Evaluation method, device, computer storage medium and terminal device

Abstract

The present disclosure provides an evaluation method, an evaluation device, a computer storage medium and a terminal device. The method comprises: firstly obtaining first images corresponding to each target scene in a target scene set, then obtaining second images corresponding to each first image, wherein the second image is an image with the highest matching degree with the first user preference data of the first image in a sample image set corresponding to the target scene, then obtaining second user preference data corresponding to each second image, generating first user preference data of the first image corresponding to each second image based on each second user preference data, and finally generating evaluation data based on each first user preference data, thereby realizing the evaluation of the imaging capability of the terminal from the perspective of user preference and based on multi-target scenes.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to an evaluation method, an evaluation device, a computer storage medium, and a terminal device. Background Art

[0002] Image capture is an important part of smartphones and has gradually become a focus of consumers. The quality of images captured directly affects the sales of smartphones. Therefore, creating an image capture function that satisfies consumers is the key to increasing smartphone sales.

[0003] Public Content

[0004] The purpose of the present disclosure is to provide an evaluation method, an evaluation device, a computer storage medium and a terminal device, which can evaluate and score the imaging capabilities of the terminal from the perspective of user preferences, and the scoring results are in line with the user preferences.

[0005] According to one aspect of the present disclosure, there is provided an evaluation method, the method comprising:

[0006] Acquire first images corresponding to each target scene in the target scene set;

[0007] Acquire a second image corresponding to each of the first images, where the second image is an image with the highest matching degree with the first user preference data of the first image in the sample image set corresponding to the target scene;

[0008] acquiring second user preference data corresponding to each second image, and generating first user preference data of the first image corresponding to each second image based on each second user preference data;

[0009] Evaluation data is generated according to each of the first user preference data.

[0010] According to another aspect of the present disclosure, there is provided an evaluation device, the device comprising:

[0011] A first image acquisition module, used to acquire first images corresponding to each target scene in the target scene set;

[0012] A second image acquisition module, used to acquire a second image corresponding to each of the first images, wherein the second image is an image in a set of sample images corresponding to the target scene that has the highest matching degree with the first user preference data of the first image;

[0013] a first data generating module, configured to obtain second user preference data corresponding to each second image, and generate first user preference data of the first image corresponding to each second image based on each second user preference data;

[0014] An evaluation data generating module is used to generate evaluation data according to each of the first user preference data.

[0015] According to another aspect of the present disclosure, a terminal device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the evaluation method in the above embodiment when executing the computer program.

[0016] According to another aspect of the present disclosure, a computer storage medium is provided, on which a computer program is stored, characterized in that when the computer program is executed by a processor, the evaluation method in the above embodiment is implemented.

[0017] The evaluation method, evaluation device, computer storage medium, and terminal device provided by the embodiments of the present disclosure have the following technical effects:

[0018] This technical solution obtains first images corresponding to each target scene in a target scene set, obtains second images corresponding to each of the first images, then obtains second user preference data corresponding to each of the second images, and generates first user preference data of the first images corresponding to each of the second images based on each of the second user preference data. Finally, evaluation data is generated based on each of the first user preference data, thereby evaluating the imaging capability of the terminal from the perspective of user preference. The evaluation result depends on the user preference and is more representative. The imaging capability of the terminal is evaluated based on multiple target scenes, thereby ensuring the authenticity and reliability of the scoring result. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the present disclosure, and together with the specification, are used to explain the principles of the present disclosure. Obviously, the drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work;

[0020] Figure 1 A schematic diagram of a flow chart of an evaluation method is provided for an embodiment of the present application;

[0021] Figure 2 A flow chart of another evaluation method is provided for the embodiment of the present application;

[0022] Figure 3 A schematic diagram of another evaluation method provided in an embodiment of the present application;

[0023] Figure 4 An example schematic diagram of training a user preference model is provided for an embodiment of the present application;

[0024] Figure 5 An example schematic diagram of calculating the first user's preference data is provided for an embodiment of the present application;

[0025] Figure 6 A schematic diagram of the structure of an evaluation device is provided for an embodiment of the present application;

[0026] Figure 7 A schematic diagram of the structure of an evaluation device is provided for an embodiment of the present application;

[0027] Figure 8 A structural schematic diagram of a model training module is provided for an embodiment of the present application;

[0028] Fig. 9 A structural schematic diagram of a first data generation module is provided for an embodiment of the present application;

[0029] Fig.10 A structural block diagram of a terminal device is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0030] In order to make the objectives, technical solutions and advantages of the present disclosure more clear, the embodiments of the present disclosure will be further described in detail below with reference to the accompanying drawings.

[0031] When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are only examples of devices and methods consistent with some aspects of the present disclosure as detailed in the attached claims.

[0032] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as being limited to the examples set forth herein; on the contrary, these embodiments are provided so that the present disclosure will be more comprehensive and complete, and the concepts of the example embodiments are fully conveyed to those skilled in the art. The described features, structures, or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, known technical solutions are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0033] In addition, the accompanying drawings are only schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0034] In the prior art, most of the solutions for evaluating the terminal imaging capability are based on an objective perspective, and objective evaluation is performed through various objective parameters to generate a terminal imaging capability score. For example, Vodafone's VCX image evaluation system generates a camera capability score through various objective evaluations. Or in some solutions, there is also a combination of objective evaluation of objective tools and subjective evaluation of people, and the final mobile phone imaging capability score is generated through the subjective and objective scores, such as DxOMark.

