WO2019015645A1 - 图像处理方法及装置 - Google Patents

图像处理方法及装置 Download PDF

Info

Publication number
WO2019015645A1
WO2019015645A1 PCT/CN2018/096278 CN2018096278W WO2019015645A1 WO 2019015645 A1 WO2019015645 A1 WO 2019015645A1 CN 2018096278 W CN2018096278 W CN 2018096278W WO 2019015645 A1 WO2019015645 A1 WO 2019015645A1
Authority
WO
WIPO (PCT)
Prior art keywords
target object
image
similarity
feature
images
Prior art date
Application number
PCT/CN2018/096278
Other languages
English (en)
French (fr)
Inventor
江南
郭明宇
Original Assignee
阿里巴巴集团控股有限公司
江南
郭明宇
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 阿里巴巴集团控股有限公司, 江南, 郭明宇 filed Critical 阿里巴巴集团控股有限公司
Priority to SG11201907444YA priority Critical patent/SG11201907444YA/en
Priority to EP18835563.0A priority patent/EP3579146A4/en
Priority to JP2019545984A priority patent/JP6945639B2/ja
Priority to KR1020197025770A priority patent/KR102316230B1/ko
Priority to MYPI2019004655A priority patent/MY201739A/en
Publication of WO2019015645A1 publication Critical patent/WO2019015645A1/zh
Priority to PH12019501920A priority patent/PH12019501920A1/en
Priority to US16/577,191 priority patent/US11093792B2/en
Priority to US16/777,696 priority patent/US10769490B2/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G06V40/172—Classification, e.g. identification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/40—Software arrangements specially adapted for pattern recognition, e.g. user interfaces or toolboxes therefor
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
    • G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
    • G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching

