Method and system for perceptual and evaluation of compression artifacts in high efficiency video coding

Abstract

The application provides a compression effect perception and evaluation method and system in high-efficiency video coding, a deep compression effect perception model combining attention learning and data reweighting is constructed, positive and negative samples of each type of compression effect are input into the network, and the perception model of the corresponding compression effect is obtained through training; based on the compression effect perception model obtained through training, and considering the influence of visual saliency and visual fovea effect, the video compression effect intensity value is calculated; through subjective test, the average subjective score of the video is calculated to establish the mapping relationship between the video compression effect intensity value and the average subjective quality score of the video, and video quality evaluation oriented to compression effect perception is realized.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of video quality evaluation, and particularly relates to a compression effect perception and evaluation method and system in high-efficiency video coding. BACKGROUND

[0002] Videos are usually compressed to save storage and transmission bandwidth, and popular lossy video compression inevitably leads to video compression effects, also known as perceptible coding artifacts. Compression effects affect the subjective experience of users on videos and significantly reduce the video experience quality of users. However, there are many types of video compression effects and complex features, and it is difficult to construct a perception model by manually extracting features. Existing methods usually lack comprehensive consideration of spatial and temporal compression effects, have great defects in perception and integration of compression effect features, and cannot provide reliable data reference for compressed video quality evaluation, thereby having limitations in compressed video quality evaluation. Therefore, compression effect perception and analysis in high-efficiency video coding, exploration of more accurate perception methods, and analysis of compressed video quality are helpful to promote optimization and improvement of video coding compression technology. SUMMARY

[0003] Therefore, the application aims to provide a compression effect perception and analysis method and system based on high-efficiency video coding, wherein the method comprises the following steps: step S1: constructing a deep compression effect perception model, inputting positive and negative samples of each type of compression effect into the network, and respectively training to obtain a corresponding compression effect perception and recognition model; step S2: based on the compression effect perception model and considering the influence of visual saliency and visual fovea effect, calculating a video compression effect intensity value; step S3: calculating the average subjective score of the video through subjective testing, establishing a mapping relationship between the video compression effect intensity value and the average subjective quality score of the video, and finally realizing video quality evaluation oriented to compression effect perception according to the mapping relationship between the compression effect intensity measurement value and the average subjective quality score of the video; and the method can realize accurate compression effect perception to guide video quality evaluation.

[0004] The application solves the technical problems and specifically adopts the technical solutions of:

[0005] A compression effect perception and evaluation method in high-efficiency video coding comprises the following steps:

[0006] A deep compression effect perception model combining attention learning and data reweighting is constructed, positive and negative samples of each type of compression effect are input into the network, and a corresponding compression effect perception model is respectively trained;

[0007] Based on the trained compression effect perception model and considering the influence of visual saliency and visual fovea effect, a video compression effect intensity value is calculated;

[0008] The average subjective score of the video is calculated through subjective test to establish a mapping relationship between the video compression effect intensity value and the average subjective quality score of the video, so as to realize the video quality evaluation oriented to compression effect perception.

[0009] Further, when training the deep compression effect perception model, video frame samples including multiple compression effects are preprocessed as inputs, the sample data are input into a YOLO detection backbone network to extract compression effect features, the compression effect features are created into a feature pyramid through a Transformer encoding block, and then feature fusion is realized through a multi-scale structure, and data reweighting is performed to solve the long-tail distribution problem of the sample.

[0010] The trained network model is subjected to model verification to confirm whether the preset requirements are met, if yes, the model is saved as a compression effect detection model, and if not, the training is restarted until the preset requirements are met.

[0011] Further, the types of the compression effect samples include blur effect samples, blocking effect samples, ringing effect samples, color bleeding effect samples, flicker effect samples and floating effect samples.

[0012] Further, the compression effect intensity value of the video is calculated based on the trained compression effect perception model and considering the influence of visual saliency and visual fovea effect, and the calculation includes:

[0013] The total pixel value of the compression effect of the video is counted based on the perception model;

[0014] The saliency factor and the fovea factor are calculated considering the visual saliency and the visual fovea effect;

[0015] The weighted compression effect intensity value is recalculated.