[0035] However, the terminal imaging capability evaluation schemes existing in the prior art cannot reflect user preferences. For example, the subjective evaluation in DxOMark is also professionally scored by trained professionals according to the image clarity, brightness, color and other dimensions, which cannot fully fit the actual situation of users. The final imaging capability score is not representative of users. In addition, the calculation of each objective indicator in the objective evaluation requires professional lighting equipment and specified picture cards, which is extremely inconvenient and inefficient.

[0036] Based on this, the present application provides an evaluation method, which obtains first images corresponding to each target scene in a target scene set, obtains second images corresponding to each first image, and then obtains second user preference data corresponding to each second image, and generates first user preference data of the first image corresponding to each second image based on each second user preference data, and finally generates evaluation data based on each first user preference data, thereby realizing the evaluation of the imaging capability of the terminal from the perspective of user preference. The evaluation result depends on the user preference and is more representative. The imaging capability of the terminal is evaluated based on multiple target scenes, which ensures the authenticity and reliability of the scoring result. There is no need to calculate objective indicators, and the efficiency is higher than that of traditional evaluation schemes based on objective indicator calculation.

[0037] The following is a detailed description in conjunction with specific embodiments. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims. The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily follow the steps shown. For example, some steps are parallel and do not have a strict logical order, so the actual execution order is variable.

[0038] In one embodiment, an evaluation method is proposed. The method can be implemented by a computer program and can be run on an evaluation device based on the von Neumann architecture. The computer program can be integrated into an application or run as an independent tool application. Figure 1 , which is a flow chart of an evaluation method provided in the embodiment of the present application. Figure 1 As shown, the evaluation method may include the following steps:

[0039] S101, acquiring first images corresponding to respective target scenes in a target scene set;

[0040] The target scene refers to the user's usage scene, the scene in which the user takes the image, for example, the target scene can be a portrait, night scene, sea view, city, sunrise, sunset, blue sky, etc. The target scene set refers to a set covering all usage scenes of a user.

[0041] In one embodiment, a first image of each target scene in a target scene set of a user is obtained respectively.

[0042] It should be understood that different users have different user habits, and their corresponding target scenes are also different. Each user has a target scene set that is unique to the user. For example, user A likes to take selfies and often takes portraits, while user B likes nature and often takes photos in scenes such as sunrise, sunset, blue sky and white clouds. In this embodiment, the target scene set is not a fixed set, and the target scene set corresponding to different users may be different.

[0043] In one embodiment, a user's target scene set is first determined, and then a first image is acquired for each target scene in the user's target scene set. The first image is used to evaluate the imaging capability of the terminal.

[0044] Optionally, the obtaining of the first image corresponding to each target scene in the target scene set may be taking the first image of each target scene on site, or may be selecting the first image of each target scene from a picture set saved by the user.

[0045] S102, acquiring a second image corresponding to each of the first images, where the second image is an image in a set of sample images corresponding to the target scene that has the highest matching degree with first user preference data of the first image;

[0046] In one embodiment, before obtaining the first images respectively corresponding to each target scene in the target scene set, a sample image set respectively corresponding to each target scene in the target scene set is obtained, and the user's score for each sample image in the sample image set is obtained to obtain second user preference data corresponding to each sample image, and based on each sample image in the sample image set and the second user preference data corresponding to each sample image, a user preference model corresponding to each target scene in the target scene set is trained.

[0047] The first user preference data refers to the user preference data corresponding to the first image, and the user preference data is used to indicate the user's preference degree for the image. For example, if the user preference data for image A is better than the user preference data for image B, it indicates that the user likes image A more than image B, where image A and image B are different images under the same target scene.

[0048] In one embodiment, the user preference data may be a user preference score for an image, and a higher user preference score indicates that the user likes the image more highly.

[0049] It should be understood that the target scenes of the second image corresponding to the first image should be the same, that is, for a first image belonging to target scene 1, its corresponding second image should also belong to target scene 1.

[0050] In one embodiment, a second image corresponding to the first image is obtained, wherein the second image and the first image belong to the same target scene, the second image is a sample image with the highest matching degree with the first user preference data of the first image in a sample image set corresponding to the target scene of the first image, and the second user preference data of each sample image in the sample image set is obtained by user scoring and is known data, that is, in the sample image set, the second user preference data of the sample image corresponding to the second image is closest to the first user preference data corresponding to the first image.

[0051] Optionally, the acquiring of the second image corresponding to the first image may be by determining, from a set of sample images corresponding to a target scene of the first image, a sample image having the highest matching degree with first user preference data of the first image, and using the sample image as the second image.

[0052] Optionally, the method of obtaining the second image corresponding to the first image may be to input each sample image in the sample image set and the first image into the user preference model in turn, obtain the user preference ratio between the first image and each sample image, determine the sample image with the highest degree of match with the first user preference data of the first image based on each user preference ratio, and use the sample image as the second image. Furthermore, the method of determining the sample image with the highest degree of match with the first user preference data of the first image based on each user preference ratio may be to use the sample image with the user preference ratio closest to 1 with the first image as the sample image with the highest degree of match with the first user preference data of the first image.

[0053] Optionally, the method of determining, from a sample image set corresponding to a target scene of the first image, a sample image with the highest degree of match with the first user preference data of the first image, and using the sample image as the second image, may be specifically implemented as follows: sorting each sample image in the sample image set based on the second user preference data corresponding to each sample image in the sample image set to obtain a sample image queue, and obtaining, from the sample image queue, the second image with the highest degree of match with the user preference data of the first image based on a binary method.