Definitions

  • the present application relates to the field of image processing, and in particular, to an image processing method and apparatus.
  • image processing has been applied in various fields, such as the face payment field and the identity recognition field.
  • image processing it is usually necessary to use the image reserved in the system as a processing basis, and processing the image reserved in the system to obtain a processing result, for example, the base map reserved in the system and the collected user image. Compare to verify the user's identity.
  • An object of the embodiments of the present invention is to provide an image processing method and apparatus, and based on standard features of a target object, determine an image with a high degree of similarity to a standard feature of the target object in multiple images of the target object, thereby The selected images are suitable for image processing to improve image processing.
  • An embodiment of the present application provides an image processing method, including:
  • An embodiment of the present application provides another image processing method, including:
  • the trusted image of the first target object is an image determined in a plurality of images of the first target object, and a feature of the trusted image of the first target object and a standard of the first target object The similarity between the features satisfies a first preset similarity requirement;
  • the trusted image of the second target object is an image determined in a plurality of images of the second target object, and the second target object may be The similarity between the feature of the letter image and the standard feature of the second target object satisfies a second predetermined similarity requirement.
  • the embodiment of the present application provides another image processing method, including:
  • An embodiment of the present application provides an image processing apparatus, including:
  • a feature acquisition module configured to acquire features of the plurality of images of the target object and standard features of the target object
  • An image filtering module configured to determine, in the plurality of images, a trusted image of the target object according to a similarity between a feature of the plurality of images and the standard feature; wherein the trusted image is The similarity between the feature and the standard feature satisfies the preset similarity requirement.
  • An embodiment of the present application provides another image processing apparatus, including:
  • An image acquisition module configured to acquire a trusted image of the first target object and a trusted image of the second target object
  • An image comparison module configured to determine, according to the trusted image of the first target object and the trusted image of the second target object, whether the first target object and the second target object are similar;
  • the trusted image of the first target object is an image determined in a plurality of images of the first target object, and a feature of the trusted image of the first target object and a standard of the first target object The similarity between the features satisfies a first preset similarity requirement;
  • the trusted image of the second target object is an image determined in a plurality of images of the second target object, and the second target object may be The similarity between the feature of the letter image and the standard feature of the second target object satisfies a second predetermined similarity requirement.
  • the embodiment of the present application provides another image processing apparatus, including:
  • a data acquisition module configured to acquire features of the plurality of images of the first target object, standard features of the first target object, features of the plurality of images of the second target object, and standard features of the second target object;
  • An image determining module configured to determine, according to a similarity between a feature of the plurality of images of the first target object and a standard feature of the first target object, in the plurality of images of the first target object a trusted image of the first target object, and a plurality of features of the second target object according to a similarity between a feature of the plurality of images of the second target object and a standard feature of the second target object Determining a trusted image of the second target object in the image; wherein a similarity between a feature of the trusted image of the first target object and its standard feature satisfies a first preset similarity requirement, the second target The similarity between the feature of the trusted image of the object and its standard feature satisfies the second preset similarity requirement;
  • the image determining module is configured to determine whether the first target object and the second target object are similar according to the trusted image of the first target object and the trusted image of the second target object.
  • An embodiment of the present application provides an image processing apparatus, including:
  • a memory arranged to store computer executable instructions that, when executed, cause the processor to:
  • An embodiment of the present application provides another image processing apparatus, including:
  • a memory arranged to store computer executable instructions that, when executed, cause the processor to:
  • the trusted image of the first target object is an image determined in a plurality of images of the first target object, and a feature of the trusted image of the first target object and a standard of the first target object The similarity between the features satisfies a first preset similarity requirement;
  • the trusted image of the second target object is an image determined in a plurality of images of the second target object, and the second target object may be The similarity between the feature of the letter image and the standard feature of the second target object satisfies a second predetermined similarity requirement.
  • the embodiment of the present application provides another image processing apparatus, including:
  • a memory arranged to store computer executable instructions that, when executed, cause the processor to:
  • the embodiment of the present application provides a storage medium for storing computer executable instructions, and the executable instructions implement the following processes when executed:
  • An embodiment of the present application provides another storage medium for storing computer executable instructions that, when executed, implement the following processes:
  • the trusted image of the first target object is an image determined in a plurality of images of the first target object, and a feature of the trusted image of the first target object and a standard of the first target object The similarity between the features satisfies a first preset similarity requirement;
  • the trusted image of the second target object is an image determined in a plurality of images of the second target object, and the second target object may be The similarity between the feature of the letter image and the standard feature of the second target object satisfies a second predetermined similarity requirement.
  • the embodiment of the present application provides another storage medium for storing computer executable instructions, where the executable instructions implement the following processes when executed:
  • the technical solution in the embodiment it is possible to determine, in the plurality of images of the target object, a high degree of similarity with the standard features of the target object and reflect the standard features of the target object based on the standard features of the target object.
  • the image so that the selected trusted image is suitable for image processing, improving the image processing effect.
  • FIG. 1 is a schematic diagram of a first process of an image processing method according to an embodiment of the present disclosure
  • 2a is a schematic diagram of determining a trusted image of a target object according to a standard image of a target object according to an embodiment of the present invention
  • 2b is a schematic diagram showing a distribution of similarities between features and standard features of multiple images according to an embodiment of the present invention
  • FIG. 3 is a schematic diagram of a second process of an image processing method according to an embodiment of the present disclosure
  • FIG. 4 is a schematic diagram of comparing whether a first target object and a second target object are similar according to a trusted image according to an embodiment of the present invention
  • FIG. 5 is a schematic diagram of a third process of an image processing method according to an embodiment of the present disclosure.
  • FIG. 6 is a fourth schematic flowchart of an image processing method according to an embodiment of the present disclosure.
  • FIG. 7 is a schematic diagram of a fifth process of an image processing method according to an embodiment of the present disclosure.
  • FIG. 8 is a schematic structural diagram of a first module of an image processing apparatus according to an embodiment of the present disclosure.
  • FIG. 9 is a schematic diagram of a second module composition of an image processing apparatus according to an embodiment of the present disclosure.
  • FIG. 10 is a schematic diagram of a third module composition of an image processing apparatus according to an embodiment of the present disclosure.
  • FIG. 11 is a schematic structural diagram of an image processing apparatus according to an embodiment of the present application.
  • the embodiment of the present application provides an image processing method and apparatus, which can filter out a trusted image of a target object in multiple images of a target object, and can also determine whether two target objects are similar based on the trusted image of the target object, wherein The selected target object's trusted image is highly similar to the standard feature of the target object and can reflect the standard features of the target object.
  • FIG. 1 is a schematic diagram of a first process of an image processing method according to an embodiment of the present disclosure.
  • the execution body of the method is a server. As shown in FIG. 1 , the method includes at least the following steps:
  • Step S102 acquiring features of the plurality of images of the target object and standard features of the target object;
  • Step S104 Determine a trusted image of the target object in the plurality of images according to the similarity between the feature of the plurality of images and the standard feature; wherein the similarity between the feature of the trusted image and the standard feature satisfies Set similarity requirements.
  • the image processing method in this embodiment first acquires the features of the plurality of images of the target object and the standard features of the target object, and then determines the target in the plurality of images according to the similarity between the features of the plurality of images and the standard features.
  • the trusted image of the object the similarity between the feature of the trusted image and the standard feature satisfies the preset similarity requirement. It can be seen that the image processing method in this embodiment can determine, according to the standard features of the target object, that the standard features of the target object have high similarity and can reflect the standard features of the target object in multiple images of the target object.
  • the trusted image makes the selected trusted image suitable for image processing and improves the image processing effect.
  • the target object may be a natural person or an item.
  • a plurality of images of the pre-stored target object may be read from the local database or the remote database, and features of the plurality of images of the target object are acquired, wherein each image of the target object has corresponding features. .
  • the standard feature of the target object refers to the ability to accurately reflect the characteristics of the target object, and the use of features that accurately reflect the target object to filter the trusted image in multiple images of the target object, which can ensure that the trusted image also accurately reflects the target object.
  • the standard features of the target object are obtained by the following methods (a1) or (a2):
  • the specific averaging manner can be determined according to the actual implementation scenario, and is not limited herein.
  • the obtained image may be obtained.
  • the eigenvectors of the same dimension of all images are averaged to obtain the averaging characteristics of all images.
  • the averaged features of all acquired images are taken as standard features of the target object.
  • the averaging characteristics of the plurality of images are first acquired, and the process is the same as the method (a1), and is not described herein. Since each image of the target object has corresponding features, among the features of the plurality of images, the feature with the highest similarity to the averaged feature is determined, and the feature with the highest similarity is determined as the standard feature of the target object. Specifically, in the plurality of images of the target object, the similarity between the feature of each image and the averaging feature is respectively calculated, and the image with the highest similarity is obtained, and the feature of the image with the highest similarity is used as the standard of the target object. feature. In this way, an image with the highest similarity to the averaging feature can be obtained.
  • the image can be called a standard image, and the standard image reflects the standard features of the target object, which can be used for other image processing of the target object. In the process.
  • the method (a1) is simple in determining the standard features and convenient in operation, and the method (a2) can obtain the standard with the highest similarity with the standard features in the process of determining the standard features.
  • the image is to facilitate other image processing processes of the target object, and one of ordinary skill in the art may select one of mode (a1) and mode (a2) to determine the standard feature of the target object according to actual conditions.
  • the trusted image of the target object may be determined in multiple images according to the similarity between the features of the plurality of images and the standard features in the following manner (b1) or (b2):
  • the similarity between the feature of each of the plurality of images of the target object and the standard features of the target object is respectively calculated, and then the plurality of images whose similarities are greater than the preset similarity threshold are determined.
  • the plurality of images are very close to the standard features of the target object, and can reflect the standard features of the target object.
  • the calculation speed will be affected, so in this method, Regardless of the calculation speed, all the images in the determined plurality of images can be used as the trusted image of the target object. If the calculation speed is considered, part of the determined plurality of images can be used as the target object. Trusted image.
  • the preset similarity threshold is a preset value of the server, and the preset manner may be set by the server according to the implementation scenario of the embodiment.
  • a specific example is to calculate the similarity between the features of each image in the plurality of images of the target object and the standard features of the target object, and to image multiple images of the target object according to the highest to lowest order of similarity. Sorting is performed, and in the sorting, an image whose similarity is greater than a preset similarity threshold is determined. Considering the requirement of the calculation speed, if the number of images whose similarity is greater than the preset similarity threshold is greater than or equal to a certain value, such as a certain value of 100, the similarity is greater than the preset similarity threshold according to the highest to lowest order of similarity.
  • the image select the first half of the certain number of images (here, 50 images), as the trusted image of the target object, if the number of images whose similarity is greater than the preset similarity threshold is less than a certain value, then the similarity is the highest. To the lowest order, in the image whose similarity is greater than the preset similarity threshold, the first half of the image is selected as the trusted image of the target object.
  • the standard feature of the target object is a feature of the standard image of the target object.
  • FIG. 2a is a schematic diagram of determining a trusted image of the target object according to the standard image of the target object according to an embodiment of the present invention, as shown in FIG. 2a.
  • the image a is a standard image of the target object
  • the images a1, a2, a3, and a4 are partial images of the acquired plurality of images of the target object, respectively, and the standard image a and the images a1, a2, a3, and a4 are respectively calculated.
  • the similarity between the standard image a and the images a1, a2, a3, and a4 is 85%, 65%, 87%, and 86%, respectively, and the preset similarity threshold is 80%.
  • the trusted images of the target object are determined to be a1, a3, and a4, and a2 is an untrusted image.
  • the similarity between the feature of each image and the standard feature of the target object is calculated, and a plurality of similarity data are obtained, and the distribution of the plurality of similarity data is counted.
  • the distribution data is obtained, and the distribution data may be embodied in the form of a statistical distribution histogram, wherein the distribution data includes a plurality of similarity intervals, and marks the number of images corresponding to each similarity interval.
  • a similarity interval in which the image density is greater than the preset density is determined, where the image density may be represented by a ratio of the number of images in the similarity interval to the similarity range of the similarity interval, and the larger the image density, the The more the number of images corresponding to each similarity unit in the similarity interval, the preset density is the value preset by the server, and the server can set according to the scene needs.
  • the similarity interval of the image density greater than the preset density similarly to the method (b1), if the calculation speed is not considered, all the images in the determined similarity interval are determined as the trusted image of the target object. If the calculation speed is considered, the partial image within the determined similarity interval is determined as the trusted image of the target object.
  • the image within the similarity interval in which the image density is greater than the preset density is not necessarily the image most similar to the standard feature.
  • the image distributed within the similarity interval of the image density greater than the preset density is mainly a normal image of the target object, and if the target object is a natural person, when the photograph is used to laughing, the distribution of the target object is statistically distributed.
  • an image distributed within a similarity interval in which the image density is greater than the preset density is mainly a laughing image of the target object.
  • FIG. 2b is a schematic diagram showing the distribution of similarity between features and standard features of multiple images according to an embodiment of the present invention, as shown in FIG. 2b, according to the feature of each image of the target object and the standard features of the target object.
  • the distribution data of the similarity obtains a statistical distribution histogram, which includes a plurality of similarity intervals, and marks the number of images corresponding to each similarity interval.
  • the similarity interval in which the image density is greater than the preset density is an interval in which the similarity score is between 60 and 80, and the interval in which the similarity score is between 80 and 100, which is indicated by a diagonal line in the figure, Part or all of the images corresponding to the two intervals are determined as trusted images of the target object.
  • the above method (b1) mainly determines the trusted image of the target object in multiple images of the target object based on the preset similarity threshold value, and the above manner (b2) mainly Based on the similarity distribution data, a trusted image of the target object is determined in a plurality of images of the target object.
  • the above method (b1) can easily and quickly obtain a trusted image of the target object that is very close to the standard feature of the target object, when the target object is large.
  • the above method (b2) can avoid the interference of the impurity image, and the normal image of the target object is used as the trusted image of the target object.
  • One of ordinary skill in the art can select one of mode (b1) and mode (b2) to determine a trusted image of the target object according to actual conditions.
  • the trusted image when the trusted image is determined based on the similarity distribution manner, if an isolated high-similar image is determined in the similarity distribution data, if only one image and the standard feature have a similarity of 98 points The maximum similarity between the remaining images and the standard features is 90 points.
  • the isolated high-similar image may completely coincide with the standard features, or may be an image with errors. In this case, to ensure image selection accuracy, The image can be excluded, and when it is hot, it can be excluded, depending on the specific scene.
  • the trusted image of the target object is determined in multiple images of the target object, and the low-quality image that is not similar to the standard feature of the target object can be removed from the multiple images of the target object, thereby obtaining high quality.
  • FIG. 3 is a second schematic flowchart of an image processing method according to an embodiment of the present disclosure.
  • the method is preferably applied to a scenario where the target object is a natural person, as shown in FIG. 3, and the method further includes the following steps on the basis of FIG. 1 :
  • Step S106 determining a similarity between the trusted image of the first target object and the trusted image of the second target object, to obtain a plurality of similarity data
  • Step S108 Determine, according to the plurality of similarity data, whether the first target object is similar to the second target object.
  • the target object in step S102 may include a plurality of target objects, which are a first target object, a second target object, a third target object, and the like, respectively. Therefore, by step S102 and step S104, the trusted image of the first target object and the trusted image of the second target object can be determined separately.
  • step S106 the specific manner of determining the similarity between the trusted image of the first target object and the trusted image of the second target object is to determine each trusted image of the first target object and each of the second target objects.
  • the similarity between the images of the trusted images, the similarity calculation method is not specifically limited here, and can be selected as needed.
  • the calculated number of similarity data is equal to the number of trusted images of the first target object and the second The product of the number of trusted images of the target object.
  • step S108 whether the first target object and the second target object are similar according to the plurality of similarity data may be determined by the following manner (c1) or (c2):
  • calculating an average value of the plurality of similarity data if the average value of the plurality of similarity data is greater than a preset average threshold, determining that the first target object is similar to the second target object; otherwise, determining A target object is not similar to the second target object.
  • the preset average threshold is a preset value of the server, and the preset manner may be set by the server according to the implementation scenario of the embodiment.
  • a distribution of a plurality of similarity data is calculated, and the distribution may be represented by a statistical distribution histogram, wherein the distribution includes a plurality of similarity intervals, and marks each similarity interval corresponding to The number of image pairs.
  • the preset similarity distribution requirement may be that the ratio of the number of image pairs whose similarity is greater than a certain value to the total number of image comparison times is greater than a preset ratio, and in the above distribution, if the similarity is greater than a certain value It is determined that the first target object is similar to the second target object, and the first target object is not similar to the second target object. If the ratio of the number of image pairs with the similarity greater than 75 points to the total number of image comparisons is greater than 80%, it is determined that the first target object is similar to the second target object; otherwise, determining that the first target object is not similar to the second target object .
  • the number of image pairs whose similarity is greater than a certain value is determined, the total number of image comparisons is equal to the number of trusted images of the first target object, and the second The product of the number of trusted images of the target object.
  • the preset similarity distribution requirement may be that the similarity range of the similarity interval with the largest number of corresponding images is within a preset similarity range, and in the foregoing distribution, if the corresponding number of images is the largest The similarity range of the similarity interval is within the preset similarity range, and then the first target object is determined to be similar to the second target object; otherwise, the first target object is determined to be dissimilar to the second target object. If the similarity interval of the corresponding number of corresponding images has a similarity range of 80 minutes to 85 minutes, and the preset similarity range is between 70 minutes and 90 minutes, it is determined that the first target object is similar to the second target object; otherwise, It is determined that the first target object is not similar to the second target object.
  • the mode (c1) can easily and quickly determine whether the first target object and the second target object are similar by using the preset average threshold value, and the mode (c2) can be based on the similarity data.
  • the distribution of the first target object is similar to the second target object.
  • one of the mode (c1) and the mode (c2) may be determined according to the implementation scenario to determine the first target object and Whether the second target object is similar.
  • step S108 determining whether the first target object and the second target object are similar according to the plurality of similarity data, specifically: determining whether the number of the plurality of similarity data meets the preset quantity requirement And if yes, determining whether the first target object is similar to the second target object according to the plurality of similarity data.
  • the accuracy of the plurality of similarity data obtained in step S106 may be uncertain. Satisfying the preset quantity requirement, determining whether the first target object and the second target object are similar according to the similarity data. If the algorithm determines whether the first target object and the second target object are similar according to the similarity data, the algorithm has lower accuracy. It is necessary to determine whether the number of the plurality of similarity data obtained in step S106 satisfies the preset quantity requirement, and if yes, determine whether the first target object and the second target object are similar according to the similarity data, and if not, the ending method Process.
  • FIG. 4 is a schematic diagram of comparing whether a first target object and a second target object are similar according to a trusted image according to an embodiment of the present invention.
  • the first target object includes p1, p2, p3, .
  • the picture illustrated by three pictures in the figure
  • the second target object includes q1, q2, q3...qn n trusted pictures (three pictures are taken as an example in the figure), and each of the first target objects is calculated separately.
  • the similarity between each trusted picture and each trusted picture of the second target object obtains multiple similarity scores, the partial similarity score is marked in the figure, and the average score of all similarity scores is counted as 85 points. Greater than the preset average threshold, determining that the first target object is similar to the second target object.
  • FIG. 5 is an image processing method provided by an embodiment of the present application.
  • the third process diagram is shown in Figure 5. The method includes:
  • step S302 an image is acquired. Acquiring multiple images of the first target object and multiple images of the second target object.
  • Step S304 determining a first averaging feature and a second averaging feature.
  • An averaging feature of the plurality of images of the first target object is determined as a first averaging feature
  • an averaging feature of the plurality of images of the second target object is determined as a second averaging feature.
  • Step S306 calculating a first similarity and a second similarity.
  • the plurality of images of the first target object respectively calculating a first similarity between the feature of each image of the first target object and the first averaging feature, and calculating the first in the plurality of images of the second target object A second similarity between the feature of each image of the two target objects and the second averaged feature.
  • Step S308 determining a first standard image and a second standard image.
  • the image having the largest first degree of similarity is used as the first standard image of the first target object, and the image having the second highest degree of similarity is used as the second standard image of the second target object.
  • Step S310 calculating a third similarity and a fourth similarity.
  • Step S310 calculating a third similarity between each image of the first target object and the first standard image, and calculating a second in the plurality of images of the second target object respectively A fourth similarity between each image of the target object and the second standard image.
  • Step S312 determining a first trusted image and a second trusted image.
  • the image with the third similarity greater than the certain value is used as the first trusted image of the first target object, and the image with the fourth similarity greater than the certain value is used as the second trusted image of the second target object.
  • step S314 the image number product is calculated. A product of the number of first trusted images and the number of second trusted images is calculated.
  • step S316 the similarity average value is calculated. Calculating the similarity between each first trusted image and each second trusted image separately, obtaining a plurality of similarity data, and calculating an average of the plurality of similarity data.
  • step S318 it is determined whether the product calculated in step S314 is greater than a preset product. If it is greater, step S320 is performed; otherwise, the comparison is ended.
  • step S320 it is determined whether the average value calculated in step S316 is greater than a preset average value. If it is greater, step S322 is performed; otherwise, the comparison is ended.
  • Step S322 determining that the first target object is similar to the second target object.
  • the similarity calculation method when calculating the similarity, such as calculating the similarity between each image and the averaging feature, calculating the similarity between each image and the standard image, or calculating the similarity
  • the similarity calculation method is not limited, and the Euclidean distance or the cosine distance between the feature vectors may be calculated to calculate the similarity in the target object.
  • feature face-based comparison methods and the like For natural persons, it is also possible to use feature face-based comparison methods and the like, and only indicators are needed to quantify the degree of similarity between features.
  • a principal component analysis (PCA), a local feature analysis (Local Face Analysis), a neural network method (Neural Networks), or the like may be used.
  • the target object may be a natural person.
  • the low-quality image of the natural person can be excluded from the base map of the natural person reserved in the system, and the trusted image of the natural person can be selected.
  • the image of the letter can be an image with suitable light, a correct face, a clear image, no heavy makeup and exaggerated accessories, and can reflect the facial features of natural persons through a trusted image; the method in this embodiment can also be based on the trust of two natural persons. The image determines whether two natural persons are similar.
  • the method in the embodiment is simple in process and accurate in effect, and can be widely applied in the field of face recognition, especially in the fields of identical twin screening, target person positioning, etc., and has potential social value, such as recognizing that both sides have faces. Collecting identical twins in history, looking for fugitives who are being trafficked or chasing surnames.
  • FIG. 6 is a fourth schematic flowchart of an image processing method according to an embodiment of the present application.
  • the main body is a server.
  • FIG. 6 the difference between the method and the foregoing methods in FIG. 1 to FIG. 5 is mainly introduced.
  • the method includes at least the following steps:
  • Step S402 acquiring a trusted image of the first target object and a trusted image of the second target object;
  • Step S404 determining whether the first target object and the second target object are similar according to the trusted image of the first target object and the trusted image of the second target object;
  • the trusted image of the first target object is an image determined in multiple images of the first target object, and the similarity between the feature of the trusted image of the first target object and the standard feature of the first target object satisfies a predetermined similarity requirement;
  • the trusted image of the second target object is an image determined in a plurality of images of the second target object, and the feature of the trusted image of the second target object is compared with a standard feature of the second target object The similarity satisfies the second preset similarity requirement.
  • the image processing method of this embodiment since the trusted image of the first target object is filtered in the plurality of images of the first target object, the similarity between the feature and the standard feature of the first target object satisfies the first pre-preparation The image of the similarity requirement is set, and the trusted image of the second target object is filtered in the plurality of images of the second target object, and the similarity between the feature and the standard feature of the second target object satisfies the second preset similarity The image required by the image is processed. Therefore, the image processing method in the embodiment is based on the standard features of the target object, and the plurality of images of the target object are determined to have high similarity with the standard features of the target object and can reflect the target object.
  • the image of the standard feature so that the selected image is suitable for image processing, improving the image processing effect, and, since the method in this embodiment is a trusted image of the first target object and the second target object is trusted Based on the image, it is determined whether the first target object and the second target object are similar, so the image processing method in the embodiment has good image processing effect.
  • the judgment result is accurate, and it can accurately determine whether the first target object is similar to the second target object.
  • the standard feature of the first target object is an averaging feature of the plurality of images of the first target object; or the standard feature of the first target object is the first Among the features of the plurality of images of the target object, the features having the highest similarity with the averaged features of the plurality of images of the first target object; the standard features of the second target object are the averaged features of the plurality of images of the second target object; Alternatively, the standard feature of the second target object is a feature that has the highest similarity of the averaged features of the plurality of images of the second target object among the features of the plurality of images of the second target object.
  • the first preset similarity requirement and the second preset similarity requirement may be the same or different.
  • the trusted image of the first target object and the trusted image of the second target object are determined using a unified standard, it is possible to ensure the accuracy of determining whether the first target object and the second target object are similar, preferably, the first target object
  • the standard features and the standard features of the second target object are determined in the same way, both of which are averaged features of multiple images, or both are averaged feature similarities between multiple images and features of multiple images.
  • the highest feature, and the first preset similarity requirement is the same as the second preset similarity requirement.
  • the specific explanation about the first preset similarity requirement and the second preset similarity requirement is consistent with the description in FIG. 1 to FIG. 5, in the above step S402, acquiring the trusted image of the first target object and the second target object
  • the specific process of the trusted image can also refer to the description in FIG. 1 to FIG. 5, and details are not described herein.
  • step S404 it is determined whether the first target object and the second target object are similar according to the trusted image of the first target object and the trusted image of the second target object in the following manner (d1) or (d2):
  • the first target object is a database of the target object.
  • Any target object in the second target object is any target object other than the first target object in the database of the target object, so that two target objects with similar features are found in the database.
  • the target object that is similar to the specified target object is to be found in the database, such as another natural person who is similar to the designated natural person.
  • the first target object is specified by the user.
  • the target object, the second target object is any target object in the target object's database, so that the target object similar to the specified target object is found in the database.
  • FIG. 7 is a schematic flowchart of a fifth processing method of the image processing method provided by the embodiment of the present application.
  • the main body is a server.
  • FIG. 7 the difference between the method and the foregoing methods in FIG. 1 to FIG. 5 is mainly introduced.
  • the method includes at least the following steps:
  • Step S502 acquiring features of the plurality of images of the first target object, standard features of the first target object, features of the plurality of images of the second target object, and standard features of the second target object;