[0016] Further, the total pixel value of the compression effect of the video counted based on the perception model is specifically:

[0017] The compressed video to be tested is input into the trained compression effect perception model, the area where the compression effect is located in each video frame is detected, and then the total pixel value I of the compression effect of each video frame of the video is counted. p :

[0018] wherein T represents the compression effect type, M represents the total number of video frame compression effect areas, w_box and h_box represent the width and height of the detection box respectively;

[0019] The saliency factor and the fovea factor calculated considering the visual saliency and the visual fovea effect are specifically:

[0020] By introducing a video saliency detection algorithm ACLNet to extract the region of interest in each video frame, the area of the saliency region in each video frame is counted, and a saliency detection factor I is obtained S ;

[0021] In combination with the detection frame height of the perception model, a visual fovea factor I is calculated E :

[0022]

[0023] wherein H represents the height of the video frame, and y' represents the height of the prediction frame of the perception model;

[0024] The calculation of the weighted compression effect intensity value is specifically as follows:

[0025] Based on the detection result of the compression effect perception model, considering the visual saliency and the visual fovea effect, a weighted compression effect intensity value I of the entire compressed video is calculated:

[0026]

[0027] wherein x and y represent the center horizontal coordinate and the vertical coordinate of the detection frame of the perception model, I P represents the total pixel value of the compression effect, I S represents the saliency detection factor, I E represents the visual fovea factor, W and H represent the width and height of the video frame respectively, and N represents the total frame number of the compressed video.

[0028] Further, the subjective test adopts a double stimulus method, and after the subjective test is completed, the MOS values of the experimenters are screened; the collected subjective scores are processed, and the MOS value of each test video is calculated according to the following formula: wherein S i represents the score of each tester for the same video; N represents the number of testers; then the correlation coefficient between the score of each tester and MOS is calculated to ensure the effectiveness of the obtained data; finally, according to ITU-R BT.1788, the test data is screened.

[0029] Further, the correlation coefficient includes PLCC and SROCC;

[0030] Firstly, the PLCC and SRCC values of all testers are inputted; and the mean values r1 and r2 of PLCC and SRCC are calculated;

[0031] Then, it is judged whether r1 is greater than r2, if yes, r2 is taken, the standard deviation s2 of SRCC is calculated, and if no, r1 is taken, the standard deviation s1 of PLCC is calculated;

[0032] Then, the difference q between r and s is calculated, and it is determined whether q is greater than the maximum threshold:

[0033] If yes, the discard threshold is set to the maximum threshold, and it is determined whether PLCC(i) / SRCC(i) is greater than the maximum threshold, if yes, the rejection result is output, and if no, the ith test person is rejected;

[0034] If no, the discard threshold is set to q, and it is determined whether PLCC(i) / SRCC(i) is greater than q, if yes, the rejection result is output, and if no, the ith test person is rejected.

[0035] The mapping relationship between the compression effect intensity value and the average subjective quality score of the video is specifically a mapping relationship between the compression effect intensity value I and the average subjective opinion score after the screening processing.

[0036] A compression effect perception and evaluation system in high-efficiency video coding, comprising:

[0037] A training module is configured to train a deep compression effect perception model combined with attention learning and data reweighting, input positive and negative samples of each type of compression effect into the network, and train to obtain a corresponding compression effect perception model for each type of compression effect;

[0038] A video compression effect intensity value calculation module is configured to calculate a video compression effect intensity value based on the trained compression effect perception model and taking into account the influence of visual saliency and visual fovea effect;

[0039] A subjective test module is configured to calculate an average subjective score of a video through subjective test,

[0040] An evaluation module is configured to establish a mapping relationship between the video compression effect intensity value and the average subjective quality score of the video, and realize video quality evaluation oriented to compression effect perception.