[0054] Optionally, the method of obtaining the second image with the highest degree of match with the first user preference data of the first image in the sample image queue based on the dichotomy method may include: inputting the sample images in the sample image queue and the first image into a trained user preference model in sequence based on the dichotomy method, determining the sample image with the highest degree of match with the first user preference data of the first image in the sample image queue based on the user preference ratio output by the user preference model, and using the sample image as the second image.

[0055] S103, acquiring second user preference data corresponding to each second image, and generating first user preference data of the first image corresponding to each second image based on each second user preference data;

[0056] The second image is the sample image in the sample image set that has the highest matching degree with the first user preference data of the first image. The second image is an image that has been surveyed and rated by users, so the second user preference data corresponding to the second image is known.

[0057] In one embodiment, the generation of the first user preference data of the first image corresponding to each second image based on each second user preference data may be: each first image and the second image corresponding to each first image are respectively input into a user preference model corresponding to a target scene of each first image, to generate the first user preference data corresponding to each first image.

[0058] The user preference model is generated based on each sample image in the sample image set and the second user preference data corresponding to each sample image. For any two images input into the user preference model, it can output the user preference ratio between the two images.

[0059] The user preference ratio refers to the ratio of the user's preference for two images. For example, image A and image B are input into the user preference model together, and the user preference model can calculate the ratio of the user's preference for image A to the user's preference for image B according to the settings. In some embodiments, if the ratio is the user's preference for image A to the user's preference for image B, if the ratio is greater than 1, it indicates that the user prefers image A, and if the ratio is less than 1, it indicates that the user prefers image B; if the ratio is the user's preference for image B to the user's preference for image A, if the ratio is greater than 1, it indicates that the user prefers image B, and if the ratio is less than 1, it indicates that the user prefers image A.

[0060] In one embodiment, the first image and a second image corresponding to the first image are input into a user preference model corresponding to a target scene of the first image, the user preference model calculates a target user preference ratio between the first image and the second image, and then generates first user preference data corresponding to the first image based on the target user preference ratio and the known second user preference data of the second image.

[0061] It is not difficult to understand that the second image is the sample image with the highest matching degree with the first user preference data of the first image in the sample image set corresponding to the target scene of the first image, and therefore, the second user preference data of the second image is known. The first image and the second image corresponding to the first image are input into the user preference model corresponding to the target scene of the first image, and after obtaining the target user preference ratio between the first image and the second image, if the target user preference ratio is the ratio of the first image to the second image, the product of the target user preference ratio and the second user preference data is calculated, and the product is used as the first user preference data corresponding to the first image; if the target user preference ratio is the ratio of the second image to the first image, the quotient of the target user preference ratio and the second user preference data is calculated, and the quotient is used as the first user preference data corresponding to the first image.

[0062] Further, optionally, the user preference data is a user preference score. Since the target user preference ratio may not be very accurate, after obtaining the target user preference ratio between the first image and the second image, the second user preference data corresponding to the second image is fine-tuned based on the target user preference ratio, and the fine-tuned second user preference data is used as the first user preference data of the first image. For example, if the target user preference ratio indicates that the user prefers the first image, the second user preference score corresponding to the second image is fine-tuned upward, and the fine-tuned second user preference data is used as the first user preference data of the first image; if the target user preference ratio indicates that the user prefers the second image, the second user preference data corresponding to the second image is fine-tuned downward, and the fine-tuned second user preference data is used as the first user preference data of the first image.

[0063] S104: Generate evaluation data according to each of the first user preference data.

[0064] In one embodiment, evaluation data of the imaging capability of the terminal is generated based on first user preference data of each first image corresponding to each target scene.

[0065] Optionally, the evaluation data may be quantified into a score.

[0066] Optionally, generating evaluation data according to each of the first user preference data includes: calculating a sum of the first user preference data of each of the first images corresponding to each of the target scenes, and using the sum as the evaluation data of the terminal.

[0067] Optionally, generating evaluation data according to each of the first user preference data includes: calculating an average value of the first user preference data of each of the first images corresponding to each of the target scenes, and using the average value as the evaluation data of the terminal.

[0068] Optionally, generating evaluation data based on the first user preference data includes: obtaining the user's usage frequency of each target scene, generating a weight corresponding to each target scene in the target scene set based on each usage frequency, and weighted summing the first user preference data corresponding to each first image according to the weight corresponding to each target scene to obtain the evaluation data of the terminal.

[0069] In an embodiment of the present application, by acquiring first images corresponding to each target scene in a target scene set, and then acquiring second images corresponding to each of the first images, the second user preference data of the second image is known, and then the first image and the second image are input into a user preference model to obtain the first user preference data of each first image, and finally generating evaluation data of the terminal based on the first user preference data corresponding to the first images corresponding to each target scene, the imaging capability of the terminal is evaluated from the perspective of user preference, the evaluation result depends on the user preference and is more representative, and the imaging capability of the terminal is evaluated based on multiple target scenes, thereby ensuring the authenticity and reliability of the evaluation result.

[0070] In one embodiment, a training process for a user preference model is included. Figure 2 , provides a flow chart of another evaluation method for the embodiment of the present application. Figure 2 As shown, the evaluation method may include the following steps:

[0071] S201, obtaining a set of sample images corresponding to each target scene in a set of target scenes;

[0072] In one embodiment, a target scene set covering all target scenes is first constructed, and then corresponding sample image sets are obtained for different target scenes in the target scene set. The sample image sets are used to train the user preference model under the corresponding target scene.

