquisition circuit 201, a signal amplification circuit 202 and a comparison circuit 203. Among them, the signal acquisition circuit 201 is used to collect the intensity of the incident light and generate a corresponding logarithmic voltage signal; the signal amplification circuit 202 is used to amplify the voltage signal; the comparison circuit 203 is used to compare the amplified voltage signal with the threshold range. If it exceeds the preset threshold range (higher or lower), it is considered that the pixel is effectively excited, and its brightness change data is valid data. The pixel brightness value represents its excited binary data is read out. Figure 3 The output mechanism diagram of loudness brightness is shown in FIG. 1 , where the light intensity change degree preset by the threshold range represents the condition for generating new effective brightness data. If the pixel brightness change does not exceed this range, it is considered that the pixel brightness has not changed and does not need to be output. When the pixel brightness change exceeds this range, it means that new data has appeared in the pixel at this time and needs to be read out.

[0060] In order to realize color recognition based on the above-mentioned dynamic vision sensor, in an embodiment of the present invention, a color recognition method 400 of a dynamic vision sensor is provided. The flowchart of the method 400 is as follows: Figure 4 As shown, the following steps are included:

[0061] S401, obtaining histogram data of the first color, the second color and the third color events within a preset time, wherein the data includes the maximum value N of the histogram of the first color, the second color and the third color events. Gmax 、N Ymax and N Rmax The histogram data can be obtained by setting filters of the first color, the second color, and the third color in the pixel circuit of the dynamic vision sensor, and generating event information according to the photometric change signal output by the filters. The horizontal axis of the histogram is the first color, the second color, and the third color, and the vertical axis is the photometric change value of the filters of the first color, the second color, and the third color.

[0062] In the embodiment of the present invention, the color filters are arranged in a certain manner, such as Figure 5As shown, the filter is set on the photodiode in the pixel circuit of the dynamic vision sensor and then the microlens is used. The microlens can make the quantum efficiency higher. More importantly, the filter will transmit the light of a specific wavelength through the photodiode device for photoelectric conversion. The light intensity changes after the current filter, such as red, green, blue and other colors, are realized. The color filters with different layouts arranged on the photodiode can be realized in the pixel array, and the four different color modules are generated according to the regular layout. There are many ways to layout. For example, the first color filter row and the second color filter row are arranged alternately in each row of the pixel circuit, and the third color filter is arranged alternately in the filters of the first color filter row and the second color filter row; or, the first and second color filter rows and the third color filter row are arranged alternately in each row of the pixel circuit, and the first color filter and the second color filter row are arranged alternately in the first and second color filter rows. Figure 6-Figure 11 The possible layout is shown in FIG. 1 . In the above figures, G represents a green filter, Y represents a yellow filter, R represents a red filter, and R represents a red filter. G and R Y Represents the red filters of the G and Y rows respectively. Figure 6-Figure 9 In the figure, green filter rows and yellow filter rows are arranged alternately in each row of the pixel circuit, and red filters are arranged alternately in the middle of the green filters and the middle of the yellow filters in the green filter row and the yellow filter row, respectively. From the column direction, the red filters of each row can be arranged in the same column or alternately. Fig.10 and Fig.11 In each row of the pixel circuit, yellow, green and red filters are alternately arranged, wherein the number of red filters is twice the number of yellow and green filters, or in each row, green and yellow filter rows and red filter rows including only red filters are alternately arranged. The above filters can also be selected as filters of other colors, and can be set according to the requirements of color recognition in actual application scenarios. The photometric signal output by the above filters is obtained according to the following steps:

[0063] R′=R G Or R′=R Y ;or,

[0064] R′=R G |R Y , R G and R Y When any of the events is valid, R′ is valid; or,

[0065] R′=R G +R Y ;

[0066] Among them, R Gis the photometric signal output by the first color filter; R Y is the photometric signal output by the second color filter; R′ is the processed signal of the photometric signals output by the two third color filters.

[0067] When the event information is obtained through the photometric signal, the color recognition method involved in the embodiment of the present invention is triggered, and different areas are processed independently to fully realize color and shape recognition. There will be some differences in different or same layout processing. For example, the implementation of two-choice is simple, but some information will be lost. Combining the two into one at the front-end PD can improve sensitivity but also lose resolution.

[0068] According to some embodiments, after acquiring event information, a bandpass filtering method is first used to filter out high-frequency and ultra-low-frequency event signals, and useful events or events that need to be identified are extracted through target matching.

[0069] S402, processing the histogram data, such as performing coefficient correction and normalization processing, to provide accurate data for final judgment, so as to obtain a first color recognition base N′ G , the second color recognition cardinality N′ Y and the third color recognition cardinality N′ R :

[0070]

[0071] Among them, α G , α Y , α R are the first color response coefficient, the second color response coefficient and the third color response coefficient respectively; θ G ,θ Y ,θ R They are respectively the first color threshold, the second color threshold and the third color threshold. The corresponding coefficients and color thresholds of the above colors can be obtained through preliminary tests. This step processes the histogram data to realize color recognition. Each color is grouped into a module. According to the event information generated by each module, if the events generated between the modules are quite different, and the event volume of a certain module is much larger than that of the other modules, it can be judged that the event is generated by a certain color of red, green or yellow. Specifically, the histogram statistics of events in different modules can be used to determine the characteristic response of a certain color after the event volume exceeds a certain proportion. The coefficients of the histogram data are corrected and normalized through the above formula to provide accurate data for the final judgment.