  • Step S504 determining a trusted image of the first target object in the plurality of images of the first target object according to the similarity between the feature of the plurality of images of the first target object and the standard feature of the first target object, and Determining a trusted image of the second target object in the plurality of images of the second target object according to a similarity between a feature of the plurality of images of the second target object and a standard feature of the second target object; wherein, the first target The similarity between the feature of the trusted image of the object and its standard feature satisfies the first preset similarity requirement, and the similarity between the feature of the trusted image of the second target object and its standard feature satisfies the second preset similarity requirement ;
  • Step S506 determining whether the first target object and the second target object are similar according to the trusted image of the first target object and the trusted image of the second target object.
  • the image processing method in this embodiment first acquires the features of the plurality of images of the target object and the standard features of the target object, and then determines the target in the plurality of images according to the similarity between the features of the plurality of images and the standard features.
  • the trusted image of the object, the similarity between the feature of the trusted image and the standard feature satisfies the preset similarity requirement, and finally determines whether the two target objects are similar according to the trusted image of the two target objects.
  • the image processing method in this embodiment determines, based on the standard features of the target object, a trusted image that is highly similar to the standard feature of the target object and can reflect the standard features of the target object in the plurality of images of the target object.
  • the selected trusted image is suitable for image processing, and the image processing effect is improved, and the first target object and the second target are determined based on the trusted image of the first target object and the trusted image of the second target object. Whether the objects are similar can improve the accuracy of the similarity judgment of the two target objects, so as to accurately determine whether the two target objects are similar.
  • step S502 and step S504 For the specific process of step S502 and step S504, reference may be made to the descriptions of step S102 and step S104 in FIG. 1 and FIG. 2, and the specific process of step S506 may refer to the description of step S404 in FIG. 6, which is not described here.
  • the first preset similarity requirement and the second preset similarity requirement are the same, or the first preset similarity requirement and the second preset similarity requirement are different, and the first standard is determined according to the unified standard.
  • the trusted image of the target object and the trusted image of the second target object can ensure the accuracy of determining whether the first target object and the second target object are similar, preferably, the standard features of the first target object and the second target object.
  • the standard features are determined in the same way, both of which are averaged features of multiple images, or both are features with the highest similarity to the averaged features of multiple images in the features of multiple images, and the first preset The similarity requirement is the same as the second preset similarity requirement.
  • the method shown in FIG. 1 to FIG. 7 in the embodiment can determine the trusted image of the target object, and determine whether the two target objects are similar based on the trusted image of the target object, the determination process is simple, and the determination effect is accurate. Can be widely used in face recognition, twin screening, target person positioning and other fields.
  • FIG. 8 is a schematic diagram of the first module of the image processing apparatus according to the embodiment of the present application, as shown in FIG. 8 .
  • the device includes:
  • a feature acquiring module 61 configured to acquire features of the plurality of images of the target object and standard features of the target object;
  • An image filtering module 62 configured to determine, in the plurality of images, a trusted image of the target object according to a similarity between a feature of the plurality of images and the standard feature; wherein the trusted image The similarity between the feature and the standard feature satisfies the preset similarity requirement.
  • the feature acquiring module 61 is specifically configured to: obtain an averaged feature of the multiple images, and use an averaged feature of the multiple images as a standard feature of the target object; or The averaging feature of the plurality of images, wherein the feature having the highest similarity to the averaging feature among the features of the plurality of images is used as a standard feature of the target object.
  • the image screening module 62 is specifically configured to: in the plurality of images, part or all of the image whose similarity between the feature and the standard feature is greater than a preset similarity threshold, as the a trusted image of the target object; or, determining distribution data of similarity between the feature of the plurality of images and the standard feature, determining a similarity interval in which the image density is greater than a preset density in the distribution data, A part or all of the image corresponding to the determined similarity interval is used as a trusted image of the target object.
  • the device further includes:
  • a similarity determination module configured to determine a similarity between the trusted image of the first target object and the trusted image of the second target object, to obtain a plurality of similarity data
  • the similarity determining module is configured to determine, according to the multiple similarity data, whether the first target object is similar to the second target object.
  • the similarity determining module is specifically configured to: if the average value of the plurality of similarity data is greater than a preset average threshold, determine that the first target object is similar to the second target object; or And if the distribution of the plurality of similarity data satisfies a preset similarity distribution requirement, determining that the first target object is similar to the second target object.
  • the similarity determining module is specifically configured to: determine whether the quantity of the plurality of similarity data meets a preset quantity requirement, and if yes, determine the first target according to the multiple similarity data. Whether the object is similar to the second target object.
  • the image processing apparatus in this embodiment first acquires the features of the plurality of images of the target object and the standard features of the target object, and then determines the target in the plurality of images according to the similarity between the features of the plurality of images and the standard features.
  • the trusted image of the object the similarity between the feature of the trusted image and the standard feature satisfies the preset similarity requirement. It can be seen that the image processing apparatus in this embodiment can determine, based on the standard features of the target object, that the standard features of the target object have high similarity and can reflect the standard features of the target object in multiple images of the target object.
  • the trusted image makes the selected trusted image suitable for image processing and improves the image processing effect.
  • FIG. 9 is a schematic diagram of a second module composition of the image processing apparatus provided by the embodiment of the present application, such as As shown in Figure 9, the device includes:
  • the image obtaining module 71 is configured to acquire a trusted image of the first target object and a trusted image of the second target object;
  • the image comparison module 72 is configured to determine, according to the trusted image of the first target object and the trusted image of the second target object, whether the first target object and the second target object are similar;
  • the trusted image of the first target object is an image determined in a plurality of images of the first target object, and a feature of the trusted image of the first target object and a standard of the first target object The similarity between the features satisfies a first preset similarity requirement;
  • the trusted image of the second target object is an image determined in a plurality of images of the second target object, and the second target object may be The similarity between the feature of the letter image and the standard feature of the second target object satisfies a second predetermined similarity requirement.
  • the standard feature of the first target object is an averaging feature of the plurality of images of the first target object; or the standard feature of the first target object is the first target object a feature of the plurality of images that has the highest similarity to the averaged features of the plurality of images of the first target object; the standard feature of the second target object is a plurality of images of the second target object Or averaging features; or the standard feature of the second target object is a feature that has the highest similarity between the averaged features of the plurality of images of the second target object among the features of the plurality of images of the second target object .
  • the image comparison module 72 is specifically configured to: if an average value of the plurality of similarity data between the trusted image of the first target object and the trusted image of the second target object is greater than a pre- Setting an average threshold, determining that the first target object is similar to the second target object; or, if multiple of the trusted image of the first target object and the trusted image of the second target object The distribution of the similarity data satisfies the preset similarity distribution requirement, and then the first target object is determined to be similar to the second target object.
  • the first target object is any target object in the database of the target object; the second target object is any target object in the database other than the first target object; or
  • the first target object is a user-specified target object, and the second target object is any target object in a database of the target object.
  • the image processing apparatus of this embodiment since the trusted image of the first target object is filtered in the plurality of images of the first target object, the similarity between the feature and the standard feature of the first target object satisfies the first pre-preparation
  • the image of the similarity requirement is set, and the trusted image of the second target object is filtered in the plurality of images of the second target object, and the similarity between the feature and the standard feature of the second target object satisfies the second preset similarity
  • the image required by the image processing apparatus in the present embodiment is realized based on the standard features of the target object, and the plurality of images of the target object are determined to have a high degree of similarity with the standard features of the target object, and can reflect the target object.
  • the image of the standard feature so that the selected image is suitable for image processing, improving the image processing effect, and, since the device in this embodiment is the trusted image of the first target object and the trustedness of the second target object Based on the image, it is determined whether the first target object and the second target object are similar, so the image processing apparatus in the embodiment has good image processing effect. Accurate determination result, it is possible to accurately determine the first target object and the second target object are similar.
  • FIG. 10 is a schematic diagram of a third module composition of the image processing apparatus according to the embodiment of the present application, such as As shown in Figure 10, the device includes:
  • a data acquisition module 81 configured to acquire features of the plurality of images of the first target object, standard features of the first target object, features of the plurality of images of the second target object, and standard features of the second target object ;
  • the image determining module 82 is configured to determine, according to the similarity between the feature of the plurality of images of the first target object and the standard feature of the first target object, among the multiple images of the first target object a trusted image of the first target object, and a similarity between the feature of the plurality of images of the second target object and the standard feature of the second target object, Determining a trusted image of the second target object in the image; wherein a similarity between a feature of the trusted image of the first target object and its standard feature satisfies a first preset similarity requirement, the second The similarity between the feature of the trusted image of the target object and its standard feature satisfies the second preset similarity requirement;
  • the image determining module 83 is configured to determine whether the first target object and the second target object are similar according to the trusted image of the first target object and the trusted image of the second target object.
  • the first preset similarity requirement and the second preset similarity requirement are the same, or the first preset similarity requirement and the second preset similarity requirement are different.
  • the image processing apparatus in this embodiment first acquires the features of the plurality of images of the target object and the standard features of the target object, and then determines the target in the plurality of images according to the similarity between the features of the plurality of images and the standard features.
  • the trusted image of the object, the similarity between the feature of the trusted image and the standard feature satisfies the preset similarity requirement, and finally determines whether the two target objects are similar according to the trusted image of the two target objects.
  • the image processing apparatus in this embodiment determines, based on the standard features of the target object, a trusted image that is highly similar to the standard feature of the target object and can reflect the standard features of the target object in the plurality of images of the target object.
  • the selected trusted image is suitable for image processing, and the image processing effect is improved, and the first target object and the second target are determined based on the trusted image of the first target object and the trusted image of the second target object. Whether the objects are similar can improve the accuracy of the similarity judgment of the two target objects, so as to accurately determine whether the two target objects are similar.
  • an embodiment of the present application further provides an image processing apparatus, as shown in FIG. 11 .
  • the image processing device may vary considerably depending on configuration or performance, and may include one or more processors 901 and memory 902 in which one or more storage applications or data may be stored.
  • the memory 902 can be short-term storage or persistent storage.
  • An application stored in memory 902 can include one or more modules (not shown), each of which can include a series of computer-executable instructions in an image processing device.
  • the processor 901 can be arranged to communicate with the memory 902 to execute a series of computer executable instructions in the memory 902 on the image processing device.
  • the image processing device may also include one or more power sources 903, one or more wired or wireless network interfaces 904, one or more input and output interfaces 905, one or more keyboards 906, and the like.
  • an image processing device includes a memory, and one or more programs, wherein one or more programs are stored in the memory, and one or more programs can include one or more modules, and each The modules can include a series of computer executable instructions in an image processing device, and are configured to be executed by one or more processors.
  • the one or more programs are included for performing the following computer executable instructions:
  • the computer executable instructions when executed, acquire standard features of the target object, including: acquiring an averaging feature of the plurality of images, and using the averaged features of the multiple images as the target object Or the averaging feature of the plurality of images, and the feature having the highest similarity to the averaging feature among the features of the plurality of images is used as a standard feature of the target object.
  • determining a trusted image of the target object in the plurality of images according to a similarity between a feature of the plurality of images and the standard feature Included, in the plurality of images, a part or all of the image having a similarity between the feature and the standard feature greater than a preset similarity threshold as a trusted image of the target object; or determining the a distribution data of a similarity between a feature of the plurality of images and the standard feature, wherein a similarity interval in which the image density is greater than a preset density is determined in the distribution data, and the determined portion corresponding to the similarity interval or All images as a trusted image of the target object.
  • the computer executable instructions when executed, may further cause the processor to determine a similarity between the trusted image of the first target object and the trusted image of the second target object, to obtain multiple similarities. Data; determining, according to the plurality of similarity data, whether the first target object is similar to the second target object.
  • determining, according to the plurality of similarity data, whether the first target object is similar to the second target object including: if the multiple similarities Determining that the first target object is similar to the second target object; or, if the distribution of the plurality of similarity data meets a preset similarity distribution requirement, determining that the average value of the data is greater than a preset average threshold The first target object is similar to the second target object.
  • determining, according to the plurality of similarity data, whether the first target object is similar to the second target object comprises: determining the multiple similarities Whether the number of data satisfies a preset number requirement, and if so, whether the first target object and the second target object are similar according to the plurality of similarity data.
  • the image processing apparatus in this embodiment first acquires the features of the plurality of images of the target object and the standard features of the target object, and then determines the target in the plurality of images according to the similarity between the features of the plurality of images and the standard features.
  • the trusted image of the object the similarity between the feature of the trusted image and the standard feature satisfies the preset similarity requirement. It can be seen that the image processing apparatus in this embodiment can determine, based on the standard features of the target object, that the standard features of the target object have high similarity and can reflect the standard features of the target object in multiple images of the target object.
  • the trusted image makes the selected trusted image suitable for image processing and improves the image processing effect.
  • an image processing apparatus includes a memory, and one or more programs, wherein one or more programs are stored in the memory, and one or more programs can include one or more modules, and Each module can include a series of computer executable instructions in an image processing device and configured to be executed by one or more processors.
  • the one or more programs are included for performing the following computer executable instructions:
  • the trusted image of the first target object is an image determined in a plurality of images of the first target object, and a feature of the trusted image of the first target object and a standard of the first target object The similarity between the features satisfies a first preset similarity requirement;
  • the trusted image of the second target object is an image determined in a plurality of images of the second target object, and the second target object may be The similarity between the feature of the letter image and the standard feature of the second target object satisfies a second predetermined similarity requirement.
  • the standard feature of the first target object is an averaging feature of the plurality of images of the first target object; or the standard feature of the first target object is the first target object a feature of the plurality of images having the highest similarity to the averaged features of the plurality of images of the first target object; the standard feature of the second target object is an average of the plurality of images of the second target object Or a standard feature of the second target object, wherein the feature of the plurality of images of the second target object has the highest similarity to the averaged feature of the plurality of images of the second target object.
  • determining the first target object and the second according to the trusted image of the first target object and the trusted image of the second target object Whether the target object is similar including: determining, if an average value of the plurality of similarity data between the trusted image of the first target object and the trusted image of the second target object is greater than a preset average threshold The first target object is similar to the second target object; or, if the distribution of the plurality of similarity data between the trusted image of the first target object and the trusted image of the second target object satisfies a preset
  • the similarity distribution requires determining that the first target object is similar to the second target object.
  • the first target object is any target object in a database of the target object; the second target object is any target object in the database other than the first target object; or
  • the first target object is a target object specified by the user, and the second target object is any target object in the database of the target object.
  • the image processing apparatus of this embodiment since the trusted image of the first target object is filtered in the plurality of images of the first target object, the similarity between the feature and the standard feature of the first target object satisfies the first pre-preparation
  • the image of the similarity requirement is set, and the trusted image of the second target object is filtered in the plurality of images of the second target object, and the similarity between the feature and the standard feature of the second target object satisfies the second preset similarity
  • the image required by the image processing apparatus in the embodiment is realized based on the standard features of the target object, and the plurality of images of the target object are determined to have high similarity with the standard features of the target object, and can reflect the target object.
  • the image of the standard feature so that the selected image is suitable for image processing, improving the image processing effect, and, since the device in this embodiment is the trusted image of the first target object and the trustedness of the second target object Based on the image, it is determined whether the first target object and the second target object are similar, so the image processing device in the embodiment has good image processing effect.
  • the judgment result is accurate, and it can accurately determine whether the first target object is similar to the second target object.
  • an image processing apparatus includes a memory, and one or more programs, wherein one or more programs are stored in the memory, and one or more programs can include one or more modules, and Each module can include a series of computer executable instructions in an image processing device and configured to be executed by one or more processors.
  • the one or more programs are included for performing the following computer executable instructions:
  • the first preset similarity requirement and the second preset similarity requirement are the same, or the first preset similarity requirement and the second preset similarity requirement are different.
  • the image processing apparatus in this embodiment first acquires the features of the plurality of images of the target object and the standard features of the target object, and then determines the target in the plurality of images according to the similarity between the features of the plurality of images and the standard features.
  • the trusted image of the object, the similarity between the feature of the trusted image and the standard feature satisfies the preset similarity requirement, and finally determines whether the two target objects are similar according to the trusted image of the two target objects.
  • the image processing apparatus in this embodiment determines, based on the standard features of the target object, a trusted image that is highly similar to the standard feature of the target object and can reflect the standard features of the target object in the plurality of images of the target object.
  • the selected trusted image is suitable for image processing, and the image processing effect is improved, and the first target object and the second target are determined based on the trusted image of the first target object and the trusted image of the second target object. Whether the objects are similar can improve the accuracy of the similarity judgment of the two target objects, so as to accurately determine whether the two target objects are similar.
  • the embodiment of the present application further provides a storage medium for storing computer executable instructions.
  • the storage medium may be a USB flash drive.
  • the optical disk, the hard disk, and the like, the computer executable instructions stored by the storage medium, when executed by the processor, can implement the following processes:
  • acquiring the standard features of the target object includes: obtaining an averaging feature of the plurality of images, and using an averaged feature of the multiple images as a standard feature of the target object; or An averaging feature of the plurality of images is used, and a feature having the highest similarity to the averaged feature among the features of the plurality of images is used as a standard feature of the target object.
  • determining, according to the similarity between the feature of the plurality of images and the standard feature, determining a trusted image of the target object in the multiple images including: In the image, a part or all of the image having a similarity between the feature and the standard feature greater than a preset similarity threshold as a trusted image of the target object; or determining a feature of the plurality of images and the The distribution data of the similarity between the standard features, the similarity interval in which the image density is greater than the preset density is determined in the distribution data, and some or all of the images corresponding to the determined similarity interval are used as the target object Trusted image.
  • the process in this embodiment further includes: determining a similarity between the trusted image of the first target object and the trusted image of the second target object, obtaining a plurality of similarity data; determining, according to the multiple similarity data, Whether the first target object is similar to the second target object.
  • the determining, according to the multiple similarity data, whether the first target object is similar to the second target object includes: if an average value of the multiple similarity data is greater than a preset average a threshold, the first target object is determined to be similar to the second target object; or, if the distribution of the plurality of similarity data meets a preset similarity distribution requirement, determining the first target object and the The second target object is similar.
  • determining whether the first target object is similar to the second target object according to the multiple similarity data includes: determining whether the quantity of the multiple similarity data meets a preset quantity If yes, determining whether the first target object and the second target object are similar according to the plurality of similarity data.
  • the executable instructions in the storage medium of the embodiment When the executable instructions in the storage medium of the embodiment are executed, first acquiring the features of the plurality of images of the target object and the standard features of the target object, and then according to the similarity between the features of the plurality of images and the standard features, A trusted image of the target object is determined in the plurality of images, and the similarity between the feature of the trusted image and the standard feature satisfies a preset similarity requirement. It can be seen that, when executed, the executable instructions in the storage medium of the embodiment can determine, according to the standard features of the target object, that the standard features of the target object have high similarity and can be determined in multiple images of the target object. A trusted image reflecting the standard features of the target object, so that the selected trusted images are suitable for image processing, and the image processing effect is improved.
  • the storage medium may be a USB flash drive, an optical disk, a hard disk, or the like, and the computer executable instructions stored by the storage medium, when executed by the processor, can implement the following processes:
  • the trusted image of the first target object is an image determined in a plurality of images of the first target object, and a feature of the trusted image of the first target object and a standard of the first target object The similarity between the features satisfies a first preset similarity requirement;
  • the trusted image of the second target object is an image determined in a plurality of images of the second target object, and the second target object may be The similarity between the feature of the letter image and the standard feature of the second target object satisfies a second predetermined similarity requirement.
  • the standard feature of the first target object is an averaged feature of the plurality of images of the first target object; or the standard feature of the first target object is a plurality of the first target object a feature of the image that has the highest similarity to the averaged feature of the plurality of images of the first target object; and a standard feature of the second target object that is an averaged feature of the plurality of images of the second target object Or the standard feature of the second target object is a feature that has the highest degree of similarity between the features of the plurality of images of the second target object and the averaged features of the plurality of images of the second target object.
  • the determining, according to the trusted image of the first target object and the trusted image of the second target object, whether the first target object is similar to the second target object including: Determining the first target object and the second target object, wherein an average of the plurality of similarity data between the trusted image of the target object and the trusted image of the second target object is greater than a preset average threshold Similarly; or, if the distribution of the plurality of similarity data between the trusted image of the first target object and the trusted image of the second target object satisfies a preset similarity distribution requirement, determining the first The target object is similar to the second target object.
  • the first target object is any target object in the database of the target object; the second target object is any target object other than the first target object in the database; or A target object is a target object specified by the user, and the second target object is any target object in the database of the target object.
  • the trusted image of the first target object is filtered in the plurality of images of the first target object, between the feature and the standard feature of the first target object
  • the similarity satisfies the image of the first preset similarity requirement
  • the trusted image of the second target object is filtered in the plurality of images of the second target object, and the similarity between the feature and the standard feature of the second target object
  • An image satisfying the second preset similarity requirement so that when the executable instruction in the storage medium of the embodiment is executed, it is determined based on the standard feature of the target object, and is determined in multiple images of the target object.
  • the execution instruction is executed, the first target object is determined based on the trusted image of the first target object and the trusted image of the second target object. If the second target object is similar, storage medium according to the present embodiment, executable instructions, when executed, image processing effect, and accurate determination result, it is possible to accurately determine the first target object and the second target object are similar.
  • the storage medium may be a USB flash drive, an optical disk, a hard disk, or the like, and the computer executable instructions stored by the storage medium, when executed by the processor, can implement the following processes:
  • the first preset similarity requirement and the second preset similarity requirement are the same, or the first preset similarity requirement and the second preset similarity requirement are different.
  • the executable instructions in the storage medium of the embodiment When the executable instructions in the storage medium of the embodiment are executed, first acquiring the features of the plurality of images of the target object and the standard features of the target object, and then according to the similarity between the features of the plurality of images and the standard features, Determining the trusted image of the target object in multiple images, the similarity between the feature of the trusted image and the standard feature satisfies the preset similarity requirement, and finally determines the two targets according to the trusted image of the two target objects. Whether the objects are similar.
  • the executable instruction in the storage medium when the executable instruction in the storage medium is executed, based on the standard feature of the target object, determining that the standard feature of the target object has high similarity and can reflect the target object in multiple images of the target object.
  • the trusted image of the standard feature can make the selected trusted image suitable for image processing, improve the image processing effect, and judge the first image based on the trusted image of the first target object and the trusted image of the second target object. Whether a target object is similar to the second target object can improve the accuracy of the similarity judgment of the two target objects, thereby accurately determining whether the two target objects are similar.
  • PLD Programmable Logic Device
  • FPGA Field Programmable Gate Array
  • HDL Hardware Description Language
  • the controller can be implemented in any suitable manner, for example, the controller can take the form of, for example, a microprocessor or processor and a computer readable medium storing computer readable program code (eg, software or firmware) executable by the (micro)processor.
  • computer readable program code eg, software or firmware
  • examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, The Microchip PIC18F26K20 and the Silicone Labs C8051F320, the memory controller can also be implemented as part of the memory's control logic.
  • the controller can be logically programmed by means of logic gates, switches, ASICs, programmable logic controllers, and embedding.
  • Such a controller can therefore be considered a hardware component, and the means for implementing various functions included therein can also be considered as a structure within the hardware component.
  • a device for implementing various functions can be considered as a software module that can be both a method of implementation and a structure within a hardware component.
  • the system, device, module or unit illustrated in the above embodiments may be implemented by a computer chip or an entity, or by a product having a certain function.
  • a typical implementation device is a computer.
  • the computer can be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or A combination of any of these devices.
  • embodiments of the present application can be provided as a method, system, or computer program product.
  • the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment in combination of software and hardware.
  • the application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) including computer usable program code.
  • the computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device.
  • the apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.
  • These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device.
  • the instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.
  • a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
  • processors CPUs
  • input/output interfaces network interfaces
  • memory volatile and non-volatile memory
  • the memory may include non-persistent memory, random access memory (RAM), and/or non-volatile memory in a computer readable medium, such as read only memory (ROM) or flash memory.
  • RAM random access memory
  • ROM read only memory
  • Memory is an example of a computer readable medium.
  • Computer readable media includes both permanent and non-persistent, removable and non-removable media.
  • Information storage can be implemented by any method or technology.
  • the information can be computer readable instructions, data structures, modules of programs, or other data.
  • Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory. (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other memory technology, compact disk read only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, Magnetic tape cartridges, magnetic tape storage or other magnetic storage devices or any other non-transportable media can be used to store information that can be accessed by a computing device.
  • computer readable media does not include temporary storage of computer readable media, such as modulated data signals and carrier waves.
  • embodiments of the present application can be provided as a method, system, or computer program product.
  • the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment in combination of software and hardware.
  • the application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) including computer usable program code.
  • the application can be described in the general context of computer-executable instructions executed by a computer, such as a program module.
  • program modules include routines, programs, objects, components, data structures, and the like that perform particular tasks or implement particular abstract data types.
  • the present application can also be practiced in distributed computing environments where tasks are performed by remote processing devices that are connected through a communication network.
  • program modules can be located in both local and remote computer storage media including storage devices.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • Multimedia (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Human Computer Interaction (AREA)
  • Computing Systems (AREA)
  • Databases & Information Systems (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Image Analysis (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Processing Or Creating Images (AREA)