[0073] Optionally, in one embodiment, a scene system covering all target scenes can be constructed based on features such as ambient brightness, location information, and subject, and a target scene set can be generated. For example, the subject can be set to a human face, a single person, a group photo, a child, an old person, an animal, a plant, a mountain, a sea, etc. The location information can be quantified based on examples of the subject and the imaging system, and the ambient brightness can be quantified based on the ambient light intensity. Finally, different target scenes can be set based on different subjects, different location information, and different ambient light intensities.

[0074] The sample image set is a set of all sample images in a certain target scene. The sample image set is used to train the user preference model in the corresponding target scene. The more sample images there are in the sample image set, the better the effect of the trained user preference model. Optionally, each sample image in the sample image set can be not only taken by the development team itself, but also obtained from user feedback. Furthermore, in order to ensure the model effect of the user preference model, the number of sample images in the sample image set should not be less than a preset threshold. Optionally, the preset threshold is 100 images.

[0075] S202, obtaining the user's score on each sample image in the sample image set, and obtaining second user preference data corresponding to each sample image;

[0076] In one embodiment, online or offline user surveys are conducted to obtain users' preference scores for each sample image in the sample image set, and the preference scores of each user for each sample image in the sample image set are summarized to obtain second user preference data corresponding to each sample image.

[0077] It should be understood that the second user preference data is obtained through online or offline user surveys and is the actual rating of users, which can represent the actual user preference for each sample image.

[0078] S203, training a user preference model corresponding to each target scene in the target scene set based on each sample image in the sample image set and the second user preference data corresponding to each sample image;

[0079] In one embodiment, the user preference model may be a machine learning model based on a convolutional neural network or a support vector machine regression prediction model.

[0080] In one embodiment, each sample image in the sample set is used as a training sample and is sequentially input into the user preference model to obtain a first result output by the user preference model, a loss value between the first result and the corresponding second user preference data is calculated, and the parameter value in the user preference output model is optimized based on the loss value, and the training is iterated repeatedly until the loss value is less than a preset threshold. Optionally, the first result may be third user preference data.

[0081] exist Figure 2 Based on Figure 3 Step S203 may include the following steps S2031 to S2034:

[0082] S2031, pairing each sample image in the sample image set in pairs to obtain sample image pairs;

[0083] In one embodiment, each sample image in the sample image set is randomly paired with each other to obtain each sample image pair.

[0084] Optionally, the sample images in the sample image set are sorted from high to low according to the second user preference data corresponding to each sample image, and paired in sequence according to the sorting results to obtain sample image pairs. For example, if the sorting result is 1, 2, 3, 4, 5, 6, then 1 and 2 are paired, 3 and 4 are paired, and 5 and 6 are paired.

[0085] Optionally, the sample images in the sample image set are sorted from high to low according to the second user preference data corresponding to each sample image, and according to the sorting result, the first and the last in the sorting result are paired, the second and the second to last are paired, and so on and so forth to obtain each sample image pair.

[0086] S2032, inputting the sample image pair into a user preference model to obtain a user preference ratio between each sample image in the sample image pair;

[0087] In one embodiment, the sample image pair is input into a user preference model, and the user preference model outputs a user preference ratio between two sample images in the sample image pair. The user preference ratio is used to indicate which sample image the user prefers.

[0088] In one embodiment, firstly, the first measurement index of each sample image in the sample image pair in each preset dimension is extracted, and the first measurement index of each dimension is input into the user preference model to generate a user preference ratio between each sample image in the sample image pair, and the dimension includes at least one of color, resolution, brightness, distortion, noise, and signal-to-noise ratio.

[0089] S2033, calculating a loss value based on the user preference ratio and second user preference data corresponding to two sample images in the sample image pair;

[0090] The second user preference data corresponding to the two sample images in the sample image pair are real user preference data obtained based on user surveys. The second user preference data can be used as a training standard, so that the user preference model obtained through training is more in line with user preferences and more representative of users.

[0091] In one embodiment, based on the sample image and the second user preference data respectively corresponding to the two sample images, a ratio of the second user preference data is obtained, and the loss value is calculated based on the ratio and the user preference ratio.

[0092] S2034: Training the user preference model based on the loss value.

[0093] In one embodiment, each sample image pair is input into the user preference model in sequence, and a loss value is calculated based on the user preference ratio output by the user preference model and the second user preference data corresponding to the two sample images in the sample image pair. The user preference model is trained based on the loss value, and the training process is iteratively performed until the loss values ​​calculated for five consecutive times are all less than a preset threshold, and the training ends.

[0094] See also Figure 4, which is a schematic diagram of an example of training a user preference model in the embodiment of the present application. Figure 4 As shown, each sample image in the sample image set is paired up to obtain sample image pairs, and the sample image pairs are input into the user preference model in sequence. The user preference model outputs a user preference ratio, and the user preference ratio is compared with the known second user preference data of the sample image to calculate a loss value, and the user preference model is trained based on the loss value.

[0095] S204, acquiring first images corresponding to each target scene in the target scene set;

[0096] The target scene refers to the scene in which the user is when shooting an image, for example, the target scene may be a portrait, night scene, sea view, city, sunrise, sunset, blue sky, etc. The target scene set refers to a set of all target scenes covering a user.

[0097] In one embodiment, a first image of each target scene in a target scene set of a user is obtained respectively.