[0041] An electronic device comprises a memory, a processor, and a computer program stored on the memory and executable on the processor, and the processor implements the steps of the compression effect perception and evaluation method in high-efficiency video coding as described above when executing the program.

[0042] A non-transitory computer readable storage medium has a computer program stored thereon, and the computer program is executed by a processor to implement the steps of the compression effect perception and evaluation method in high-efficiency video coding as described above.

[0043] Compared with the prior art, the application and the preferred schemes thereof realize the video quality evaluation oriented to compression effect perception by establishing the mapping relationship between the video compression effect intensity value and the video average subjective quality score, and realizing the video quality evaluation oriented to compression effect perception according to the mapping relationship between the compression effect intensity measurement value and the video average subjective quality score, which can realize accurate compression effect perception to guide the video quality evaluation. BRIEF DESCRIPTION OF DRAWINGS

[0044] The application will be described in further detail below in combination with the accompanying drawings and specific embodiments:

[0045] Figure 1 is the overall implementation flowchart of the embodiment of the application;

[0046] Figure 2 is the video quality evaluation model diagram guided by compression effect perception and analysis in the embodiment of the application;

[0047] Figure 3 is the subjective data screening flowchart in the embodiment of the application. DETAILED DESCRIPTION

[0048] To make the features and advantages of the application more obvious and easy to understand, the following embodiments are specifically described as follows:

[0049] It should be noted that the following detailed description is all exemplary and is intended to provide further illustration of the present application. Unless otherwise specified, all technical and scientific terms used in the specification have the same meaning as generally understood by those skilled in the art to which the present application belongs.

[0050] It should be noted that the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, the singular form is intended to include the plural form unless the context clearly indicates otherwise, and it should also be understood that when the terms "comprise" and / or "include" are used in the specification, there is a presence of the features, steps, operations, devices, components and / or combinations thereof.

[0051] As shown in Figure 1 , the embodiment of the application provides a design of compression effect perception and analysis method in high-efficiency video coding, which comprises the following steps:

[0052] Step S1: Construct a deep compression effect perception model, input positive and negative samples of each type of compression effect into the network, and respectively train to obtain the corresponding compression effect perception and recognition model;

[0053] Step S2: Based on the compression effect perception model, and considering the influence of visual saliency and visual fovea effect, calculate the video compression effect intensity value;

[0054] Step S3: Calculate the average subjective score of the video through subjective test, and establish the mapping relationship between the video compression effect intensity value and the video average subjective quality score;

[0055] Finally, according to the mapping relationship between the compression effect intensity measurement value and the video average subjective quality score, the video quality evaluation oriented to compression effect perception is realized.

[0056] In this embodiment, as shown in Figure 2 The deep compression effect perception model is trained by taking 6 compression effect samples as input. The sample data is input into the backbone network to extract the basic features of the compression effect. The compression effect features create a feature pyramid through the Transformer encoding block, and then realize feature fusion through the multi-scale structure. Finally, the sample long-tail distribution problem is solved by data reweighting. The accurate detection of compression effect is completed to guide the compression video quality evaluation. Specifically:

[0057] Step S11: Use the samples in the large-scale compression effect database as the data input for training the network;

[0058] Step S12: Take YOLO detection network as the basic framework, use attention learning and data reweighting to deal with the complexity and diversity of compression effect and the long-tail distribution problem of samples;

[0059] Step S13: Model verification is performed on the trained network model to confirm whether the preset requirements are met. If the requirements are met, the model is saved as a compression effect detection model. If the requirements are not met, retraining is performed until the preset requirements are met.

[0060] In this embodiment, the 6 types of compression effect samples include blur effect samples, block effect samples, ringing effect samples, color overflow effect samples, flicker effect samples and floating effect samples.