Abstract

本申请实施例提供了一种图像处理方法及装置，其中方法主要包括以下流程：获取目标对象的多幅图像的特征及目标对象的标准特征；根据目标对象的多幅图像的特征与其标准特征之间的相似度，在目标对象的多幅图像中确定目标对象的可信图像；其中，目标对象的可信图像的特征与目标对象的标准特征之间的相似度满足预设相似度要求。本申请实施例提供的图像处理方法可以应用在图像比对、身份识别、目标对象查找、相似目标对象判定等应用场景。

Description

图像处理方法及装置 技术领域
本申请涉及图像处理领域，尤其涉及一种图像处理方法及装置。
背景技术
随着图像处理技术的发展，图像处理已经应用在多种领域内，如人脸支付领域、身份识别领域等。在进行图像处理时，通常需要以系统内预留的图像为处理依据，通过对系统内预留的图像进行处理，得到处理结果，比如，将系统内预留的底图与采集到的用户图像进行对比，以对用户身份进行校验。
对于同一目标对象而言，由于系统内预留的该目标对象的图像质量参差不齐，因此为提高图像处理效果，需要提供一种技术方案以在该目标对象的多幅底图中筛选出该目标对象的高质量图像。
发明内容
本申请实施例的目的是提供一种图像处理方法及装置，以目标对象的标准特征为依据，在目标对象的多幅图像中确定出与目标对象的标准特征相似度较高的图像，从而使选择出的图像均适合进行图像处理，提高图像处理效果。
为解决上述技术问题，本申请实施例是这样实现的：
本申请实施例提供了一种图像处理方法，包括：
获取目标对象的多幅图像的特征及所述目标对象的标准特征；
根据所述多幅图像的特征与所述标准特征之间的相似度，在所述多幅图像中确定所述目标对象的可信图像；其中，所述可信图像的特征与所述标准特征之间的相似度满足预设相似度要求。
本申请实施例提供了另一种图像处理方法，包括：
获取第一目标对象的可信图像和第二目标对象的可信图像；
根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似；
其中，所述第一目标对象的可信图像为在所述第一目标对象的多幅图像中确定的图像，所述第一目标对象的可信图像的特征与所述第一目标对象的标准 特征之间的相似度满足第一预设相似度要求；所述第二目标对象的可信图像为在所述第二目标对象的多幅图像中确定的图像，所述第二目标对象的可信图像的特征与所述第二目标对象的标准特征之间的相似度满足第二预设相似度要求。
本申请实施例提供了又一种图像处理方法，包括：
获取第一目标对象的多幅图像的特征、所述第一目标对象的标准特征、第二目标对象的多幅图像的特征、以及所述第二目标对象的标准特征；
根据所述第一目标对象的多幅图像的特征与所述第一目标对象的标准特征之间的相似度，在所述第一目标对象的多幅图像中确定所述第一目标对象的可信图像，以及，根据所述第二目标对象的多幅图像的特征与所述第二目标对象的标准特征之间的相似度，在所述第二目标对象的多幅图像中确定所述第二目标对象的可信图像；其中，所述第一目标对象的可信图像的特征与其标准特征之间的相似度满足第一预设相似度要求，所述第二目标对象的可信图像的特征与其标准特征之间的相似度满足第二预设相似度要求；
根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似。
本申请实施例提供了一种图像处理装置，包括：
特征获取模块，用于获取目标对象的多幅图像的特征及所述目标对象的标准特征；
图像筛选模块，用于根据所述多幅图像的特征与所述标准特征之间的相似度，在所述多幅图像中确定所述目标对象的可信图像；其中，所述可信图像的特征与所述标准特征之间的相似度满足预设相似度要求。
本申请实施例提供了另一种图像处理装置，包括：
图像获取模块，用于获取第一目标对象的可信图像和第二目标对象的可信图像；
图像比较模块，用于根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似；
其中，所述第一目标对象的可信图像为在所述第一目标对象的多幅图像中确定的图像，所述第一目标对象的可信图像的特征与所述第一目标对象的标准特征之间的相似度满足第一预设相似度要求；所述第二目标对象的可信图像为在所述第二目标对象的多幅图像中确定的图像，所述第二目标对象的可信图像 的特征与所述第二目标对象的标准特征之间的相似度满足第二预设相似度要求。
本申请实施例提供了又一种图像处理装置，包括：
数据获取模块，用于获取第一目标对象的多幅图像的特征、所述第一目标对象的标准特征、第二目标对象的多幅图像的特征、以及所述第二目标对象的标准特征；
图像确定模块，用于根据所述第一目标对象的多幅图像的特征与所述第一目标对象的标准特征之间的相似度，在所述第一目标对象的多幅图像中确定所述第一目标对象的可信图像，以及，根据所述第二目标对象的多幅图像的特征与所述第二目标对象的标准特征之间的相似度，在所述第二目标对象的多幅图像中确定所述第二目标对象的可信图像；其中，所述第一目标对象的可信图像的特征与其标准特征之间的相似度满足第一预设相似度要求，所述第二目标对象的可信图像的特征与其标准特征之间的相似度满足第二预设相似度要求；
图像判断模块，用于根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似。
本申请实施例提供了一种图像处理设备，包括：
处理器；以及
被安排成存储计算机可执行指令的存储器，所述可执行指令在被执行时使所述处理器：
获取目标对象的多幅图像的特征及所述目标对象的标准特征；
根据所述多幅图像的特征与所述标准特征之间的相似度，在所述多幅图像中确定所述目标对象的可信图像；其中，所述可信图像的特征与所述标准特征之间的相似度满足预设相似度要求。
本申请实施例提供了另一种图像处理设备，包括：
处理器；以及
被安排成存储计算机可执行指令的存储器，所述可执行指令在被执行时使所述处理器：
获取第一目标对象的可信图像和第二目标对象的可信图像；
根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似；
其中，所述第一目标对象的可信图像为在所述第一目标对象的多幅图像中 确定的图像，所述第一目标对象的可信图像的特征与所述第一目标对象的标准特征之间的相似度满足第一预设相似度要求；所述第二目标对象的可信图像为在所述第二目标对象的多幅图像中确定的图像，所述第二目标对象的可信图像的特征与所述第二目标对象的标准特征之间的相似度满足第二预设相似度要求。
本申请实施例提供了又一种图像处理设备，包括：
处理器；以及
被安排成存储计算机可执行指令的存储器，所述可执行指令在被执行时使所述处理器：
获取第一目标对象的多幅图像的特征、所述第一目标对象的标准特征、第二目标对象的多幅图像的特征、以及所述第二目标对象的标准特征；
根据所述第一目标对象的多幅图像的特征与所述第一目标对象的标准特征之间的相似度，在所述第一目标对象的多幅图像中确定所述第一目标对象的可信图像，以及，根据所述第二目标对象的多幅图像的特征与所述第二目标对象的标准特征之间的相似度，在所述第二目标对象的多幅图像中确定所述第二目标对象的可信图像；其中，所述第一目标对象的可信图像的特征与其标准特征之间的相似度满足第一预设相似度要求，所述第二目标对象的可信图像的特征与其标准特征之间的相似度满足第二预设相似度要求；
根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似。
本申请实施例提供了一种存储介质，用于存储计算机可执行指令，所述可执行指令在被执行时实现以下流程：
获取目标对象的多幅图像的特征及所述目标对象的标准特征；
根据所述多幅图像的特征与所述标准特征之间的相似度，在所述多幅图像中确定所述目标对象的可信图像；其中，所述可信图像的特征与所述标准特征之间的相似度满足预设相似度要求。
本申请实施例提供了另一种存储介质，用于存储计算机可执行指令，所述可执行指令在被执行时实现以下流程：
获取第一目标对象的可信图像和第二目标对象的可信图像；
根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似；
其中，所述第一目标对象的可信图像为在所述第一目标对象的多幅图像中确定的图像，所述第一目标对象的可信图像的特征与所述第一目标对象的标准特征之间的相似度满足第一预设相似度要求；所述第二目标对象的可信图像为在所述第二目标对象的多幅图像中确定的图像，所述第二目标对象的可信图像的特征与所述第二目标对象的标准特征之间的相似度满足第二预设相似度要求。
本申请实施例提供了又一种存储介质，用于存储计算机可执行指令，所述可执行指令在被执行时实现以下流程：
获取第一目标对象的多幅图像的特征、所述第一目标对象的标准特征、第二目标对象的多幅图像的特征、以及所述第二目标对象的标准特征；
根据所述第一目标对象的多幅图像的特征与所述第一目标对象的标准特征之间的相似度，在所述第一目标对象的多幅图像中确定所述第一目标对象的可信图像，以及，根据所述第二目标对象的多幅图像的特征与所述第二目标对象的标准特征之间的相似度，在所述第二目标对象的多幅图像中确定所述第二目标对象的可信图像；其中，所述第一目标对象的可信图像的特征与其标准特征之间的相似度满足第一预设相似度要求，所述第二目标对象的可信图像的特征与其标准特征之间的相似度满足第二预设相似度要求；
根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似。
通过本实施例中的技术方案，能够以目标对象的标准特征为依据，在目标对象的多幅图像中确定出与目标对象的标准特征相似度较高、能够反映目标对象的标准特征的可信图像，从而使选择出的可信图像均适合进行图像处理，提高图像处理效果。
附图说明
为了更清楚地说明本申请实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请中记载的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。
图1为本申请实施例提供的图像处理方法的第一种流程示意图；
图2a为本发明实施例提供的根据目标对象的标准图像确定目标对象的可 信图像的示意图；
图2b为本发明实施例提供的多幅图像的特征与标准特征之间的相似度的分布示意图；
图3为本申请实施例提供的图像处理方法的第二种流程示意图；
图4为本发明实施例提供的根据可信图像比较第一目标对象和第二目标对象是否相似的示意图；
图5为本申请实施例提供的图像处理方法的第三种流程示意图；
图6为本申请实施例提供的图像处理方法的第四种流程示意图；
图7为本申请实施例提供的图像处理方法的第五种流程示意图；
图8为本申请实施例提供的图像处理装置的第一种模块组成示意图；
图9为本申请实施例提供的图像处理装置的第二种模块组成示意图；
图10为本申请实施例提供的图像处理装置的第三种模块组成示意图；
图11为本申请实施例提供的图像处理设备的结构示意图。
具体实施方式
为了使本技术领域的人员更好地理解本申请中的技术方案，下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都应当属于本申请保护的范围。
本申请实施例提供了一种图像处理方法及装置，能够在目标对象的多幅图像中筛选出目标对象的可信图像，还能够基于目标对象的可信图像确定两个目标对象是否相似，其中，筛选出的目标对象的可信图像与目标对象的标准特征相似度较高，能够反映目标对象的标准特征。
图1为本申请实施例提供的图像处理方法的第一种流程示意图，该方法的执行主体为服务器，如图1所示，该方法至少包括以下步骤：
步骤S102，获取目标对象的多幅图像的特征及目标对象的标准特征；
步骤S104，根据多幅图像的特征与该标准特征之间的相似度，在多幅图像中确定目标对象的可信图像；其中，可信图像的特征与该标准特征之间的相似度满足预设相似度要求。
本实施例中的图像处理方法，首先获取目标对象的多幅图像的特征及目标 对象的标准特征，然后根据多幅图像的特征与该标准特征之间的相似度，在多幅图像中确定目标对象的可信图像，可信图像的特征与该标准特征之间的相似度满足预设相似度要求。可见，本实施例中的图像处理方法，能够以目标对象的标准特征为依据，在目标对象的多幅图像中确定出与目标对象的标准特征相似度较高、能够反映目标对象的标准特征的可信图像，从而使选择出的可信图像均适合进行图像处理，提高图像处理效果。
本实施例中，目标对象可以是自然人，还可以是物品。上述步骤S102中，可以从本地数据库或远端数据库中读取预先存储的目标对象的多幅图像，并获取目标对象的多幅图像的特征，其中，目标对象的每幅图像都具有对应的特征。
目标对象的标准特征指的是能够准确反映目标对象的特征，利用能够准确反映目标对象的特征在目标对象的多幅图像中筛选可信图像，能够保证可信图像同样准确反映目标对象。考虑到标准特征需要准确反映目标对象，本实施例中，通过以下方式(a1)或(a2)获取目标对象的标准特征：
(a1)获取多幅图像的平均化特征，将多幅图像的平均化特征作为目标对象的标准特征；
(a2)获取多幅图像的平均化特征，将多幅图像的特征中与平均化特征相似度最高的特征，作为目标对象的标准特征。
上述方式(a1)中，对目标对象的多幅图像做特征平均化处理，具体平均化的方式可以根据实际实施场景确定，这里不做限制，一种具体的实施场景下，可以对获取到的所有图像的同一维度的特征向量做平均化，从而得到所有图像的平均化特征。本方式中，将获取的所有图像的平均化特征作为目标对象的标准特征。
上述方式(a2)中，首先获取上述多幅图像的平均化特征，这一过程与方式(a1)相同，这里不做赘述。由于目标对象的每幅图像都具有对应的特征，因此这里在多幅图像的特征中，确定与该平均化特征相似度最高的特征，将该相似度最高的特征确定为目标对象的标准特征，具体地，在目标对象的多幅图像中，分别计算每幅图像的特征与平均化特征之间的相似度，得到相似度最高的图像，将该相似度最高的图像的特征作为目标对象的标准特征。通过这种方式，能够得到一幅特征与上述平均化特征相似度最高的图像，该图像可以称为标准图像，该标准图像反映了目标对象的标准特征，其可以用在目标对象的其他图像处理过程中。
比较上述方式(a1)和方式(a2)可知，方式(a1)确定标准特征的过程简单，操作方便，方式(a2)在确定标准特征的过程中，还能够得到与标准特征相似度最高的标准图像，以便于目标对象的其他图像处理过程，本领域普通技术人员可以根据实际情况选择方式(a1)和方式(a2)中的一种以确定目标对象的标准特征。
上述步骤S104中，具体可以通过以下方式(b1)或(b2)根据多幅图像的特征与标准特征之间的相似度，在多幅图像中确定目标对象的可信图像：
(b1)在多幅图像中，将特征与标准特征之间的相似度大于预设相似度阈值的部分或全部图像，作为目标对象的可信图像；
(b2)确定多幅图像的特征与标准特征之间的相似度的分布数据，在该分布数据中确定图像密度大于预设密度的相似度区间，将确定的相似度区间所对应的部分或全部图像，作为目标对象的可信图像。
上述方式(b1)中，分别计算目标对象的多幅图像中的每幅图像的特征，与目标对象的标准特征之间的相似度，然后确定相似度大于预设相似度阈值的多幅图像，能够理解，该多幅图像与目标对象的标准特征都非常接近，都能够反映目标对象的标准特征，然而，考虑到多幅图像的数量可能非常多，将影响计算速度，因此本方式中，若不考虑计算速度，则可以将确定出来的多幅图像中的全部图像，作为目标对象的可信图像，若考虑计算速度，则可以将确定出来的多幅图像中的部分图像，作为目标对象的可信图像。本方式中，预设相似度阈值为服务器预设的值，预设方式可以由服务器根据本实施例的实施场景设置。
一种具体的例子为，计算目标对象的多幅图像中的每幅图像的特征，与目标对象的标准特征之间的相似度，按照相似度最高至最低的顺序，将目标对象的多幅图像进行排序，在排序中，确定相似度大于预设相似度阈值的图像。考虑到计算速度的需求，若相似度大于预设相似度阈值的图像的数量大于等于一定值，如一定值为100，则按照相似度最高至最低的排序，在相似度大于预设相似度阈值的图像中，选择前一半一定值数量的图像(这里为50张图像)，作为目标对象的可信图像，若相似度大于预设相似度阈值的图像的数量小于一定值，则按照相似度最高至最低的排序，在相似度大于预设相似度阈值的图像中，选择前一半数量的图像，作为目标对象的可信图像。
以目标对象的标准特征是目标对象的标准图像的特征为例，图2a为本发 明实施例提供的根据目标对象的标准图像确定目标对象的可信图像的示意图，如图2a所示，目标对象为自然人，图像a为目标对象的标准图像，图像a1、a2、a3、a4分别为获取到的目标对象的多幅图像中的部分图像，分别计算标准图像a与图像a1、a2、a3、a4之间的相似度，得到标准图像a与图像a1、a2、a3、a4之间的相似度分别为85％、65％、87％、86％，设定预设相似度阈值为80％，则确定目标对象的可信图像为a1、a3和a4，a2为不可信图像。
上述方式(b2)中，在目标对象的多幅图像中，分别计算每幅图像的特征与目标对象的标准特征之间的相似度，得到多个相似度数据，统计多个相似度数据的分布，得到分布数据，该分布数据可以以统计分布直方图的形式体现，该分布数据中，包括多个相似度区间，并标记了每个相似度区间对应的图像数量。在该分布数据中，确定图像密度大于预设密度的相似度区间，这里图像密度可以用相似度区间内的图像数量与该相似度区间的相似度范围的比值表示，图像密度越大，说明该相似度区间内每个相似度单位对应的图像数量越多，预设密度为服务器预先设定的值，服务器可以根据场景需要进行设置。确定图像密度大于预设密度的相似度区间后，与方式(b1)中同理，若不考虑计算速度，则将该确定的相似度区间内的全部图像，确定为目标对象的可信图像，若考虑计算速度，则将该确定的相似度区间内的部分图像，确定为目标对象的可信图像。其中，图像密度大于预设密度的相似度区间内的图像不一定是与标准特征最为相似的图像。
上述方式(b2)中，图像密度大于预设密度的相似度区间内分布的图像主要为目标对象的常态图像，如目标对象是自然人，其拍照时习惯大笑，则在统计目标对象的上述分布数据时，图像密度大于预设密度的相似度区间内分布的图像主要为目标对象的大笑图像。通过将图像密度大于预设密度的相似度区间内的部分或全部图像确定为可信图像，能够使目标对象的可信图像为其常态图像。
图2b为本发明实施例提供的多幅图像的特征与标准特征之间的相似度的分布示意图，如图2b所示，根据目标对象的每幅图像的特征与目标对象的标准特征之间的相似度的分布数据，得到统计分布直方图，该图包括多个相似度区间，并标记了每个相似度区间对应的图像数量。该图中，图像密度大于预设密度的相似度区间为相似度得分在60和80之间的区间，以及相似度得分在80和100之间的区间，图中以斜线方式示意，则将这两个区间对应的部分或全部 图像，确定为目标对象的可信图像。
比较上述方式(b1)和方式(b2)可知，上述方式(b1)主要以预设相似度阈值为依据，在目标对象的多幅图像中确定目标对象的可信图像，上述方式(b2)主要以相似度的分布数据为依据，在目标对象的多幅图像中确定目标对象的可信图像。当目标对象的多幅图像中都是目标图像本身的图像时，则采用上述方式(b1)能够简单快速的获得与目标对象的标准特征非常接近的目标对象的可信图像，当目标对象的多幅图像中混入少量的杂质图像(如无意义图像或者其他目标对象的图像)时，则采用上述方式(b2)能够规避杂质图像的干扰，将目标对象的常态图像作为目标对象的可信图像。本领域普通技术人员可以根据实际情况选择方式(b1)和方式(b2)中的一种以确定目标对象的可信图像。
上述方式(b2)中，基于相似度分布的方式确定可信图像时，若在相似度的分布数据中确定孤立的高相似图像，如只有一幅图像与标准特征之间的相似度达到98分，其余图像与标准特征之间的相似度最大为90分，则该孤立的高相似图像可能与标准特征完全重合，也可能为计算存在误差的图像，这种情况下，为保证图像选取精度，可以排除该图像，当热，也可以不排除，视具体场景而定。
通过以上过程可知，在目标对象的多幅图像中确定目标对象的可信图像，能够在目标对象的多幅图像中去除低质量的、与目标对象的标准特征不相近的图像，从而得到高质量的、能够反映目标对象的标准特征的图像，从而在利用目标对象的可信图像进行图像处理时，提高图像处理的精准度。
图3为本申请实施例提供的图像处理方法的第二种流程示意图，通过图3中的方法，能够在获取目标对象的可信图像的基础上，比较两个目标对象是否相似，图3中的方法优选适用在目标对象是自然人的场景下，如图3所示，该方法在图1的基础上，还包括以下步骤：
步骤S106，确定第一目标对象的可信图像与第二目标对象的可信图像之间的相似度，得到多个相似度数据；
步骤S108，根据该多个相似度数据，确定第一目标对象与第二目标对象是否相似。
步骤S102中的目标对象可以包括多个目标对象，分别为第一目标对象、第二目标对象、第三目标对象等等。因此通过步骤S102和步骤S104，能够分 别确定第一目标对象的可信图像和第二目标对象的可信图像。
步骤S106中，确定第一目标对象的可信图像与第二目标对象的可信图像之间的相似度的具体方式为，确定第一目标对象的每幅可信图像与第二目标对象的每幅可信图像之间的相似度，相似度计算方法这里不做具体限制，可以根据需要选择。在第一目标对象的每幅可信图像与第二目标对象的每幅可信图像都计算过相似度后，计算得到的相似度数据的数量等于第一目标对象的可信图像数量与第二目标对象的可信图像数量的乘积。
步骤S108中，可以通过以下方式(c1)或(c2)根据多个相似度数据，确定第一目标对象与第二目标对象是否相似：
(c1)若多个相似度数据的平均值大于预设平均阈值，则确定第一目标对象与第二目标对象相似；
(c2)若多个相似度数据的分布满足预设相似度分布要求，则确定第一目标对象与第二目标对象相似。
上式方式(c1)中，计算多个相似度数据的平均值，若多个相似度数据的平均值大于预设平均阈值，则确定第一目标对象与第二目标对象相似，否则，确定第一目标对象与第二目标对象不相似。预设平均阈值为服务器预设的值，预设方式可以由服务器根据本实施例的实施场景设置。
上式方式(c2)中，计算多个相似度数据的分布，该分布可以用统计分布直方图的形式体现，该分布中，包括多个相似度区间，并标记了每个相似度区间对应的图像对的数量。