[0098] It should be understood that different users have different user habits, and their corresponding target scenes are also different. Each user has a target scene set that is unique to the user. For example, user A likes to take selfies and often takes portraits, while user B likes nature and often takes photos in scenes such as sunrise, sunset, blue sky and white clouds. In this embodiment, the target scene set is not a fixed set, and the target scene set corresponding to different users may be different.

[0099] In one embodiment, a user's target scene set is first determined, and then a first image is acquired for each target scene in the user's target scene set. The first image is used to evaluate the imaging capability of the terminal.

[0100] Optionally, the obtaining of the first image corresponding to each target scene in the target scene set may be taking the first image of each target scene on site, or may be selecting the first image of each target scene from a picture set saved by the user.

[0101] S205, sorting each sample image in the sample image set based on the second user preference data corresponding to each sample image in the sample image set to obtain a sample image queue;

[0102] In one embodiment, the sample images in the sample image set are sorted from high to low according to the second user preference data corresponding to the sample images to obtain a sample image queue.

[0103] Optionally, the sorting method may also be to sort the sample images from low to high according to the second user preference data corresponding to each sample image, which is not limited in this embodiment of the present application.

[0104] S206, obtaining, from the sample image queue, a second image with the highest matching degree with the first user preference data of the first image based on a binary search method;

[0105] The method of obtaining the second image with the highest matching degree with the first user preference data of the first image in the sample image queue based on the dichotomy method specifically includes: inputting the first image and the sample image sorted in the middle in the sample image queue into a trained user preference model, the user preference model outputting the user preference ratio between the first image and the sample image, determining the next sample image in the sample image queue based on the dichotomy method based on the user preference ratio, and executing the step of inputting the first image and the sample image into the trained user preference model, continuously narrowing the optional range of the sample image based on the user preference ratio until the position of the first image in the sample image queue is determined, and taking any of the sample images adjacent to it as the second image.

[0106] For example, the sample image queue includes: sample image 1, sample image 2, sample image 3, sample image 4, sample image 5 ... sample image 99, sample image 100. First, the first image and sample image 50 are input into the user preference model. If the user preference model outputs a user preference ratio of the user preference of the first image to the user preference of sample image 50, if the user preference ratio is greater than 1, it indicates that the user prefers the first image, and the position of the first image in the sample image queue should be between sample image 1 and sample image 50; then the first image and sample image 25 are input into the user preference model to obtain the user preference ratio, compare the user's preference for the first image and sample image 25, and narrow the range of the position of the first image in the sample image queue according to the user preference ratio until the position of the first image in the sample image queue is determined, and finally any sample image adjacent to it is used as the second image.

[0107] S207, extracting second measurement indicators of the first image and the second image in each preset dimension, and inputting the second measurement indicators of each dimension into a user preference model corresponding to a target scene of the first image;

[0108] The dimension includes at least one of color, resolution, brightness, distortion, noise, and signal-to-noise ratio. The second measurement index refers to an index value in the form of a vector corresponding to each dimension. For example, in the brightness dimension, the second measurement index is an index value expressing the brightness value of the image.

[0109] S208, generating a target user preference ratio between the first image and the second image based on the second measurement indicators in each dimension;

[0110] In one embodiment, second measurement indicators of the first image and the second image in each preset dimension are extracted, and the second measurement indicators of each dimension are input into a user preference model corresponding to the target scene of the first image. The user preference model generates a target user preference ratio between the first image and the second image based on the second measurement indicators of each dimension.

[0111] It is worth mentioning that different target scenes correspond to different user preference models, and the first image and the user preference model should belong to the same target scene.

[0112] S209, calculating and obtaining first user preference data corresponding to the first image based on the target user preference ratio and second user preference data corresponding to the second image;

[0113] In one embodiment, if the target user preference ratio is the ratio of the first image to the second image, the product of the target user preference ratio and the second user preference data is calculated, and the product is used as the first user preference data corresponding to the first image; if the target user preference ratio is the ratio of the second image to the first image, the quotient of the target user preference ratio and the second user preference data is calculated, and the quotient is used as the first user preference data corresponding to the first image.

[0114] It is not difficult to understand that the target user preference ratio represents the ratio of the user's preference for the first image to the user's preference for the second image, and the second user preference data of the second image is known. Based on the ratio and the second user preference data, the first user preference data corresponding to the first image can be obtained.

[0115] See also Figure 5 , an example schematic diagram of calculating the first user preference data is provided for an embodiment of the present application. The first image and the second image are input into the trained user preference model, the user preference model outputs the target user preference ratio between the first image and the second image, and then based on the target user preference ratio and the second user preference data corresponding to the second image, the first user preference data corresponding to the first image is obtained.

[0116] S210, obtaining the usage frequency of each target scene by the user, and generating a weight corresponding to each target scene in the target scene set based on each usage frequency;

[0117] S211 , performing weighted summation on the first user preference data corresponding to each of the first images according to the weight corresponding to each target scene to obtain evaluation data.

[0118] It should be understood that the evaluation of a terminal's imaging capability cannot be conducted solely from one target scene, but rather the terminal's imaging capability in each target scene should be comprehensively evaluated to obtain the final evaluation data.

[0119] In this embodiment, by obtaining the user's usage frequency of each target scene, a weight corresponding to each target scene is generated based on the user's usage frequency of each target scene, and then based on the weight corresponding to each target scene, a weighted sum is taken of the first user preference data of each first image corresponding to each target scene to obtain the imaging capability score of the terminal.