[0061] In one specific example, the model training is as follows: In the process of training the detection model, for each type of perceptible coding effect, 6 types of compression effect samples are selected from a large-scale database, a total of 37130 samples, of which 80% are used as the training set and 20% are used as the test set. The ratio of positive and negative samples is 1:1. In the process of training the model, the SGD optimizer is used, the epoch is set to 300 batches during training, the input image size is set to 640x640, the patch is set to 8, the learning rate is set to 0.01, and the IoU threshold is set to 0.5. In the process of model training, the adjustment of network hyperparameters, the fine-tuning of network layer structure, the selection of network optimizer and other work can improve the accuracy and performance of the network.

[0062] In this embodiment, step S2 is specifically:

[0063] Step S21: taking the compressed video to be tested as the input of the perception model, detecting the area of compression effect in each video frame, and then counting the total pixel value I of the compression effect of each video frame of the video p :

[0064] wherein T represents the compression effect type, M represents the total number of video frame compression effect areas, w_box and h_box represent the width and height of the detection box respectively;

[0065] Step S22: considering that people usually pay more attention to the area of interest, which is an important factor affecting the perceptual quality of the video. By introducing the video saliency detection algorithm ACLNet, the area of interest in each video frame is extracted, and the area of saliency in each video frame, i.e., the saliency detection factor I is counted S ;

[0066] Step S23: considering the foveal effect formed by the difference in visual acuity between the central vision and the peripheral vision of the human eye, and combining the detection box height of the perception model, the visual foveal factor I is calculated E :

[0067]

[0068] wherein H represents the height of the video frame, and y' represents the predicted box height of the perception model;

[0069] Step S24: based on the detection result of the compression effect perception model, considering the visual saliency and the visual foveal effect, the weighted compression effect intensity value I of the entire compressed video is calculated:

[0070]

[0071] wherein x and y represent the center horizontal coordinate and vertical coordinate of the detection box of the perception model respectively, I P represents the total pixel value of the compression effect, I S represents the saliency detection factor, I E represents the visual foveal factor, W and H represent the width and height of the video frame respectively, and N represents the total number of frames of the compressed video.

[0072] In this embodiment, step S3 is specifically:

[0073] Step S31: the double stimulation method in the subjective test method is adopted, i.e., a reference video and a test video are provided for the test personnel each time. The reference video and the test video have a duration of 10 seconds, and 1 second of black screen is set between the videos. After the test video is played, 3 seconds of scoring time is set, and the test personnel are asked to score the video subjectively. The experimental environment is set and calibrated according to the recommendations of ITU-R BT.500.

[0074] Step S32: After the subjective experiment is completed, the MOS values of the experimenters are screened. The collected subjective scores are processed, and the MOS value of each test video is calculated according to the following formula: Wherein S i represents the score of each tester for the same video; N represents the number of testers;

[0075] Step S33: The correlation coefficient between the score of each tester and the MOS is calculated to ensure the effectiveness of the obtained data.

[0076] Step S34: According to the calculation result of step S33, the test data is screened according to ITU-R BT.1788.

[0077] Step S35: The compression effect intensity value I obtained in step S3 is mapped with the average subjective opinion score after the screening processing in step S34.

[0078] As a preferred solution, in the embodiment, the correlation coefficient includes PLCC and SROCC.

[0079] Wherein, PLCC (Pearson Linear Correlation Coefficient) and SROCC (Spearman Rank Correlation Coefficient) are both indexes for measuring the correlation between subjective scores and objective scores.

[0080] PLCC is used to describe the linear correlation between the two groups of quality evaluation data. Its value range is (-1, 1), and the value 0 indicates that the two groups of data are completely irrelevant, and the values 1 or -1 indicate that the two groups of data are completely relevant.

[0081] SROCC is used to describe the consistency of subjective and objective evaluation results, and its value range is (0, 1), and the value 0 indicates that the subjective and objective evaluation results are completely inconsistent, and the value 1 indicates that they are completely consistent. SROCC is not a measure of the linear correlation of the two groups of data, but a measure of the correlation between the order of data.