一种具体的实施例中，预设相似度分布要求可以为相似度大于一定值的图像对的数量占总图像比对次数的比例大于预设比例，在上述分布中，若相似度大于一定值的图像对的数量占总图像比对次数的比例大于预设比例，则确定第一目标对象与第二目标对象相似，否则，确定第一目标对象与第二目标对象不相似。如相似度大于75分的图像对的数量占总图像比对次数的比例大于80％，则确定第一目标对象与第二目标对象相似，否则，确定第一目标对象与第二目标对象不相似。其中，由于第一目标对象和第二目标对象是成对比对图像的，因此确定相似度大于一定值的图像对的数量，总图像比对次数等于第一目标对象的可信图像数量与第二目标对象的可信图像数量的乘积。
另一种具体的实施例中，预设相似度分布要求可以为对应的图像数量最多的相似度区间的相似度范围在预设相似度范围内，在上述分布中，若对应的图 像数量最多的相似度区间的相似度范围在预设相似度范围内，则确定第一目标对象与第二目标对象相似，否则，确定第一目标对象与第二目标对象不相似。如对应的图像数量最多的相似度区间的相似度范围为80分到85分，在预设相似度范围70分到90分之间，则确定第一目标对象与第二目标对象相似，否则，确定第一目标对象与第二目标对象不相似。
比较上述方式(c1)和上述方式(c2)，通过方式(c1)能够利用预设平均阈值简单快速的判断第一目标对象与第二目标对象是否相似，通过方式(c2)能够根据相似度数据的分布情况确定第一目标对象与第二目标对象是否相似，在具体实施本实施例时，可以根据实施场景需要确定方式(c1)和方式(c2)中的一种以确定第一目标对象与第二目标对象是否相似。
考虑到第一目标对象的图像数量较少，和/或，第二目标对象的图像数量较少时，上述步骤S106中得到的多个相似度数据的数量较少，可能会影响步骤S108判断的准确性，因此本实施例中，步骤S108，根据该多个相似度数据，确定第一目标对象与第二目标对象是否相似，具体为：确定多个相似度数据的数量是否满足预设数量要求，若是，则根据该多个相似度数据，确定第一目标对象与第二目标对象是否相似。通过确定步骤S106中得到的多个相似度数据的数量满足预设数量要求，能够保证第一目标对象和第二目标对象相似性判断的准确性。
一种具体的实施例中，若服务器根据相似度数据确定第一目标对象与第二目标对象是否相似的算法准确度较高，则可以不确定步骤S106中得到的多个相似度数据的数量是否满足预设数量要求，直接根据相似度数据确定第一目标对象与第二目标对象是否相似，若服务器根据相似度数据确定第一目标对象与第二目标对象是否相似的算法准确度较低，则在需要先确定步骤S106中得到的多个相似度数据的数量是否满足预设数量要求，若满足，则根据相似度数据确定第一目标对象与第二目标对象是否相似，若不满足，结束方法流程。
图4为本发明实施例提供的根据可信图像比较第一目标对象和第二目标对象是否相似的示意图，如图4所示，第一目标对象包括p1、p2、p3…pm m个可信图片(图中以三个图片为例说明)，第二目标对象包括q1、q2、q3…qn n个可信图片(图中以三个图片为例说明)，分别计算第一目标对象的每个可信图片与第二目标对象的每个可信图片之间的相似度，得到多个相似度得分，图中标注了部分相似度得分，统计所有相似度得分的平均分，为85分，大于预 设平均阈值，确定第一目标对象与第二目标对象相似。
根据上述公开的方法，本实施例提供了另一种图像处理方法，既包括可信图像确定过程，也包括两个目标对象是否相似的确定过程，图5为本申请实施例提供的图像处理方法的第三种流程示意图，如图5所示，该方法包括：
步骤S302，获取图像。获取第一目标对象的多幅图像和第二目标对象的多幅图像。
步骤S304，确定第一平均化特征和第二平均化特征。确定第一目标对象的多幅图像的平均化特征作为第一平均化特征，并确定第二目标对象的多幅图像的平均化特征作为第二平均化特征。
步骤S306，计算第一相似度和第二相似度。在第一目标对象的多幅图像中，分别计算第一目标对象的每幅图像的特征与第一平均化特征的第一相似度，并在第二目标对象的多幅图像中，分别计算第二目标对象的每幅图像的特征与第二平均化特征的第二相似度。
步骤S308，确定第一标准图像和第二标准图像。将第一相似度最大的图像作为第一目标对象的第一标准图像，将第二相似度最大的图像作为第二目标对象的第二标准图像。
步骤S310，计算第三相似度和第四相似度。在第一目标对象的多幅图像中，分别计算第一目标对象的每幅图像与第一标准图像之间的第三相似度，并在第二目标对象的多幅图像中，分别计算第二目标对象的每幅图像与第二标准图像之间的第四相似度。
步骤S312，确定第一可信图像和第二可信图像。将第三相似度大于一定值的图像作为第一目标对象的第一可信图像，将第四相似度大于一定值的图像作为第二目标对象的第二可信图像。
步骤S314，计算图像数量乘积。计算第一可信图像的数量与第二可信图像的数量的乘积。
步骤S316，计算相似度平均值。分别计算每幅第一可信图像与每幅第二可信图像之间的相似度，得到多个相似度数据，计算该多个相似度数据的平均值。
步骤S318，判断步骤S314计算得到的乘积是否大于预设乘积，若大于，则执行步骤S320，否则，结束对比。
步骤S320，判断步骤S316计算得到的平均值是否大于预设平均值，若大于，则执行步骤S322，否则，结束对比。
步骤S322，确定第一目标对象与第二目标对象相似。
图1至图5所示的图像处理方法中，在计算相似度时，如计算每幅图像与平均化特征之间的相似度，计算每幅图像与标准图像之间的相似度，或者计算第一目标对象的可信图像与第二目标对象的可信图像之间的相似度时，不限制相似度计算方法，可以计算特征向量之间的欧式距离或余弦距离以计算相似度，在目标对象为自然人时，还可以采用基于特征脸的比对方式等等，只需要有指标用于量化特征间的相似程度即可。其中，计算欧式距离或余弦距离时，可以采用主成分分析法(principal component analysis,简称PCA)，或者局部特征分析方法(Local Face Analysis)，神经网络方法(Neural Networks)等。
本实施例中，目标对象可以为自然人，通过本实施例中的方法，能够在系统内预留的自然人的底图中排除掉自然人的低质量图像，筛选出自然人的可信图像，自然人的可信图像可以是光线合适、脸部位置端正、图像清晰，没有浓妆和夸张配饰的图像，通过可信图像能够反映自然人的面部特征；本实施例中的方法还能够基于两个自然人的可信图像判断两个自然人是否相似。本实施例中的方法过程简单，效果准确，能够广泛应用在人脸识别领域，尤其应用在同卵双胞胎筛选、目标人物定位等领域，具有潜在的社会价值，如识别出双方都有过人脸采集历史的同卵双胞胎，寻找被拐卖儿童或追缉更名换姓的逃犯等。
进一步地，基于上述图1至图5中的方法，本申请实施例还提供了一种图像处理方法，图6为本申请实施例提供的图像处理方法的第四种流程示意图，该方法的执行主体为服务器，针对图6中的方法，这里重点介绍其与前述图1至图5中的方法的不同之处，相同之处可以参考前述图1至图5的描述，如图6所示，该方法至少包括以下步骤：
步骤S402，获取第一目标对象的可信图像和第二目标对象的可信图像；
步骤S404，根据第一目标对象的可信图像与第二目标对象的可信图像，确定第一目标对象与第二目标对象是否相似；
其中，第一目标对象的可信图像为在第一目标对象的多幅图像中确定的图像，第一目标对象的可信图像的特征与第一目标对象的标准特征之间的相似度满足第一预设相似度要求；第二目标对象的可信图像为在第二目标对象的多幅图像中确定的图像，第二目标对象的可信图像的特征与第二目标对象的标准特征之间的相似度满足第二预设相似度要求。
本实施例中的图像处理方法，由于第一目标对象的可信图像是在第一目标 对象的多幅图像中筛选的，特征与第一目标对象的标准特征之间的相似度满足第一预设相似度要求的图像，且第二目标对象的可信图像是在第二目标对象的多幅图像中筛选的，特征与第二目标对象的标准特征之间的相似度满足第二预设相似度要求的图像，因此本实施例中的图像处理方法实现了以目标对象的标准特征为依据，在目标对象的多幅图像中确定出与目标对象的标准特征相似度较高、能够反映目标对象的标准特征的图像，从而使选择出的图像均适合进行图像处理，提高图像处理效果，并且，由于本实施例中的方法是以第一目标对象的可信图像和第二目标对象的可信图像为依据，判断第一目标对象与第二目标对象是否相似，因此本实施例中的图像处理方法图像处理效果好，判断结果准确，能够准确确定第一目标对象与第二目标对象是否相似。
根据图1至图5的描述可知，本实施例中，第一目标对象的标准特征，为第一目标对象的多幅图像的平均化特征；或者，第一目标对象的标准特征，为第一目标对象的多幅图像的特征中与第一目标对象的多幅图像的平均化特征相似度最高的特征；第二目标对象的标准特征，为第二目标对象的多幅图像的平均化特征；或者，第二目标对象的标准特征，为第二目标对象的多幅图像的特征中与第二目标对象的多幅图像的平均化特征相似度最高的特征。上述第一预设相似度要求和第二预设相似度要求可以相同，也可以不同。
考虑到采用统一的标准确定第一目标对象的可信图像和第二目标对象的可信图像，能够保证判断第一目标对象和第二目标对象是否相似的准确性，优选地，第一目标对象的标准特征和第二目标对象的标准特征采用相同的方式确定，二者都为多幅图像的平均化特征，或者二者都为多幅图像的特征中与多幅图像的平均化特征相似度最高的特征，且第一预设相似度要求与第二预设相似度要求相同。
有关第一预设相似度要求与第二预设相似度要求的具体解释与图1至图5中的描述一致，上述步骤S402中，获取第一目标对象的可信图像和第二目标对象的可信图像的具体过程还可以参考图1至图5中的描述，这里不做赘述。
上述步骤S404中，具体通过以下方式(d1)或(d2)根据第一目标对象的可信图像与第二目标对象的可信图像，确定第一目标对象与第二目标对象是否相似：
(d1)若第一目标对象的可信图像与第二目标对象的可信图像之间的多个相似度数据的平均值大于预设平均阈值，则确定第一目标对象与第二目标对象 相似；
(d2)若第一目标对象的可信图像与第二目标对象的可信图像之间的多个相似度数据的分布满足预设相似度分布要求，则确定第一目标对象与第二目标对象相似。
上述方式(d1)和(d2)的具体解释可参考前述方式(c1)和(c2)，这里不过多赘述。
一种应用场景下，需要在目标对象的数据库内查找出特征相似的两个目标对象，如长相相似的两个自然人，这种情景下，本实施例中，第一目标对象为目标对象的数据库中的任一目标对象，第二目标对象为目标对象的数据库中除第一目标对象以外的任一目标对象，从而在数据库内查找出特征相似的两个目标对象。
另一种应用场景下，需要在数据库查找出与指定目标对象相似的目标对象，如与指定自然人长相相似的另一个自然人，这种场景下，本实施例中，第一目标对象为用户指定的目标对象，第二目标对象为目标对象的数据库中的任一目标对象，从而在数据库查找出与指定目标对象相似的目标对象。
进一步地，基于上述图1至图5中的方法，本申请实施例还提供了一种图像处理方法，图7为本申请实施例提供的图像处理方法的第五种流程示意图，该方法的执行主体为服务器，针对图7中的方法，这里重点介绍其与前述图1至图5中的方法的不同之处，相同之处可以参考前述图1至图5的描述，如图7所示，该方法至少包括以下步骤：
步骤S502，获取第一目标对象的多幅图像的特征、第一目标对象的标准特征、第二目标对象的多幅图像的特征、以及第二目标对象的标准特征；
步骤S504，根据第一目标对象的多幅图像的特征与第一目标对象的标准特征之间的相似度，在第一目标对象的多幅图像中确定第一目标对象的可信图像，以及，根据第二目标对象的多幅图像的特征与第二目标对象的标准特征之间的相似度，在第二目标对象的多幅图像中确定第二目标对象的可信图像；其中，第一目标对象的可信图像的特征与其标准特征之间的相似度满足第一预设相似度要求，第二目标对象的可信图像的特征与其标准特征之间的相似度满足第二预设相似度要求；
步骤S506，根据第一目标对象的可信图像与第二目标对象的可信图像，确定第一目标对象与第二目标对象是否相似。
本实施例中的图像处理方法，首先获取目标对象的多幅图像的特征及目标对象的标准特征，然后根据多幅图像的特征与该标准特征之间的相似度，在多幅图像中确定目标对象的可信图像，可信图像的特征与该标准特征之间的相似度满足预设相似度要求，最后根据两个目标对象的可信图像，判断这两个目标对象是否相似。本实施例中的图像处理方法，以目标对象的标准特征为依据，在目标对象的多幅图像中确定出与目标对象的标准特征相似度较高、能够反映目标对象的标准特征的可信图像，能够使选择出的可信图像均适合进行图像处理，提高图像处理效果，以第一目标对象的可信图像和第二目标对象的可信图像为依据，判断第一目标对象与第二目标对象是否相似，能够提高两个目标对象相似度判断的准确性，从而准确判断两个目标对象是否相似。
步骤S502和步骤S504的具体过程可以参考图1和图2中步骤S102和步骤S104的描述，步骤S506的具体过程可参考图6中步骤S404的描述，这里不过多赘述。
本实施例中有关标准特征的具体解释可以参考图1至图6的描述。本实施例中，第一预设相似度要求和第二预设相似度要求相同，或者，第一预设相似度要求和第二预设相似度要求不同，考虑到采用统一的标准确定第一目标对象的可信图像和第二目标对象的可信图像，能够保证判断第一目标对象和第二目标对象是否相似的准确性，优选地，第一目标对象的标准特征和第二目标对象的标准特征采用相同的方式确定，二者都为多幅图像的平均化特征，或者二者都为多幅图像的特征中与多幅图像的平均化特征相似度最高的特征，且第一预设相似度要求与第二预设相似度要求相同。
综上，本实施例中图1至图7所示的方法，能够确定目标对象的可信图像，并基于目标对象的可信图像确定两个目标对象是否相似，判断过程简单，判断效果准确，可以广泛应用在人脸识别、双胞胎筛选、目标人物定位等领域。
进一步地，基于上述图1至图7所示的方法，本申请实施例提供了一种图像处理装置，图8为本申请实施例提供的图像处理装置的第一种模块组成示意图，如图8所示，该装置包括：
特征获取模块61，用于获取目标对象的多幅图像的特征及所述目标对象的标准特征；
图像筛选模块62，用于根据所述多幅图像的特征与所述标准特征之间的相 似度，在所述多幅图像中确定所述目标对象的可信图像；其中，所述可信图像的特征与所述标准特征之间的相似度满足预设相似度要求。
本实施例中，所述特征获取模块61具体用于：获取所述多幅图像的平均化特征，将所述多幅图像的平均化特征作为所述目标对象的标准特征；或者，获取所述多幅图像的平均化特征，将所述多幅图像的特征中与所述平均化特征相似度最高的特征，作为所述目标对象的标准特征。
本实施例中，所述图像筛选模块62具体用于：在所述多幅图像中，将特征与所述标准特征之间的相似度大于预设相似度阈值的部分或全部图像，作为所述目标对象的可信图像；或者，确定所述多幅图像的特征与所述标准特征之间的相似度的分布数据，在所述分布数据中确定图像密度大于预设密度的相似度区间，将确定的所述相似度区间所对应的部分或全部图像，作为所述目标对象的可信图像。
本实施例中，该装置还包括：
相似度确定模块，用于确定第一目标对象的可信图像与第二目标对象的可信图像之间的相似度，得到多个相似度数据；
相似性判断模块，用于根据所述多个相似度数据，确定所述第一目标对象与所述第二目标对象是否相似。
本实施例中，所述相似性判断模块具体用于：若所述多个相似度数据的平均值大于预设平均阈值，则确定所述第一目标对象与所述第二目标对象相似；或者，若所述多个相似度数据的分布满足预设相似度分布要求，则确定所述第一目标对象与所述第二目标对象相似。
本实施例中，所述相似性判断模块具体用于：确定所述多个相似度数据的数量是否满足预设数量要求，若是，则根据所述多个相似度数据，确定所述第一目标对象与所述第二目标对象是否相似。
本实施例中的图像处理装置，首先获取目标对象的多幅图像的特征及目标对象的标准特征，然后根据多幅图像的特征与该标准特征之间的相似度，在多幅图像中确定目标对象的可信图像，可信图像的特征与该标准特征之间的相似度满足预设相似度要求。可见，本实施例中的图像处理装置，能够以目标对象的标准特征为依据，在目标对象的多幅图像中确定出与目标对象的标准特征相似度较高、能够反映目标对象的标准特征的可信图像，从而使选择出的可信图像均适合进行图像处理，提高图像处理效果。
进一步地，基于上述图1至图7所示的方法，本申请实施例还提供了另一种图像处理装置，图9为本申请实施例提供的图像处理装置的第二种模块组成示意图，如图9所示，该装置包括：
图像获取模块71，用于获取第一目标对象的可信图像和第二目标对象的可信图像；
图像比较模块72，用于根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似；
其中，所述第一目标对象的可信图像为在所述第一目标对象的多幅图像中确定的图像，所述第一目标对象的可信图像的特征与所述第一目标对象的标准特征之间的相似度满足第一预设相似度要求；所述第二目标对象的可信图像为在所述第二目标对象的多幅图像中确定的图像，所述第二目标对象的可信图像的特征与所述第二目标对象的标准特征之间的相似度满足第二预设相似度要求。
本实施例中，所述第一目标对象的标准特征，为所述第一目标对象的多幅图像的平均化特征；或者，所述第一目标对象的标准特征，为所述第一目标对象的多幅图像的特征中与所述第一目标对象的多幅图像的平均化特征相似度最高的特征；所述第二目标对象的标准特征，为所述第二目标对象的多幅图像的平均化特征；或者，所述第二目标对象的标准特征，为所述第二目标对象的多幅图像的特征中与所述第二目标对象的多幅图像的平均化特征相似度最高的特征。
本实施例中，所述图像比较模块72具体用于：若所述第一目标对象的可信图像与所述第二目标对象的可信图像之间的多个相似度数据的平均值大于预设平均阈值，则确定所述第一目标对象与所述第二目标对象相似；或者，若所述第一目标对象的可信图像与所述第二目标对象的可信图像之间的多个相似度数据的分布满足预设相似度分布要求，则确定所述第一目标对象与所述第二目标对象相似。
本实施例中，所述第一目标对象为目标对象的数据库中的任一目标对象；所述第二目标对象为所述数据库中除所述第一目标对象以外的任一目标对象；或者，所述第一目标对象为用户指定的目标对象，所述第二目标对象为目标对象的数据库中的任一目标对象。
本实施例中的图像处理装置，由于第一目标对象的可信图像是在第一目标对象的多幅图像中筛选的，特征与第一目标对象的标准特征之间的相似度满足第一预设相似度要求的图像，且第二目标对象的可信图像是在第二目标对象的多幅图像中筛选的，特征与第二目标对象的标准特征之间的相似度满足第二预设相似度要求的图像，因此本实施例中的图像处理装置实现了以目标对象的标准特征为依据，在目标对象的多幅图像中确定出与目标对象的标准特征相似度较高、能够反映目标对象的标准特征的图像，从而使选择出的图像均适合进行图像处理，提高图像处理效果，并且，由于本实施例中的装置是以第一目标对象的可信图像和第二目标对象的可信图像为依据，判断第一目标对象与第二目标对象是否相似，因此本实施例中的图像处理装置图像处理效果好，判断结果准确，能够准确确定第一目标对象与第二目标对象是否相似。
进一步地，基于上述图1至图7所示的方法，本申请实施例还提供了又一种图像处理装置，图10为本申请实施例提供的图像处理装置的第三种模块组成示意图，如图10所示，该装置包括：
数据获取模块81，用于获取第一目标对象的多幅图像的特征、所述第一目标对象的标准特征、第二目标对象的多幅图像的特征、以及所述第二目标对象的标准特征；
图像确定模块82，用于根据所述第一目标对象的多幅图像的特征与所述第一目标对象的标准特征之间的相似度，在所述第一目标对象的多幅图像中确定所述第一目标对象的可信图像，以及，根据所述第二目标对象的多幅图像的特征与所述第二目标对象的标准特征之间的相似度，在所述第二目标对象的多幅图像中确定所述第二目标对象的可信图像；其中，所述第一目标对象的可信图像的特征与其标准特征之间的相似度满足第一预设相似度要求，所述第二目标对象的可信图像的特征与其标准特征之间的相似度满足第二预设相似度要求；
图像判断模块83，用于根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似。
本实施例中，所述第一预设相似度要求和所述第二预设相似度要求相同，或者，所述第一预设相似度要求和所述第二预设相似度要求不同。
本实施例中的图像处理装置，首先获取目标对象的多幅图像的特征及目标对象的标准特征，然后根据多幅图像的特征与该标准特征之间的相似度，在多幅图像中确定目标对象的可信图像，可信图像的特征与该标准特征之间的相似 度满足预设相似度要求，最后根据两个目标对象的可信图像，判断这两个目标对象是否相似。本实施例中的图像处理装置，以目标对象的标准特征为依据，在目标对象的多幅图像中确定出与目标对象的标准特征相似度较高、能够反映目标对象的标准特征的可信图像，能够使选择出的可信图像均适合进行图像处理，提高图像处理效果，以第一目标对象的可信图像和第二目标对象的可信图像为依据，判断第一目标对象与第二目标对象是否相似，能够提高两个目标对象相似度判断的准确性，从而准确判断两个目标对象是否相似。
进一步地，基于上述图1至图7所示的方法，本申请实施例还提供了一种图像处理设备，如图11所示。
图像处理设备可因配置或性能不同而产生比较大的差异，可以包括一个或一个以上的处理器901和存储器902，存储器902中可以存储有一个或一个以上存储应用程序或数据。其中，存储器902可以是短暂存储或持久存储。存储在存储器902的应用程序可以包括一个或一个以上模块(图示未示出)，每个模块可以包括对图像处理设备中的一系列计算机可执行指令。更进一步地，处理器901可以设置为与存储器902通信，在图像处理设备上执行存储器902中的一系列计算机可执行指令。图像处理设备还可以包括一个或一个以上电源903，一个或一个以上有线或无线网络接口904，一个或一个以上输入输出接口905，一个或一个以上键盘906等。
在一个具体的实施例中，图像处理设备包括有存储器，以及一个或一个以上的程序，其中一个或者一个以上程序存储于存储器中，且一个或者一个以上程序可以包括一个或一个以上模块，且每个模块可以包括对图像处理设备中的一系列计算机可执行指令，且经配置以由一个或者一个以上处理器执行该一个或者一个以上程序包含用于进行以下计算机可执行指令：
获取目标对象的多幅图像的特征及所述目标对象的标准特征；
根据所述多幅图像的特征与所述标准特征之间的相似度，在所述多幅图像中确定所述目标对象的可信图像；其中，所述可信图像的特征与所述标准特征之间的相似度满足预设相似度要求。