[0120] It is understandable that the more frequently a user uses a target scenario, the greater its corresponding weight will be, and vice versa.

[0121] In an embodiment of the present application, second user preference data of a sample image is obtained based on a user survey, and then a user preference model is trained based on the sample image and the corresponding second user preference data, thereby ensuring that the user preference model conforms to the user's preference and is representative of the user; by obtaining each first image in a multi-target scene, and determining the second image corresponding to the first image in a sample image set based on the trained user preference model, the first image and the second image are input into the user preference model to obtain a target user preference ratio, and the first user preference data corresponding to the first image is calculated based on the target user preference ratio and the second user preference data corresponding to the second image, and finally, a weight corresponding to each target scene is generated based on the user's frequency of use of each target scene, and the first user preference data in each target scene is weightedly summed based on the weight to obtain evaluation data corresponding to the terminal, thereby realizing the evaluation of the imaging capability of the terminal from the perspective of user preference, and the evaluation result depends on the user preference and is more representative, and the imaging capability of the terminal is comprehensively evaluated based on the multi-target scene, thereby ensuring the authenticity and reliability of the evaluation result.

[0122] The following are embodiments of the device disclosed herein, which can be used to execute the method embodiments disclosed herein. For details not disclosed in the device embodiments disclosed herein, please refer to the method embodiments disclosed herein.

[0123] See also Figure 6 , is a schematic diagram of the structure of an evaluation device provided in an embodiment of the present application. Figure 6 As shown, the evaluation device 1 can be implemented as all or part of the terminal device through software, hardware or a combination of both. According to some embodiments, the evaluation device 1 includes a first image acquisition module 11, a second image acquisition module 12, a first data generation module 13, and an evaluation data generation module 14, specifically including:

[0124] A first image acquisition module 11 is used to acquire first images corresponding to each target scene in the target scene set;

[0125] A second image acquisition module 12 is used to acquire a second image corresponding to each of the first images, where the second image is an image with the highest matching degree with the first user preference data of the first image in the sample image set corresponding to the target scene;

[0126] A first data generating module 13 is used to obtain second user preference data corresponding to each second image, and generate first user preference data of the first image corresponding to each second image based on each second user preference data;

[0127] The evaluation data generating module 14 is used to generate evaluation data according to each of the first user preference data.

[0128] Optional, see Figure 7 , which is a schematic diagram of the structure of an evaluation device provided in an embodiment of the present application, such as Figure 7 As shown, the evaluation device further includes:

[0129] The sample image acquisition module 15 is used to acquire a sample image set corresponding to each target scene in the target scene set;

[0130] The second data generating module 16 is used to obtain the user's score on each sample image in the sample image set, and obtain the second user preference data corresponding to each sample image;

[0131] The model training module 17 is used to train the user preference model corresponding to each target scene in the target scene set based on each sample image in the sample image set and the second user preference data corresponding to each sample image.

[0132] Optional, see Figure 8 , which is a schematic diagram of the structure of a model training module provided in the embodiment of the present application, such as Figure 8 As shown, the model training module 17 includes:

[0133] An image pairing unit 171 is used to pair each sample image in the sample image set with each other to obtain sample image pairs;

[0134] A first preference ratio generating unit 172, configured to input the sample image pair into a user preference model to obtain a user preference ratio between each sample image in the sample image pair;

[0135] A loss value calculation unit 173, configured to calculate a loss value based on the user preference ratio and second user preference data corresponding to the two sample images in the sample image pair;

[0136] The model training unit 174 is used to train the user preference model based on the loss value.

[0137] Optionally, the first preference ratio generating unit 172 is specifically configured to:

[0138] Extract the first measurement index of each sample image in the sample image pair in each preset dimension, input the first measurement index of each dimension into the user preference model, and generate a user preference ratio between each sample image in the sample image pair, wherein the dimension includes at least one of color, resolution, brightness, distortion, noise, and signal-to-noise ratio.

[0139] Optionally, the second image acquisition module 12 is specifically used to:

[0140] Sort each sample image in the sample image set based on the second user preference data corresponding to each sample image in the sample image set to obtain a sample image queue;

[0141] Based on the binary method, a second image with the highest matching degree with the first user preference data of the first image is obtained from the sample image queue.

[0142] Optional, see Fig. 9 , which is a schematic diagram of the structure of a first data generation module provided in an embodiment of the present application, such as Fig. 9 As shown, the first data generating module 13 includes:

[0143] A measurement index extraction unit 131 is used to extract second measurement indexes of the first image and the second image in each preset dimension, and input the second measurement index of each dimension into a user preference model corresponding to a target scene of the first image;

[0144] A second preference ratio generating unit 132, configured to generate a target user preference ratio between the first image and the second image based on the second measurement index in each dimension;

[0145] The first data calculation unit 133 is configured to calculate the first user preference data corresponding to the first image based on the target user preference ratio and the second user preference data corresponding to the second image.

[0146] Optionally, the first data calculation unit 133 is specifically configured to:

[0147] If the target user preference ratio is a ratio of the first image to the second image, then calculating the product of the target user preference ratio and the second user preference data, and using the product as the first user preference data corresponding to the first image;

[0148] If the target user preference ratio is a ratio of the second image to the first image, the quotient of the target user preference ratio and the second user preference data is calculated, and the quotient is used as the first user preference data corresponding to the first image.