[0082] As Figure 3 shown, step S33 specifically includes the following implementation process:

[0083] First, input the PLCC and SRCC values of all testers; and calculate the mean values r1 and r2 of PLCC and SRCC;

[0084] Then, determine whether r1 is greater than r2, if yes, take r2, calculate the standard deviation s2 of SRCC, if no, take r1, calculate the standard deviation s1 of PLCC;

[0085] Then, calculate the difference q of r and s, and determine whether q is greater than the maximum threshold:

[0086] If yes, then let the discard threshold = the maximum threshold, and then determine whether PLCC(i) / SRCC(i) is greater than the maximum threshold. If yes, then output the rejection result, and if no, then reject the ith test subject;

[0087] If no, then let the discard threshold = q, and then determine whether PLCC(i) / SRCC(i) is greater than q. If yes, then output the rejection result, and if no, then reject the ith test subject.

[0088] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code.

[0089] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and a combination of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus generate an apparatus that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in the flowcharts and / or block diagrams Figure 1 The functions specified in the flowcharts and / or block diagrams

[0090] These computer program instructions can also be stored in a computer-readable memory that can direct the computer or other programmable data processing apparatus to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including instruction apparatus, which implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in the flowcharts and / or block diagrams Figure 1 The functions specified in the flowcharts and / or block diagrams

[0091] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable data processing apparatus to produce a computer-implemented process, so that the instructions executed on the computer or other programmable data processing apparatus provide a process for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in the flowcharts and / or block diagrams Figure 1 The functions specified in the flowcharts and / or block diagrams

[0092] It should be noted that the technical terms or scientific terms used in the present application should be understood as the common meanings understood by those skilled in the art unless otherwise defined. The terms "first", "second", and the like used in the present application do not represent any order, number, or importance, but are only used to distinguish different components. The terms "include", "contain", and the like mean that the elements or objects before the terms cover the elements or objects listed after the terms and their equivalents, and do not exclude other elements or objects. The terms "connect" or "connected" and the like are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. The terms "upper", "lower", "left", "right", and the like are only used to represent relative positional relationships, and when the absolute positions of the described objects change, the relative positional relationships can also change accordingly.

[0093] The above is only the preferred embodiment of the present application, and does not limit the other forms of the present application. Any skilled person in the art can use the disclosed technical content to make changes or modifications to equivalent embodiments. However, any simple modification, equivalent change, and modification of the above embodiments made without departing from the technical solution of the present application, in accordance with the technical essence of the present application, still falls within the protection scope of the present application.

[0094] The present application is not limited to the above-mentioned best mode, and anyone can derive other various forms of compression effect perception and evaluation methods and systems in high-efficiency video coding under the inspiration of the present application. Any equivalent changes and modifications made in accordance with the scope of the present application should be within the scope of the present application.

Claims

1. A method for compression artifact perception and evaluation in high-efficiency video coding, characterized in that: a deep compression artifact perception model combining attention learning and data reweighting is constructed, positive and negative samples of each type of compression artifact are input into the network, and the corresponding compression artifact perception model is trained respectively; based on the trained compression artifact perception model, the influence of visual saliency and visual fovea effect is considered, and the video compression artifact intensity value is calculated; through subjective testing, the average subjective score of the video is calculated to establish the mapping relationship between the video compression artifact intensity value and the average subjective quality score, and the video quality evaluation oriented to compression artifact perception is realized; when training the deep compression artifact perception model, video frame samples including multiple compression artifacts are preprocessed as input, the sample data is input into the YOLO detection backbone network to extract compression artifact features, the compression artifact features are created into a feature pyramid through a Transformer encoding block, and then feature fusion is realized through a multi-scale structure, and data reweighting is performed to solve the long-tail distribution problem of the sample; the calculation of the video compression artifact intensity value based on the trained compression artifact perception model and considering the influence of visual saliency and visual fovea effect includes: calculating the total pixel value of the compression artifact of the video based on the perception model; calculating the saliency factor and the fovea factor considering the visual saliency and the visual fovea effect; and calculating the weighted compression artifact intensity value; the calculation of the total pixel value of the compression artifact of the video based on the perception model specifically includes: calculating the total pixel value of the compression artifact of the video based on the perception model; the calculation of the saliency factor and the fovea factor considering the visual saliency and the visual fovea effect specifically includes: calculating the saliency factor and the fovea factor considering the visual saliency and the visual fovea effect; and the calculation of the weighted compression artifact intensity value specifically includes: calculating the weighted compression artifact intensity value; based on the detection result of the compression artifact perception model, considering the visual saliency and the visual fovea effect, the weighted compression artifact intensity value I of the entire compressed video is calculated; the types of the compression artifact samples include blur effect samples, block effect samples, ringing effect samples, color overflow effect samples, flicker effect samples, and floating effect samples.