可选地，计算机可执行指令在被执行时，获取所述目标对象的标准特征，包括：获取所述多幅图像的平均化特征，将所述多幅图像的平均化特征作为所述目标对象的标准特征；或者，获取所述多幅图像的平均化特征，将所述多幅图像的特征中与所述平均化特征相似度最高的特征，作为所述目标对象的标准 特征。
可选地，计算机可执行指令在被执行时，所述根据所述多幅图像的特征与所述标准特征之间的相似度，在所述多幅图像中确定所述目标对象的可信图像，包括：在所述多幅图像中，将特征与所述标准特征之间的相似度大于预设相似度阈值的部分或全部图像，作为所述目标对象的可信图像；或者，确定所述多幅图像的特征与所述标准特征之间的相似度的分布数据，在所述分布数据中确定图像密度大于预设密度的相似度区间，将确定的所述相似度区间所对应的部分或全部图像，作为所述目标对象的可信图像。
可选地，计算机可执行指令在被执行时，还可以使所述处理器，确定第一目标对象的可信图像与第二目标对象的可信图像之间的相似度，得到多个相似度数据；根据所述多个相似度数据，确定所述第一目标对象与所述第二目标对象是否相似。
可选地，计算机可执行指令在被执行时，所述根据所述多个相似度数据，确定所述第一目标对象与所述第二目标对象是否相似，包括：若所述多个相似度数据的平均值大于预设平均阈值，则确定所述第一目标对象与所述第二目标对象相似；或者，若所述多个相似度数据的分布满足预设相似度分布要求，则确定所述第一目标对象与所述第二目标对象相似。
可选地，计算机可执行指令在被执行时，所述根据所述多个相似度数据，确定所述第一目标对象与所述第二目标对象是否相似，包括：确定所述多个相似度数据的数量是否满足预设数量要求，若是，则根据所述多个相似度数据，确定所述第一目标对象与所述第二目标对象是否相似。
本实施例中的图像处理设备，首先获取目标对象的多幅图像的特征及目标对象的标准特征，然后根据多幅图像的特征与该标准特征之间的相似度，在多幅图像中确定目标对象的可信图像，可信图像的特征与该标准特征之间的相似度满足预设相似度要求。可见，本实施例中的图像处理设备，能够以目标对象的标准特征为依据，在目标对象的多幅图像中确定出与目标对象的标准特征相似度较高、能够反映目标对象的标准特征的可信图像，从而使选择出的可信图像均适合进行图像处理，提高图像处理效果。
在另一个具体的实施例中，图像处理设备包括有存储器，以及一个或一个以上的程序，其中一个或者一个以上程序存储于存储器中，且一个或者一个以上程序可以包括一个或一个以上模块，且每个模块可以包括对图像处理设备中 的一系列计算机可执行指令，且经配置以由一个或者一个以上处理器执行该一个或者一个以上程序包含用于进行以下计算机可执行指令：
获取第一目标对象的可信图像和第二目标对象的可信图像；
根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似；
其中，所述第一目标对象的可信图像为在所述第一目标对象的多幅图像中确定的图像，所述第一目标对象的可信图像的特征与所述第一目标对象的标准特征之间的相似度满足第一预设相似度要求；所述第二目标对象的可信图像为在所述第二目标对象的多幅图像中确定的图像，所述第二目标对象的可信图像的特征与所述第二目标对象的标准特征之间的相似度满足第二预设相似度要求。
可选地，所述第一目标对象的标准特征，为所述第一目标对象的多幅图像的平均化特征；或者，所述第一目标对象的标准特征，为所述第一目标对象的多幅图像的特征中与所述第一目标对象的多幅图像的平均化特征相似度最高的特征；所述第二目标对象的标准特征，为所述第二目标对象的多幅图像的平均化特征；或者，所述第二目标对象的标准特征，为所述第二目标对象的多幅图像的特征中与所述第二目标对象的多幅图像的平均化特征相似度最高的特征。
可选地，计算机可执行指令在被执行时，所述根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似，包括：若所述第一目标对象的可信图像与所述第二目标对象的可信图像之间的多个相似度数据的平均值大于预设平均阈值，则确定所述第一目标对象与所述第二目标对象相似；或者，若所述第一目标对象的可信图像与所述第二目标对象的可信图像之间的多个相似度数据的分布满足预设相似度分布要求，则确定所述第一目标对象与所述第二目标对象相似。
可选地，所述第一目标对象为目标对象的数据库中的任一目标对象；所述第二目标对象为所述数据库中除所述第一目标对象以外的任一目标对象；或者，所述第一目标对象为用户指定的目标对象，所述第二目标对象为目标对象的数据库中的任一目标对象。
本实施例中的图像处理设备，由于第一目标对象的可信图像是在第一目标对象的多幅图像中筛选的，特征与第一目标对象的标准特征之间的相似度满足 第一预设相似度要求的图像，且第二目标对象的可信图像是在第二目标对象的多幅图像中筛选的，特征与第二目标对象的标准特征之间的相似度满足第二预设相似度要求的图像，因此本实施例中的图像处理设备实现了以目标对象的标准特征为依据，在目标对象的多幅图像中确定出与目标对象的标准特征相似度较高、能够反映目标对象的标准特征的图像，从而使选择出的图像均适合进行图像处理，提高图像处理效果，并且，由于本实施例中的设备是以第一目标对象的可信图像和第二目标对象的可信图像为依据，判断第一目标对象与第二目标对象是否相似，因此本实施例中的图像处理设备图像处理效果好，判断结果准确，能够准确确定第一目标对象与第二目标对象是否相似。
在又一个具体的实施例中，图像处理设备包括有存储器，以及一个或一个以上的程序，其中一个或者一个以上程序存储于存储器中，且一个或者一个以上程序可以包括一个或一个以上模块，且每个模块可以包括对图像处理设备中的一系列计算机可执行指令，且经配置以由一个或者一个以上处理器执行该一个或者一个以上程序包含用于进行以下计算机可执行指令：
获取第一目标对象的多幅图像的特征、所述第一目标对象的标准特征、第二目标对象的多幅图像的特征、以及所述第二目标对象的标准特征；
根据所述第一目标对象的多幅图像的特征与所述第一目标对象的标准特征之间的相似度，在所述第一目标对象的多幅图像中确定所述第一目标对象的可信图像，以及，根据所述第二目标对象的多幅图像的特征与所述第二目标对象的标准特征之间的相似度，在所述第二目标对象的多幅图像中确定所述第二目标对象的可信图像；其中，所述第一目标对象的可信图像的特征与其标准特征之间的相似度满足第一预设相似度要求，所述第二目标对象的可信图像的特征与其标准特征之间的相似度满足第二预设相似度要求；
根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似。
可选地，所述第一预设相似度要求和所述第二预设相似度要求相同，或者，所述第一预设相似度要求和所述第二预设相似度要求不同。
本实施例中的图像处理设备，首先获取目标对象的多幅图像的特征及目标对象的标准特征，然后根据多幅图像的特征与该标准特征之间的相似度，在多幅图像中确定目标对象的可信图像，可信图像的特征与该标准特征之间的相似度满足预设相似度要求，最后根据两个目标对象的可信图像，判断这两个目标 对象是否相似。本实施例中的图像处理设备，以目标对象的标准特征为依据，在目标对象的多幅图像中确定出与目标对象的标准特征相似度较高、能够反映目标对象的标准特征的可信图像，能够使选择出的可信图像均适合进行图像处理，提高图像处理效果，以第一目标对象的可信图像和第二目标对象的可信图像为依据，判断第一目标对象与第二目标对象是否相似，能够提高两个目标对象相似度判断的准确性，从而准确判断两个目标对象是否相似。
进一步地，基于上述图1至图7所示的方法，本申请实施例还提供了一种存储介质，用于存储计算机可执行指令，一种具体的实施例中，该存储介质可以为U盘、光盘、硬盘等，该存储介质存储的计算机可执行指令在被处理器执行时，能实现以下流程：
获取目标对象的多幅图像的特征及所述目标对象的标准特征；
根据所述多幅图像的特征与所述标准特征之间的相似度，在所述多幅图像中确定所述目标对象的可信图像；其中，所述可信图像的特征与所述标准特征之间的相似度满足预设相似度要求。
本实施例中，获取所述目标对象的标准特征，包括：获取所述多幅图像的平均化特征，将所述多幅图像的平均化特征作为所述目标对象的标准特征；或者，获取所述多幅图像的平均化特征，将所述多幅图像的特征中与所述平均化特征相似度最高的特征，作为所述目标对象的标准特征。
本实施例中，所述根据所述多幅图像的特征与所述标准特征之间的相似度，在所述多幅图像中确定所述目标对象的可信图像，包括：在所述多幅图像中，将特征与所述标准特征之间的相似度大于预设相似度阈值的部分或全部图像，作为所述目标对象的可信图像；或者，确定所述多幅图像的特征与所述标准特征之间的相似度的分布数据，在所述分布数据中确定图像密度大于预设密度的相似度区间，将确定的所述相似度区间所对应的部分或全部图像，作为所述目标对象的可信图像。
本实施例中的流程还包括：确定第一目标对象的可信图像与第二目标对象的可信图像之间的相似度，得到多个相似度数据；根据所述多个相似度数据，确定所述第一目标对象与所述第二目标对象是否相似。
本实施例中，所述根据所述多个相似度数据，确定所述第一目标对象与所述第二目标对象是否相似，包括：若所述多个相似度数据的平均值大于预设平均阈值，则确定所述第一目标对象与所述第二目标对象相似；或者，若所述多 个相似度数据的分布满足预设相似度分布要求，则确定所述第一目标对象与所述第二目标对象相似。
本实施例中，所述根据所述多个相似度数据，确定所述第一目标对象与所述第二目标对象是否相似，包括：确定所述多个相似度数据的数量是否满足预设数量要求，若是，则根据所述多个相似度数据，确定所述第一目标对象与所述第二目标对象是否相似。
本实施例的存储介质中的可执行指令在被执行时，首先获取目标对象的多幅图像的特征及目标对象的标准特征，然后根据多幅图像的特征与该标准特征之间的相似度，在多幅图像中确定目标对象的可信图像，可信图像的特征与该标准特征之间的相似度满足预设相似度要求。可见，本实施例的存储介质中的可执行指令在被执行时，能够以目标对象的标准特征为依据，在目标对象的多幅图像中确定出与目标对象的标准特征相似度较高、能够反映目标对象的标准特征的可信图像，从而使选择出的可信图像均适合进行图像处理，提高图像处理效果。
在另一种具体的实施例中，该存储介质可以为U盘、光盘、硬盘等，该存储介质存储的计算机可执行指令在被处理器执行时，能实现以下流程：
获取第一目标对象的可信图像和第二目标对象的可信图像；
根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似；
其中，所述第一目标对象的可信图像为在所述第一目标对象的多幅图像中确定的图像，所述第一目标对象的可信图像的特征与所述第一目标对象的标准特征之间的相似度满足第一预设相似度要求；所述第二目标对象的可信图像为在所述第二目标对象的多幅图像中确定的图像，所述第二目标对象的可信图像的特征与所述第二目标对象的标准特征之间的相似度满足第二预设相似度要求。
其中，所述第一目标对象的标准特征，为所述第一目标对象的多幅图像的平均化特征；或者，所述第一目标对象的标准特征，为所述第一目标对象的多幅图像的特征中与所述第一目标对象的多幅图像的平均化特征相似度最高的特征；所述第二目标对象的标准特征，为所述第二目标对象的多幅图像的平均化特征；或者，所述第二目标对象的标准特征，为所述第二目标对象的多幅图像的特征中与所述第二目标对象的多幅图像的平均化特征相似度最高的特征。
其中，所述根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似，包括：若所述第一目标对象的可信图像与所述第二目标对象的可信图像之间的多个相似度数据的平均值大于预设平均阈值，则确定所述第一目标对象与所述第二目标对象相似；或者，若所述第一目标对象的可信图像与所述第二目标对象的可信图像之间的多个相似度数据的分布满足预设相似度分布要求，则确定所述第一目标对象与所述第二目标对象相似。
其中，所述第一目标对象为目标对象的数据库中的任一目标对象；所述第二目标对象为所述数据库中除所述第一目标对象以外的任一目标对象；或者，所述第一目标对象为用户指定的目标对象，所述第二目标对象为目标对象的数据库中的任一目标对象。
本实施例的存储介质中的可执行指令在被执行时，由于第一目标对象的可信图像是在第一目标对象的多幅图像中筛选的，特征与第一目标对象的标准特征之间的相似度满足第一预设相似度要求的图像，且第二目标对象的可信图像是在第二目标对象的多幅图像中筛选的，特征与第二目标对象的标准特征之间的相似度满足第二预设相似度要求的图像，因此本实施例的存储介质中的可执行指令在被执行时，实现了以目标对象的标准特征为依据，在目标对象的多幅图像中确定出与目标对象的标准特征相似度较高、能够反映目标对象的标准特征的图像，从而使选择出的图像均适合进行图像处理，提高图像处理效果，并且，由于本实施例的存储介质中的可执行指令在被执行时，是以第一目标对象的可信图像和第二目标对象的可信图像为依据，判断第一目标对象与第二目标对象是否相似，因此本实施例的存储介质中的可执行指令在被执行时，图像处理效果好，判断结果准确，能够准确确定第一目标对象与第二目标对象是否相似。
在又一种具体的实施例中，该存储介质可以为U盘、光盘、硬盘等，该存储介质存储的计算机可执行指令在被处理器执行时，能实现以下流程：
获取第一目标对象的多幅图像的特征、所述第一目标对象的标准特征、第二目标对象的多幅图像的特征、以及所述第二目标对象的标准特征；
根据所述第一目标对象的多幅图像的特征与所述第一目标对象的标准特征之间的相似度，在所述第一目标对象的多幅图像中确定所述第一目标对象的可信图像，以及，根据所述第二目标对象的多幅图像的特征与所述第二目标对 象的标准特征之间的相似度，在所述第二目标对象的多幅图像中确定所述第二目标对象的可信图像；其中，所述第一目标对象的可信图像的特征与其标准特征之间的相似度满足第一预设相似度要求，所述第二目标对象的可信图像的特征与其标准特征之间的相似度满足第二预设相似度要求；
根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似。
其中，所述第一预设相似度要求和所述第二预设相似度要求相同，或者，所述第一预设相似度要求和所述第二预设相似度要求不同。
本实施例的存储介质中的可执行指令在被执行时，首先获取目标对象的多幅图像的特征及目标对象的标准特征，然后根据多幅图像的特征与该标准特征之间的相似度，在多幅图像中确定目标对象的可信图像，可信图像的特征与该标准特征之间的相似度满足预设相似度要求，最后根据两个目标对象的可信图像，判断这两个目标对象是否相似。本实施例中存储介质中的可执行指令在被执行时，以目标对象的标准特征为依据，在目标对象的多幅图像中确定出与目标对象的标准特征相似度较高、能够反映目标对象的标准特征的可信图像，能够使选择出的可信图像均适合进行图像处理，提高图像处理效果，以第一目标对象的可信图像和第二目标对象的可信图像为依据，判断第一目标对象与第二目标对象是否相似，能够提高两个目标对象相似度判断的准确性，从而准确判断两个目标对象是否相似。
在20世纪90年代，对于一个技术的改进可以很明显地区分是硬件上的改进(例如，对二极管、晶体管、开关等电路结构的改进)还是软件上的改进(对于方法流程的改进)。然而，随着技术的发展，当今的很多方法流程的改进已经可以视为硬件电路结构的直接改进。设计人员几乎都通过将改进的方法流程编程到硬件电路中来得到相应的硬件电路结构。因此，不能说一个方法流程的改进就不能用硬件实体模块来实现。例如，可编程逻辑器件(Programmable Logic Device，PLD)(例如现场可编程门阵列(Field Programmable Gate Array，FPGA))就是这样一种集成电路，其逻辑功能由用户对器件编程来确定。由设计人员自行编程来把一个数字系统“集成”在一片PLD上，而不需要请芯片制造厂商来设计和制作专用的集成电路芯片。而且，如今，取代手工地制作集成电路芯片，这种编程也多半改用“逻辑编译器(logic compiler)”软件来实现，它与程序开发撰写时所用的软件编译器相类似，而要编译之前的原始代码也得用 特定的编程语言来撰写，此称之为硬件描述语言(Hardware Description Language，HDL)，而HDL也并非仅有一种，而是有许多种，如ABEL(Advanced Boolean Expression Language)、AHDL(Altera Hardware Description Language)、Confluence、CUPL(Cornell University Programming Language)、HDCal、JHDL(Java Hardware Description Language)、Lava、Lola、MyHDL、PALASM、RHDL(Ruby Hardware Description Language)等，目前最普遍使用的是VHDL(Very-High-Speed Integrated Circuit Hardware Description Language)与Verilog。本领域技术人员也应该清楚，只需要将方法流程用上述几种硬件描述语言稍作逻辑编程并编程到集成电路中，就可以很容易得到实现该逻辑方法流程的硬件电路。
控制器可以按任何适当的方式实现，例如，控制器可以采取例如微处理器或处理器以及存储可由该(微)处理器执行的计算机可读程序代码(例如软件或固件)的计算机可读介质、逻辑门、开关、专用集成电路(Application Specific Integrated Circuit，ASIC)、可编程逻辑控制器和嵌入微控制器的形式，控制器的例子包括但不限于以下微控制器：ARC 625D、Atmel AT91SAM、Microchip PIC18F26K20以及Silicone Labs C8051F320，存储器控制器还可以被实现为存储器的控制逻辑的一部分。本领域技术人员也知道，除了以纯计算机可读程序代码方式实现控制器以外，完全可以通过将方法步骤进行逻辑编程来使得控制器以逻辑门、开关、专用集成电路、可编程逻辑控制器和嵌入微控制器等的形式来实现相同功能。因此这种控制器可以被认为是一种硬件部件，而对其内包括的用于实现各种功能的装置也可以视为硬件部件内的结构。或者甚至，可以将用于实现各种功能的装置视为既可以是实现方法的软件模块又可以是硬件部件内的结构。
上述实施例阐明的系统、装置、模块或单元，具体可以由计算机芯片或实体实现，或者由具有某种功能的产品来实现。一种典型的实现设备为计算机。具体的，计算机例如可以为个人计算机、膝上型计算机、蜂窝电话、相机电话、智能电话、个人数字助理、媒体播放器、导航设备、电子邮件设备、游戏控制台、平板计算机、可穿戴设备或者这些设备中的任何设备的组合。
为了描述的方便，描述以上装置时以功能分为各种单元分别描述。当然，在实施本申请时可以把各单元的功能在同一个或多个软件和/或硬件中实现。
本领域内的技术人员应明白，本申请的实施例可提供为方法、系统、或计 算机程序产品。因此，本申请可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且，本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本申请是参照根据本申请实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
在一个典型的配置中，计算设备包括一个或多个处理器(CPU)、输入/输出接口、网络接口和内存。
内存可能包括计算机可读介质中的非永久性存储器，随机存取存储器(RAM)和/或非易失性内存等形式，如只读存储器(ROM)或闪存(flash RAM)。内存是计算机可读介质的示例。
计算机可读介质包括永久性和非永久性、可移动和非可移动媒体可以由任何方法或技术来实现信息存储。信息可以是计算机可读指令、数据结构、程序的模块或其他数据。计算机的存储介质的例子包括，但不限于相变内存(PRAM)、静态随机存取存储器(SRAM)、动态随机存取存储器(DRAM)、其他类型的随机存取存储器(RAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、快闪记忆体或其他内存技术、只读光盘只读存储器(CD-ROM)、数字多功能光盘(DVD)或其他光学存储、磁盒式磁带，磁带 磁磁盘存储或其他磁性存储设备或任何其他非传输介质，可用于存储可以被计算设备访问的信息。按照本文中的界定，计算机可读介质不包括暂存电脑可读媒体(transitory media)，如调制的数据信号和载波。
还需要说明的是，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、商品或者设备不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、商品或者设备所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括所述要素的过程、方法、商品或者设备中还存在另外的相同要素。
本领域技术人员应明白，本申请的实施例可提供为方法、系统或计算机程序产品。因此，本申请可采用完全硬件实施例、完全软件实施例或结合软件和硬件方面的实施例的形式。而且，本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本申请可以在由计算机执行的计算机可执行指令的一般上下文中描述，例如程序模块。一般地，程序模块包括执行特定任务或实现特定抽象数据类型的例程、程序、对象、组件、数据结构等等。也可以在分布式计算环境中实践本申请，在这些分布式计算环境中，由通过通信网络而被连接的远程处理设备来执行任务。在分布式计算环境中，程序模块可以位于包括存储设备在内的本地和远程计算机存储介质中。
本说明书中的各个实施例均采用递进的方式描述，各个实施例之间相同相似的部分互相参见即可，每个实施例重点说明的都是与其他实施例的不同之处。尤其，对于系统实施例而言，由于其基本相似于方法实施例，所以描述的比较简单，相关之处参见方法实施例的部分说明即可。
以上所述仅为本申请的实施例而已，并不用于限制本申请。对于本领域技术人员来说，本申请可以有各种更改和变化。凡在本申请的精神和原理之内所作的任何修改、等同替换、改进等，均应包含在本申请的权利要求范围之内。