[0149] Optionally, the evaluation data generating module 14 is specifically used for:

[0150] Obtain the user's usage frequency of each target scene, and generate a weight corresponding to each target scene in the target scene set based on each usage frequency;

[0151] The first user preference data corresponding to each of the first images is weightedly summed according to the weights corresponding to each target scene to obtain evaluation data.

[0152] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0153] By adopting the evaluation method provided in the present application, the second user preference data of the sample image is obtained based on the user survey, and then the user preference model is trained based on the sample image and the corresponding second user preference data, so as to ensure that the user preference model conforms to the user preference and is representative of the user; by obtaining each first image under a multi-target scene, and determining the second image corresponding to the first image in the sample image set based on the trained user preference model, the first image and the second image are input into the user preference model to obtain the target user preference ratio, and the first user preference data corresponding to the first image is calculated based on the target user preference ratio and the second user preference data corresponding to the second image, and finally, the weight corresponding to each target scene is generated based on the user's frequency of use of each target scene, and the first user preference data under each target scene is weightedly summed based on the weight to obtain the evaluation data of the terminal, so as to realize the evaluation of the imaging capability of the terminal from the perspective of user preference, and the evaluation result depends on the user preference and is more representative, and the imaging capability of the terminal is evaluated comprehensively based on the multi-target scene, so as to ensure the authenticity and reliability of the evaluation result.

[0154] The present application also provides a computer storage medium, which can store multiple instructions, and the instructions are suitable for being loaded and executed by a processor as described above. Figure 1 to Figure 5 The evaluation method of the embodiment shown in the figure can be specifically implemented by referring to Figure 1 to Figure 5 The specific description of the illustrated embodiment will not be repeated here.

[0155] The present application also provides a computer program product, which stores at least one instruction, and the at least one instruction is loaded and executed by the processor as described above. Figure 1 to Figure 5 The evaluation method of the embodiment shown in the figure can be specifically implemented by referring to Figure 1 to Figure 5The specific description of the illustrated embodiment will not be repeated here.

[0156] Please refer to Fig.10 , provides a block diagram of the structure of a terminal device for an embodiment of the present application. The terminal device in the present application may include one or more of the following components: a processor 110, a memory 120, an input device 130, an output device 140, and a bus 150. The processor 110, the memory 120, the input device 130, and the output device 140 may be connected via a bus 150.

[0157] The processor 110 may include one or more processing cores. The processor 110 uses various interfaces and lines to connect various parts of the entire terminal device, and executes various functions and processes data of the terminal device 100 by running or executing instructions, programs, code sets or instruction sets stored in the memory 120, and calling data stored in the memory 120. Optionally, the processor 110 can be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), and programmable logic array (PLA). The processor 110 can integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. Among them, the CPU mainly processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing display content; and the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 110, but may be implemented separately through a communication chip.

[0158] The memory 120 may include a random access memory (RAM) or a read-only memory (ROM). Optionally, the memory 120 includes a non-transitory computer-readable storage medium. The memory 120 may be used to store instructions, programs, codes, code sets, or instruction sets.

[0159] The input device 130 is used to receive input commands or data, and the input device 130 includes but is not limited to a keyboard, a mouse, a camera, a microphone, or a touch device. The output device 140 is used to output commands or data, and the output device 140 includes but is not limited to a display device and a speaker. In the embodiment of the present application, the input device 130 may be a temperature sensor for obtaining the operating temperature of the terminal device. The output device 140 may be a speaker for outputting audio signals.

[0160] In addition, those skilled in the art will appreciate that the structure of the terminal device shown in the above drawings does not constitute a limitation on the terminal device, and the terminal device may include more or fewer components than shown, or combine certain components, or arrange the components differently. For example, the terminal device also includes a radio frequency circuit, an input unit, a sensor, an audio circuit, a wireless fidelity (WiFi) module, a power supply, a Bluetooth module, and other components, which will not be described in detail here.

[0161] In the embodiment of the present application, the execution subject of each step may be the terminal device described above. Optionally, the execution subject of each step is the operating system of the terminal device. The operating system may be an Android system, an IOS system, or other operating systems, which is not limited in the embodiment of the present application.

[0162] exist Fig.10 In the terminal device shown, the processor 110 can be used to call the image capability assessment program stored in the memory 120 and execute it to implement the assessment method as described in the various method embodiments of the present application.

[0163] By adopting the evaluation method provided in the present application, the second user preference data of the sample image is obtained based on the user survey, and then the user preference model is trained based on the sample image and the corresponding second user preference data, so as to ensure that the user preference model conforms to the user preference and is representative of the user; by obtaining each first image in a multi-target scene, and determining the second image corresponding to the first image in the sample image set based on the trained user preference model, the first image and the second image are input into the user preference model to obtain the target user preference ratio, and the first user preference data corresponding to the first image is calculated based on the target user preference ratio and the second user preference data corresponding to the second image, and finally, the weight corresponding to each target scene is generated based on the user's frequency of use of each target scene, and the first user preference data in each target scene is weightedly summed based on the weight to obtain the evaluation data of the terminal, so as to realize the evaluation of the imaging capability of the terminal from the perspective of user preference, and the evaluation result depends on the user preference and is more representative, and the imaging capability of the terminal is comprehensively evaluated based on the multi-target scene, so as to ensure the authenticity and reliability of the evaluation result.

[0164] Those skilled in the art can clearly understand that the technical solution of the present application can be implemented with the help of software and / or hardware. The "unit" and "module" in this specification refer to software and / or hardware that can independently complete or cooperate with other components to complete specific functions, where the hardware can be, for example, a field programmable gate array (FPGA), an integrated circuit (IC), etc.