[0072] S403, identifying the color according to the first color identification cardinality, the second color identification cardinality and the third color identification cardinality, wherein the first color is, for example, green, the second color is, for example, yellow, and the third color is, for example, red. The identification step is specifically as follows:

[0073] When N′ G >σN′ Y And N′ G >σN′ R When the first color is identified as the main response color, a first color interrupt signal is output;

[0074] When N′ Y >σN′ G And N′ Y >σN′ R When the second color is identified as the main response color, a second color interrupt signal is output;

[0075] When N′ R >σN′ Y And N′ R >σN′ G When the third color is identified as the main response color, a third color interrupt signal is output;

[0076] When N′ G / N′ Y , N′ G / N′ R or N′ R / N′ Y When the second judgment threshold is met, white is identified as the main response color;

[0077] When none of the above conditions are met, it is judged as street view mode and other application processing is performed.

[0078] Wherein, σ is the first judgment threshold, which can be set according to the quantum efficiency of different colors, and in this embodiment, σ = 1.567 can be taken. The second judgment threshold is a data range, which can be set according to the average standard deviation, and in this embodiment, (0.967, 1.033) can be taken.

[0079] According to some embodiments, after color recognition is completed, morphological recognition can be performed on modules with a large number of events based on the color recognition results, such as the circle and arrow of a traffic light. If it meets the requirements, it is determined to be a traffic identification light of the corresponding color.

[0080] An embodiment of the present invention further provides a color recognition device for a dynamic vision sensor, the device comprising a histogram data acquisition module, a color recognition cardinality calculation module, and a color recognition module.

[0081] A histogram data acquisition module is used to acquire the histogram data of the first color, the second color and the third color events within a preset time, wherein the data includes the maximum value N of the histogram of the first color, the second color and the third color events. Gmax 、N Ymax and N Rmax ;

[0082] A color recognition cardinality calculation module is used to process the histogram data to obtain a first color recognition cardinality N′ G , the second color recognition cardinality N′ Y and the third color recognition cardinality N′ R :

[0083]

[0084] Among them, α G , α Y , α R are the first color response coefficient, the second color response coefficient and the third color response coefficient respectively; θ G ,θ Y ,θ R are respectively a first color threshold, a second color threshold and a third color threshold;

[0085] The color recognition module is used to recognize the color according to the first color recognition base, the second color recognition base and the third color recognition base. The color recognition module recognizes the color, including:

[0086] When N′ G >σN′ Y And N′ G >σN′ R When the first color is identified as the main response color, a first color interrupt signal is output;

[0087] When N′ Y >σN′ G And N′ Y >σN′ R When the second color is identified as the main response color, a second color interrupt signal is output;

[0088] When N′ R >σN′ Y And N′ R >σN′ G When the third color is identified as the main response color, a third color interrupt signal is output;

[0089] When N′ G / N′ Y , N′ G / N′ R or N′ R / N′ Y When the second judgment threshold is met, white is identified as the main response color.

[0090] Wherein, σ is the first judgment threshold.

[0091] The principles and specific processes of color recognition performed by each module in the color recognition device of the dynamic vision sensor in this embodiment of the present invention are the same as the color recognition method of the dynamic vision sensor involved in the above-mentioned embodiment, and will not be described in detail here.

[0092] In summary, the embodiments of the present invention relate to a color recognition method and device for a dynamic vision sensor, the method comprising: obtaining histogram data of first color, second color and third color events within a preset time; processing the histogram data to obtain a first color recognition cardinality, a second color recognition cardinality and a third color recognition cardinality; and recognizing colors according to the first color recognition cardinality, the second color recognition cardinality and the third color recognition cardinality. The technical solution of the embodiments of the present invention adopts dynamic vision sensor technology, superimposes color filters, and effectively generates event information related to colors; and uses the histogram data of the generated event information for processing, so as to effectively recognize color information in fast-moving scenes.

[0093] It should be understood that the discussion of any of the above embodiments is only exemplary and is not intended to imply that the scope of the present invention (including claims) is limited to these examples; under the idea of ​​the present invention, the technical features in the above embodiments or different embodiments may also be combined, the steps may be implemented in any order, and there are many other changes in different aspects of one or more embodiments of the present invention as described above, which are not provided in detail for the sake of simplicity. The above specific embodiments of the present invention are only used to illustrate or explain the principles of the present invention, and do not constitute a limitation of the present invention. Therefore, any modifications, equivalent substitutions, improvements, etc. made without departing from the spirit and scope of the present invention should be included in the scope of protection of the present invention. In addition, the claims attached to the present invention are intended to cover all changes and modifications that fall within the scope and boundaries of the attached claims, or the equivalent forms of such scope and boundaries.