3. The method for compression artifact perception and evaluation in high-efficiency video coding according to claim 1, characterized in that: the correlation coefficient includes PLCC and SROCC; first, input the PLCC and SRCC values of all test personnel; and calculate the mean values r1 and r2 of PLCC and SRCC; then, determine whether r1 is greater than r2, if yes, take r2, calculate the standard deviation s2 of SRCC, and if no, take r1, calculate the standard deviation s1 of PLCC; next, calculate the difference q of r and s, and determine whether q is greater than the maximum threshold: if yes, let the discard threshold = the maximum threshold, and then determine whether PLCC(i) / SRCC(i) is greater than the maximum threshold, if yes, output the rejection result, and if no, reject the ith test personnel; if no, let the discard threshold = q, and then determine whether PLCC(i) / SRCC(i) is greater than q, if yes, output the rejection result, and if no, reject the ith test personnel. including: ​ ​ ​ ​ ​ ​ ​ The compressed video to be tested is input into the compression artifact perception model obtained by training, the area where the compression artifact is located in each video frame is detected, and then the total pixel value I of the compression artifact of each video frame of the video is counted p : wherein T represents the compression effect category, M represents the total number of video frame compression effect regions, and respectively represent the width and height of the detection frame. ​ By introducing the video saliency detection algorithm ACLNet to extract the region of interest in each video frame, the area of the salient region in each video frame is counted, and the saliency detection factor I is obtained S ; Recombining the detection box height of the perception model to calculate the visual fovea factor I E : where H denotes the height of the video frame, representing the perception model prediction box height; ​ ​ where x, y represent the center horizontal and vertical coordinates of the bounding box detected by the perception model, I P represents the total pixel value of the compression effect, I S represents the saliency detection factor, I E represents the visual fovea factor, W and H represent the width and height of the video frame, and N represents the total number of compressed video frames.

2. The method of claim 1, wherein the method further comprises: ​ ​ The subjective test adopts a double stimulation method, MOS values of the experimenters are screened after the subjective test is completed, data processing is performed on the collected subjective scores, and a MOS value of each test video is calculated according to the following formula: , wherein represents a score of each tester on the same video; represents a number of testers; then a correlation coefficient between the score of each tester and the MOS is calculated to ensure effectiveness of the obtained data; and finally, the test data is screened according to ITU-R BT.1788.

4. The method of claim 3, wherein the compression artifacts are perceptually evaluated based on the human visual system. ​ ​ ​ ​ ​ ​ 5. A system for compression artifacts perception and evaluation in high efficiency video coding for implementing the method of claim 1, characterized by, ​ The training module is configured to train a deep compression effect perception model combined with attention learning and data reweighting, input positive and negative samples of each compression effect into the network, and train to obtain a corresponding compression effect perception model. The video compression effect intensity value calculation module is configured to calculate a video compression effect intensity value based on the trained compression effect perception model and considering the influence of visual saliency and visual fovea effect. The subjective test module is configured to calculate an average subjective score of a video through a subjective test. The evaluation module is configured to establish a mapping relationship between the video compression effect intensity value and the average subjective quality score of the video, and realize video quality evaluation oriented to compression effect perception.

6. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to implement the steps of the compression effect perception and evaluation method in the high-efficiency video coding according to any one of claims 1-4.

7. A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the compression effect perception and evaluation method in the high-efficiency video coding according to any one of claims 1-4.

Images