[0165] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0166] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0167] In the several embodiments provided in the present application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of the units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0168] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0169] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0170] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by entering a program to instruct the relevant hardware. The program may be stored in a computer-readable memory, and the memory may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0171] The above is only an exemplary embodiment of the present disclosure, and the scope of the present disclosure cannot be limited thereto. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the specification and practicing the disclosure here, those skilled in the art will easily think of other embodiments of the present disclosure. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the technical field not recorded in the present disclosure. The description and examples are only regarded as exemplary, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. An evaluation method, characterized in that: The method comprises: Acquire first images corresponding to each target scene in the target scene set; Acquire a second image corresponding to each of the first images, where the second image is an image with the highest matching degree with the first user preference data of the first image in the sample image set corresponding to the target scene; acquiring second user preference data corresponding to each second image, and generating first user preference data of the first image corresponding to each second image based on each second user preference data; generating evaluation data according to each of the first user preference data; The generating evaluation data according to each of the first user preference data comprises: Obtain the user's usage frequency of each target scene, and generate a weight corresponding to each target scene in the target scene set based on each usage frequency; The first user preference data corresponding to each of the first images is weightedly summed according to the weights corresponding to each target scene to obtain evaluation data.

2. The method according to claim 1, characterized in that Before acquiring the first images respectively corresponding to the target scenes in the target scene set, the method further includes: Acquire a set of sample images corresponding to each target scene in the target scene set; Obtaining the user's score for each sample image in the sample image set, and obtaining second user preference data corresponding to each sample image; Based on each sample image in each of the sample image sets and the second user preference data corresponding to each sample image, a user preference model corresponding to each target scene in the target scene set is trained.

3. The method according to claim 2, characterized in that The training of the user preference model corresponding to each target scene in the target scene set based on each sample image in the sample image set and the second user preference data corresponding to each sample image comprises: Pairing each sample image in the sample image set in pairs to obtain sample image pairs; Inputting the sample image pair into a user preference model to obtain a user preference ratio between each sample image in the sample image pair; Calculating a loss value based on the user preference ratio and second user preference data corresponding to two sample images in the sample image pair; The user preference model is trained based on the loss value.

4. The method according to claim 3, characterized in that: The step of inputting the sample image pair into a user preference model to obtain a user preference ratio between each sample image in the sample image pair includes: Extract the first measurement index of each sample image in the sample image pair in each preset dimension, input the first measurement index of each dimension into the user preference model, and generate a user preference ratio between each sample image in the sample image pair, wherein the dimension includes at least one of color, resolution, brightness, distortion, noise, and signal-to-noise ratio.

5. The method according to claim 2, characterized in that: The acquiring of the second image corresponding to each of the first images, wherein the second image is an image in the sample image set corresponding to the target scene that has the highest matching degree with the first user preference data of the first image, includes: Sort each sample image in the sample image set based on the second user preference data corresponding to each sample image in the sample image set to obtain a sample image queue; Based on the binary method, a second image with the highest matching degree with the first user preference data of the first image is obtained from the sample image queue.

6. The method according to claim 2, characterized in that The step of generating first user preference data of the first image corresponding to each second image based on each second user preference data includes: Each of the first images and a second image corresponding to each of the first images are respectively input into a user preference model corresponding to a target scene of each of the first images to generate first user preference data corresponding to each of the first images.

7. The method according to claim 6, characterized in that The step of inputting each of the first images and the second images corresponding to each of the first images into a user preference model corresponding to a target scene of each of the first images to generate first user preference data corresponding to each of the first images includes: Extracting second measurement indicators of the first image and the second image in each preset dimension, and inputting the second measurement indicators of each dimension into a user preference model corresponding to a target scene of the first image; generating a target user preference ratio between the first image and the second image based on the second measurement indicators in each dimension; Based on the target user preference ratio and the second user preference data corresponding to the second image, the first user preference data corresponding to the first image is calculated.

8. The method according to claim 7, characterized in that The calculating the first user preference data corresponding to the first image based on the target user preference ratio and the second user preference data corresponding to the second image includes: If the target user preference ratio is a ratio of the first image to the second image, then calculating the product of the target user preference ratio and the second user preference data, and using the product as the first user preference data corresponding to the first image; If the target user preference ratio is a ratio of the second image to the first image, the quotient of the target user preference ratio and the second user preference data is calculated, and the quotient is used as the first user preference data corresponding to the first image.

9. An evaluation device, characterized in that: The device comprises: A first image acquisition module, used to acquire first images corresponding to each target scene in the target scene set; A second image acquisition module, used to acquire a second image corresponding to each of the first images, wherein the second image is an image in a set of sample images corresponding to the target scene that has the highest matching degree with the first user preference data of the first image; a first data generating module, configured to obtain second user preference data corresponding to each second image, and generate first user preference data of the first image corresponding to each second image based on each second user preference data; An evaluation data generating module, used for generating evaluation data according to each of the first user preference data; The evaluation data generation module is also used to obtain the user's usage frequency of each target scene, and generate a weight corresponding to each target scene in the target scene set based on each usage frequency; and perform weighted summation on the first user preference data corresponding to each of the first images according to the weight corresponding to each target scene to obtain evaluation data.

10. A computer storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

11. A terminal device, characterized in that: include: A processor and a memory; wherein the memory stores a computer program, and the computer program is suitable for being loaded by the processor and executing the steps of the method as claimed in any one of claims 1 to 8.

Images