Claims

1. A color recognition method for a dynamic visual sensor, characterized in that: include: Obtain histogram data of the first color, the second color, and the third color events within a preset time, wherein the data includes the maximum value N of the histogram of the first color, the second color, and the third color events Gmax 、N Ymax and N Rmax ; wherein filters of the first color, the second color and the third color are arranged in the pixel circuit of the dynamic vision sensor; the histogram data is obtained according to the event information generated by the photometric change signal output by the filter; the abscissa of the histogram is the first color, the second color and the third color, and the ordinate is the photometric change value of the filters of the first color, the second color and the third color; Process the histogram data to obtain a first color recognition base N′ G , the second color recognition cardinality N′ Y and the third color recognition cardinality N′ R : Among them, α G , α Y , α R are the first color response coefficient, the second color response coefficient and the third color response coefficient respectively; θ G ,θ Y ,θ G are respectively a first color threshold, a second color threshold and a third color threshold; The color is identified according to the first color identification base, the second color identification base and the third color identification base.

2. The method according to claim 1, characterized in that Also includes: When N′ G >σN′ Y And N′ G >σN′ R When the first color is identified as the main response color, a first color interrupt signal is output; When N′ Y >σN′ G And N′ Y >σN′ R When the second color is identified as the main response color, a second color interrupt signal is output; When N′ R >σN′ Y And N′ R >σN′ G When the third color is identified as the main response color, a third color interrupt signal is output; Wherein, σ is the first judgment threshold.

3. The method according to claim 2, characterized in that The first color is green, the second color is yellow, and the third color is red; and further comprising: When N′ G / N′ Y , N′ G / N′ R or N′ R / N′ Y When the second judgment threshold is met, white is identified as the main response color.

4. The method according to claim 3, characterized in that: The event information generated by the photometric change signal output by the filter is obtained according to the following steps: R′=R G Or R′=R Y ;or, R′=R G |R Y , R G and R Y When any of the events is valid, R′ is valid; or, R′=R G +R Y ; Among them, R G is the light intensity change signal value output by the first color filter; R Y is the light intensity change signal value output by the second color filter; R ′ It is a signal value after the event information generated by the photometric change signal values ​​output by the two third color filters is processed.

5. The method according to claim 4, characterized in that Filters of the first color, the second color, and the third color are arranged in the pixel circuit of the dynamic vision sensor according to the following layout: A first color filter row and a second color filter row are alternately arranged in each row of pixel circuits, and a third color filter is alternately arranged in the filters of the first color filter row and the second color filter row respectively; or, The first and second color filter rows and the third color filter row are arranged alternately in each row of the pixel circuit. In the first and second color filter rows, the first color filter and the second color filter are arranged alternately.

6. The method according to claim 4 or 5, characterized in that: Also includes: The recognition method is triggered according to the photometric signal output by the filter; The bandpass filtering method is used to filter out the event information generated by high-frequency and ultra-low-frequency luminosity change signals.

7. A color recognition device for a dynamic visual sensor, characterized in that: It includes a histogram data acquisition module, a color recognition cardinality calculation module, and a color recognition module; wherein, The histogram data acquisition module is used to acquire the histogram data of the first color, the second color and the third color events within a preset time, and the data includes the maximum value N of the histogram of the first color, the second color and the third color events. Gmax 、N Ymax and N Rmax ; wherein filters of the first color, the second color and the third color are arranged in the pixel circuit of the dynamic vision sensor; the histogram data is obtained according to the event information generated by the photometric change signal output by the filter; the abscissa of the histogram is the first color, the second color and the third color, and the ordinate is the photometric change value of the filters of the first color, the second color and the third color; The color recognition cardinality calculation module is used to process the histogram data to obtain a first color recognition cardinality N′ G , the second color recognition cardinality N′ Y and the third color recognition cardinality N′ R : Among them, α G , α Y , α R are the first color response coefficient, the second color response coefficient and the third color response coefficient respectively; θ G ,θ Y ,θ R are respectively a first color threshold, a second color threshold and a third color threshold; The color recognition module is used to recognize colors according to the first color recognition base, the second color recognition base and the third color recognition base.

8. The device according to claim 7, characterized in that The color recognition module recognizes the color, including: When N′ G >σN′ Y And N′ G >σN′ R When the first color is identified as the main response color, a first color interrupt signal is output; When N′ Y >σN′ G And N′ Y >σN′ R When the second color is identified as the main response color, a second color interrupt signal is output; When N′ R >σN′ Y And N′ R >σN′ G When the third color is identified as the main response color, a third color interrupt signal is output; Wherein, σ is the first judgment threshold.

9. The device according to claim 8, characterized in that The color recognition module recognizes the color, and also includes: When N′ G / N′ Y , N′ G / N′ R or N′ R / N′ Y When the second judgment threshold is met, white is identified as the main response color.

Images