Claims (21)

  1. 一种图像处理方法，其特征在于，包括：
    获取目标对象的多幅图像的特征及所述目标对象的标准特征；
    根据所述多幅图像的特征与所述标准特征之间的相似度，在所述多幅图像中确定所述目标对象的可信图像；其中，所述可信图像的特征与所述标准特征之间的相似度满足预设相似度要求。
  2. 根据权利要求1所述的方法，其特征在于，获取所述目标对象的标准特征，包括：
    获取所述多幅图像的平均化特征，将所述多幅图像的平均化特征作为所述目标对象的标准特征；
    或者，
    获取所述多幅图像的平均化特征，将所述多幅图像的特征中与所述平均化特征相似度最高的特征，作为所述目标对象的标准特征。
  3. 根据权利要求1所述的方法，其特征在于，所述根据所述多幅图像的特征与所述标准特征之间的相似度，在所述多幅图像中确定所述目标对象的可信图像，包括：
    在所述多幅图像中，将特征与所述标准特征之间的相似度大于预设相似度阈值的部分或全部图像，作为所述目标对象的可信图像；或者，
    确定所述多幅图像的特征与所述标准特征之间的相似度的分布数据，在所述分布数据中确定图像密度大于预设密度的相似度区间，将确定的所述相似度区间所对应的部分或全部图像，作为所述目标对象的可信图像。
  4. 根据权利要求1至3任一项所述的方法，其特征在于，所述方法还包括：
    确定第一目标对象的可信图像与第二目标对象的可信图像之间的相似度，得到多个相似度数据；
    根据所述多个相似度数据，确定所述第一目标对象与所述第二目标对象是否相似。
  5. 根据权利要求4所述的方法，其特征在于，所述根据所述多个相似度数据，确定所述第一目标对象与所述第二目标对象是否相似，包括：
    若所述多个相似度数据的平均值大于预设平均阈值，则确定所述第一目标对象与所述第二目标对象相似；或者，
    若所述多个相似度数据的分布满足预设相似度分布要求，则确定所述第一目标对象与所述第二目标对象相似。
  6. 根据权利要求4所述的方法，其特征在于，所述根据所述多个相似度数据，确定所述第一目标对象与所述第二目标对象是否相似，包括：
    确定所述多个相似度数据的数量是否满足预设数量要求，若是，则根据所述多个相似度数据，确定所述第一目标对象与所述第二目标对象是否相似。
  7. 一种图像处理方法，其特征在于，包括：
    获取第一目标对象的可信图像和第二目标对象的可信图像；
    根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似；
    其中，所述第一目标对象的可信图像为在所述第一目标对象的多幅图像中确定的图像，所述第一目标对象的可信图像的特征与所述第一目标对象的标准特征之间的相似度满足第一预设相似度要求；所述第二目标对象的可信图像为在所述第二目标对象的多幅图像中确定的图像，所述第二目标对象的可信图像的特征与所述第二目标对象的标准特征之间的相似度满足第二预设相似度要求。
  8. 根据权利要求7所述的方法，其特征在于，
    所述第一目标对象的标准特征，为所述第一目标对象的多幅图像的平均化特征；或者，所述第一目标对象的标准特征，为所述第一目标对象的多幅图像的特征中与所述第一目标对象的多幅图像的平均化特征相似度最高的特征；
    所述第二目标对象的标准特征，为所述第二目标对象的多幅图像的平均化特征；或者，所述第二目标对象的标准特征，为所述第二目标对象的多幅图像的特征中与所述第二目标对象的多幅图像的平均化特征相似度最高的特征。
  9. 根据权利要求7所述的方法，其特征在于，所述根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似，包括：
    若所述第一目标对象的可信图像与所述第二目标对象的可信图像之间的多个相似度数据的平均值大于预设平均阈值，则确定所述第一目标对象与所述第二目标对象相似；或者，
    若所述第一目标对象的可信图像与所述第二目标对象的可信图像之间的多个相似度数据的分布满足预设相似度分布要求，则确定所述第一目标对象与 所述第二目标对象相似。
  10. 根据权利要求7至9任一项所述的方法，其特征在于，
    所述第一目标对象为目标对象的数据库中的任一目标对象；所述第二目标对象为所述数据库中除所述第一目标对象以外的任一目标对象；或者，
    所述第一目标对象为用户指定的目标对象，所述第二目标对象为目标对象的数据库中的任一目标对象。
  11. 一种图像处理方法，其特征在于，包括：
    获取第一目标对象的多幅图像的特征、所述第一目标对象的标准特征、第二目标对象的多幅图像的特征、以及所述第二目标对象的标准特征；
    根据所述第一目标对象的多幅图像的特征与所述第一目标对象的标准特征之间的相似度，在所述第一目标对象的多幅图像中确定所述第一目标对象的可信图像，以及，根据所述第二目标对象的多幅图像的特征与所述第二目标对象的标准特征之间的相似度，在所述第二目标对象的多幅图像中确定所述第二目标对象的可信图像；其中，所述第一目标对象的可信图像的特征与其标准特征之间的相似度满足第一预设相似度要求，所述第二目标对象的可信图像的特征与其标准特征之间的相似度满足第二预设相似度要求；
    根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似。
  12. 根据权利要求11所述的方法，其特征在于，所述第一预设相似度要求和所述第二预设相似度要求相同，或者，所述第一预设相似度要求和所述第二预设相似度要求不同。
  13. 一种图像处理装置，其特征在于，包括：
    特征获取模块，用于获取目标对象的多幅图像的特征及所述目标对象的标准特征；
    图像筛选模块，用于根据所述多幅图像的特征与所述标准特征之间的相似度，在所述多幅图像中确定所述目标对象的可信图像；其中，所述可信图像的特征与所述标准特征之间的相似度满足预设相似度要求。
  14. 一种图像处理装置，其特征在于，包括：
    图像获取模块，用于获取第一目标对象的可信图像和第二目标对象的可信图像；
    图像比较模块，用于根据所述第一目标对象的可信图像与所述第二目标对 象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似；
    其中，所述第一目标对象的可信图像为在所述第一目标对象的多幅图像中确定的图像，所述第一目标对象的可信图像的特征与所述第一目标对象的标准特征之间的相似度满足第一预设相似度要求；所述第二目标对象的可信图像为在所述第二目标对象的多幅图像中确定的图像，所述第二目标对象的可信图像的特征与所述第二目标对象的标准特征之间的相似度满足第二预设相似度要求。
  15. 一种图像处理装置，其特征在于，包括：
    数据获取模块，用于获取第一目标对象的多幅图像的特征、所述第一目标对象的标准特征、第二目标对象的多幅图像的特征、以及所述第二目标对象的标准特征；
    图像确定模块，用于根据所述第一目标对象的多幅图像的特征与所述第一目标对象的标准特征之间的相似度，在所述第一目标对象的多幅图像中确定所述第一目标对象的可信图像，以及，根据所述第二目标对象的多幅图像的特征与所述第二目标对象的标准特征之间的相似度，在所述第二目标对象的多幅图像中确定所述第二目标对象的可信图像；其中，所述第一目标对象的可信图像的特征与其标准特征之间的相似度满足第一预设相似度要求，所述第二目标对象的可信图像的特征与其标准特征之间的相似度满足第二预设相似度要求；
    图像判断模块，用于根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似。
  16. 一种图像处理设备，其特征在于，包括：
    处理器；以及
    被安排成存储计算机可执行指令的存储器，所述可执行指令在被执行时使所述处理器：
    获取目标对象的多幅图像的特征及所述目标对象的标准特征；
    根据所述多幅图像的特征与所述标准特征之间的相似度，在所述多幅图像中确定所述目标对象的可信图像；其中，所述可信图像的特征与所述标准特征之间的相似度满足预设相似度要求。
  17. 一种图像处理设备，其特征在于，包括：
    处理器；以及
    被安排成存储计算机可执行指令的存储器，所述可执行指令在被执行时使 所述处理器：
    获取第一目标对象的可信图像和第二目标对象的可信图像；
    根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似；
    其中，所述第一目标对象的可信图像为在所述第一目标对象的多幅图像中确定的图像，所述第一目标对象的可信图像的特征与所述第一目标对象的标准特征之间的相似度满足第一预设相似度要求；所述第二目标对象的可信图像为在所述第二目标对象的多幅图像中确定的图像，所述第二目标对象的可信图像的特征与所述第二目标对象的标准特征之间的相似度满足第二预设相似度要求。
  18. 一种图像处理设备，其特征在于，包括：
    处理器；以及
    被安排成存储计算机可执行指令的存储器，所述可执行指令在被执行时使所述处理器：
    获取第一目标对象的多幅图像的特征、所述第一目标对象的标准特征、第二目标对象的多幅图像的特征、以及所述第二目标对象的标准特征；
    根据所述第一目标对象的多幅图像的特征与所述第一目标对象的标准特征之间的相似度，在所述第一目标对象的多幅图像中确定所述第一目标对象的可信图像，以及，根据所述第二目标对象的多幅图像的特征与所述第二目标对象的标准特征之间的相似度，在所述第二目标对象的多幅图像中确定所述第二目标对象的可信图像；其中，所述第一目标对象的可信图像的特征与其标准特征之间的相似度满足第一预设相似度要求，所述第二目标对象的可信图像的特征与其标准特征之间的相似度满足第二预设相似度要求；
    根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似。
  19. 一种存储介质，用于存储计算机可执行指令，其特征在于，所述可执行指令在被执行时实现以下流程：
    获取目标对象的多幅图像的特征及所述目标对象的标准特征；
    根据所述多幅图像的特征与所述标准特征之间的相似度，在所述多幅图像中确定所述目标对象的可信图像；其中，所述可信图像的特征与所述标准特征之间的相似度满足预设相似度要求。
  20. 一种存储介质，用于存储计算机可执行指令，其特征在于，所述可执行指令在被执行时实现以下流程：
    获取第一目标对象的可信图像和第二目标对象的可信图像；
    根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似；
    其中，所述第一目标对象的可信图像为在所述第一目标对象的多幅图像中确定的图像，所述第一目标对象的可信图像的特征与所述第一目标对象的标准特征之间的相似度满足第一预设相似度要求；所述第二目标对象的可信图像为在所述第二目标对象的多幅图像中确定的图像，所述第二目标对象的可信图像的特征与所述第二目标对象的标准特征之间的相似度满足第二预设相似度要求。
  21. 一种存储介质，用于存储计算机可执行指令，其特征在于，所述可执行指令在被执行时实现以下流程：
    获取第一目标对象的多幅图像的特征、所述第一目标对象的标准特征、第二目标对象的多幅图像的特征、以及所述第二目标对象的标准特征；
    根据所述第一目标对象的多幅图像的特征与所述第一目标对象的标准特征之间的相似度，在所述第一目标对象的多幅图像中确定所述第一目标对象的可信图像，以及，根据所述第二目标对象的多幅图像的特征与所述第二目标对象的标准特征之间的相似度，在所述第二目标对象的多幅图像中确定所述第二目标对象的可信图像；其中，所述第一目标对象的可信图像的特征与其标准特征之间的相似度满足第一预设相似度要求，所述第二目标对象的可信图像的特征与其标准特征之间的相似度满足第二预设相似度要求；
    根据所述第一目标对象的可信图像与所述第二目标对象的可信图像，确定所述第一目标对象与所述第二目标对象是否相似。
PCT/CN2018/096278 2017-07-20 2018-07-19 图像处理方法及装置 WO2019015645A1 (zh)

Priority Applications (8)

Application Number Priority Date Filing Date Title
SG11201907444YA SG11201907444YA (en) 2017-07-20 2018-07-19 Imaging processing method and apparatus
EP18835563.0A EP3579146A4 (en) 2017-07-20 2018-07-19 IMAGING PROCESSING METHOD AND DEVICE
JP2019545984A JP6945639B2 (ja) 2017-07-20 2018-07-19 画像処理方法及び装置
KR1020197025770A KR102316230B1 (ko) 2017-07-20 2018-07-19 이미지 처리 방법 및 장치
MYPI2019004655A MY201739A (en) 2017-07-20 2018-07-19 Image processing method and apparatus
PH12019501920A PH12019501920A1 (en) 2017-07-20 2019-08-19 Image processing method and apparatus
US16/577,191 US11093792B2 (en) 2017-07-20 2019-09-20 Image processing methods and devices
US16/777,696 US10769490B2 (en) 2017-07-20 2020-01-30 Image processing methods and devices

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710594226.2A CN107516105B (zh) 2017-07-20 2017-07-20 图像处理方法及装置
CN201710594226.2 2017-07-20

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US16/577,191 Continuation US11093792B2 (en) 2017-07-20 2019-09-20 Image processing methods and devices

Publications (1)

Publication Number Publication Date
WO2019015645A1 true WO2019015645A1 (zh) 2019-01-24

Family

ID=60721638

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/096278 WO2019015645A1 (zh) 2017-07-20 2018-07-19 图像处理方法及装置

Country Status (10)

Country Link
US (2) US11093792B2 (zh)
EP (1) EP3579146A4 (zh)
JP (1) JP6945639B2 (zh)
KR (1) KR102316230B1 (zh)
CN (1) CN107516105B (zh)
MY (1) MY201739A (zh)
PH (1) PH12019501920A1 (zh)
SG (1) SG11201907444YA (zh)
TW (1) TWI736765B (zh)
WO (1) WO2019015645A1 (zh)

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107516105B (zh) * 2017-07-20 2020-06-16 阿里巴巴集团控股有限公司 图像处理方法及装置
CN108346139A (zh) * 2018-01-09 2018-07-31 阿里巴巴集团控股有限公司 一种图像筛选方法及装置
CN110490026A (zh) * 2018-05-14 2019-11-22 阿里巴巴集团控股有限公司 验证对象的方法、装置和系统
CN108763570A (zh) * 2018-06-05 2018-11-06 北京拓世寰宇网络技术有限公司 一种识别相同房源的方法及装置
CN109726756A (zh) * 2018-12-25 2019-05-07 北京旷视科技有限公司 图像处理方法、装置、电子设备及存储介质
CN109933680A (zh) * 2019-03-19 2019-06-25 北京大学深圳医院 图像数据筛选方法、装置、计算机设备及其存储介质
CN111046933B (zh) * 2019-12-03 2024-03-05 东软集团股份有限公司 图像分类方法、装置、存储介质及电子设备
CN112906798A (zh) * 2021-02-25 2021-06-04 云从科技集团股份有限公司 图像匹配方法、装置及计算机存储介质

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2003162689A (ja) * 2001-11-27 2003-06-06 Fujitsu Ltd 類似文字認識プログラムおよび類似文字認識方法
CN105069426A (zh) * 2015-07-31 2015-11-18 小米科技有限责任公司 相似图片判断方法以及装置
CN105117692A (zh) * 2015-08-05 2015-12-02 福州瑞芯微电子股份有限公司 一种基于深度学习的实时人脸识别方法及系统
CN105320954A (zh) * 2014-07-30 2016-02-10 北京三星通信技术研究有限公司 人脸认证装置和方法
CN105678778A (zh) * 2016-01-13 2016-06-15 北京大学深圳研究生院 一种图像匹配方法和装置
CN106339695A (zh) * 2016-09-20 2017-01-18 北京小米移动软件有限公司 人脸相似检测方法、装置及终端
CN106560840A (zh) * 2015-09-30 2017-04-12 腾讯科技（深圳）有限公司 一种图像信息识别处理方法及装置
CN106875419A (zh) * 2016-12-29 2017-06-20 北京理工雷科电子信息技术有限公司 基于ncc匹配帧差的弱小动目标跟踪丢失重检方法
CN107516105A (zh) * 2017-07-20 2017-12-26 阿里巴巴集团控股有限公司 图像处理方法及装置

Family Cites Families (25)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1722331B1 (en) * 2004-03-03 2010-12-01 NEC Corporation Image similarity calculation system, image search system, image similarity calculation method, and image similarity calculation program
FR2884007A1 (fr) * 2005-03-29 2006-10-06 France Telecom Procede d'identification de visages a partir d'images de visage, dispositif et programme d'ordinateur correspondants
GB2431793B (en) * 2005-10-31 2011-04-27 Sony Uk Ltd Image processing
KR100745981B1 (ko) * 2006-01-13 2007-08-06 삼성전자주식회사 보상적 특징에 기반한 확장형 얼굴 인식 방법 및 장치
US10916043B2 (en) * 2007-11-26 2021-02-09 Fraunhofer-Gesellschaft Zur Foerderung Der Angewandten Forschung E.V. Apparatus, method and computer program for generating a template for arranging at least one object at at least one place
US8509482B2 (en) * 2009-12-21 2013-08-13 Canon Kabushiki Kaisha Subject tracking apparatus, subject region extraction apparatus, and control methods therefor
US9116924B2 (en) * 2013-01-14 2015-08-25 Xerox Corporation System and method for image selection using multivariate time series analysis
IL226219A (en) * 2013-05-07 2016-10-31 Picscout (Israel) Ltd Efficient comparison of images for large groups of images
EP2808828B1 (en) * 2013-05-31 2020-08-05 Omron Corporation Image matching method, image matching device, model template generation method, model template generation device, and program
WO2015009968A2 (en) * 2013-07-19 2015-01-22 Google Inc. Face template balancing
US9053365B2 (en) * 2013-09-16 2015-06-09 EyeVerify, Inc. Template update for biometric authentication
WO2015164584A1 (en) * 2014-04-23 2015-10-29 Google Inc. User interface control using gaze tracking
US9330329B2 (en) * 2014-07-09 2016-05-03 Ditto Labs, Inc. Systems, methods, and devices for image matching and object recognition in images using minimal feature points
US10210427B2 (en) * 2014-07-09 2019-02-19 Slyce Acquisition Inc. Systems, methods, and devices for image matching and object recognition in images
US9846948B2 (en) * 2014-07-09 2017-12-19 Ditto Labs, Inc. Systems, methods, and devices for image matching and object recognition in images using feature point optimization
US9317921B2 (en) * 2014-07-10 2016-04-19 Qualcomm Incorporated Speed-up template matching using peripheral information
CN105450411B (zh) * 2014-08-14 2019-01-08 阿里巴巴集团控股有限公司 利用卡片特征进行身份验证的方法、装置及系统
US9424470B1 (en) * 2014-08-22 2016-08-23 Google Inc. Systems and methods for scale invariant 3D object detection leveraging processor architecture
US9652688B2 (en) * 2014-11-26 2017-05-16 Captricity, Inc. Analyzing content of digital images
US10496695B2 (en) * 2016-06-06 2019-12-03 Think-Cell Software Gmbh Automated data extraction from scatter plot images
JP6723909B2 (ja) * 2016-12-09 2020-07-15 キヤノン株式会社 画像処理方法、画像処理装置、及びプログラム
EP3682366A1 (en) * 2017-10-27 2020-07-22 Koninklijke Philips N.V. Camera and image calibration for subject identification
CN108229314B (zh) * 2017-11-28 2021-05-04 深圳市商汤科技有限公司 目标人物的搜索方法、装置和电子设备
CN111125391A (zh) * 2018-11-01 2020-05-08 北京市商汤科技开发有限公司 数据库更新方法和装置、电子设备、计算机存储介质
WO2020087950A1 (zh) * 2018-11-01 2020-05-07 北京市商汤科技开发有限公司 数据库更新方法和装置、电子设备、计算机存储介质

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2003162689A (ja) * 2001-11-27 2003-06-06 Fujitsu Ltd 類似文字認識プログラムおよび類似文字認識方法
CN105320954A (zh) * 2014-07-30 2016-02-10 北京三星通信技术研究有限公司 人脸认证装置和方法
CN105069426A (zh) * 2015-07-31 2015-11-18 小米科技有限责任公司 相似图片判断方法以及装置
CN105117692A (zh) * 2015-08-05 2015-12-02 福州瑞芯微电子股份有限公司 一种基于深度学习的实时人脸识别方法及系统
CN106560840A (zh) * 2015-09-30 2017-04-12 腾讯科技（深圳）有限公司 一种图像信息识别处理方法及装置
CN105678778A (zh) * 2016-01-13 2016-06-15 北京大学深圳研究生院 一种图像匹配方法和装置
CN106339695A (zh) * 2016-09-20 2017-01-18 北京小米移动软件有限公司 人脸相似检测方法、装置及终端
CN106875419A (zh) * 2016-12-29 2017-06-20 北京理工雷科电子信息技术有限公司 基于ncc匹配帧差的弱小动目标跟踪丢失重检方法
CN107516105A (zh) * 2017-07-20 2017-12-26 阿里巴巴集团控股有限公司 图像处理方法及装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3579146A4 *

Also Published As

Publication number Publication date
MY201739A (en) 2024-03-15
EP3579146A1 (en) 2019-12-11
KR102316230B1 (ko) 2021-10-26
US20200012885A1 (en) 2020-01-09
CN107516105B (zh) 2020-06-16
TWI736765B (zh) 2021-08-21
KR20200008993A (ko) 2020-01-29
PH12019501920A1 (en) 2020-06-01
US20200167599A1 (en) 2020-05-28
US11093792B2 (en) 2021-08-17
US10769490B2 (en) 2020-09-08
SG11201907444YA (en) 2019-09-27
TW201909034A (zh) 2019-03-01
JP2020523651A (ja) 2020-08-06
CN107516105A (zh) 2017-12-26
JP6945639B2 (ja) 2021-10-06
EP3579146A4 (en) 2020-04-22

Similar Documents

Publication Publication Date Title
WO2019015645A1 (zh) 图像处理方法及装置
US11244199B2 (en) User identity determining method, apparatus, and device
WO2019137216A1 (zh) 一种图像筛选方法及装置
WO2017063530A1 (zh) 动作信息识别方法和系统
TWI767675B (zh) 隱私保護下的用戶識別方法、裝置及設備
WO2021114916A1 (zh) 风险检测方法、装置及设备
TWI754818B (zh) 支付方法、裝置及系統
CN104050449A (zh) 一种人脸识别方法及装置
CN109086734B (zh) 一种对人眼图像中瞳孔图像进行定位的方法及装置
Lee et al. Place recognition using straight lines for vision-based SLAM
US9691180B2 (en) Determination of augmented reality information
CN108664526B (zh) 检索的方法和设备
CN103207898A (zh) 一种基于局部敏感哈希的相似人脸快速检索方法
WO2021184976A1 (zh) 一种用于隐私保护的用户特征提取系统及设备
WO2016015621A1 (zh) 人脸图片人名识别方法和系统
US10614312B2 (en) Method and apparatus for determining signature actor and identifying video based on probability of appearance of signature actor
WO2018095307A1 (zh) 一种评价信息的发布方法及装置
CN109447022A (zh) 一种镜头类型识别方法及装置
US10410045B2 (en) Automated facial recognition systems and methods
WO2019174392A1 (zh) 针对rpc信息的向量处理
CN111143626A (zh) 团伙识别方法、装置、设备及计算机可读存储介质
CN113128277A (zh) 一种人脸关键点检测模型的生成方法及相关设备
CN116188919B (zh) 一种测试方法、装置、可读存储介质及电子设备
CN105701459B (zh) 一种图片显示方法及终端设备
Wu et al. A novel image retrieval approach with Bag-of-Word model and Gabor feature

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18835563

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2019545984

Country of ref document: JP

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 20197025770

Country of ref document: KR

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 2018835563

Country of ref document: EP

Effective date: 20190903

NENP Non-entry into the national phase

Ref country code: DE