WO2016101766A1 - 相似人脸图片获取和人脸图片信息获取方法和装置 - Google Patents
相似人脸图片获取和人脸图片信息获取方法和装置 Download PDFInfo
- Publication number
- WO2016101766A1 WO2016101766A1 PCT/CN2015/095861 CN2015095861W WO2016101766A1 WO 2016101766 A1 WO2016101766 A1 WO 2016101766A1 CN 2015095861 W CN2015095861 W CN 2015095861W WO 2016101766 A1 WO2016101766 A1 WO 2016101766A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- face
- face image
- name
- person
- similar
- Prior art date
- Legal status (The legal status 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 status listed.)
- Ceased
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/53—Querying
- G06F16/532—Query formulation, e.g. graphical querying
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
- G06F16/5838—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using colour
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/30—Scenes; Scene-specific elements in albums, collections or shared content, e.g. social network photos or video
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/70—Labelling scene content, e.g. deriving syntactic or semantic representations
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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
- G06V40/161—Detection; Localisation; Normalisation
- G06V40/165—Detection; Localisation; Normalisation using facial parts and geometric relationships
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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 OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/14—Digital output to display device ; Cooperation and interconnection of the display device with other functional units
Definitions
- the present invention relates to the field of Internet technologies, and in particular, to a similar face image acquisition method and apparatus, and a method and device for acquiring face image information.
- the present invention has been made in order to provide a similar face image acquisition method and apparatus that overcomes the above problems or at least partially solves the above problems, and a face image information acquisition method and apparatus.
- a method for acquiring a similar face image includes: acquiring a face image specified by a user; performing face recognition on the face image to identify from the collected face image a similar face picture of the face picture; displaying the similar face picture to the user.
- a similar face image obtaining apparatus comprising: a face image obtaining module, configured to acquire a face image specified by a user; and a face image recognition module, configured to The face image performs face recognition to identify a similar face picture of the face picture from the collected face picture; a face picture display module is configured to display the similar face picture to the user.
- a method for acquiring face image information includes: acquiring a face image specified by a user; performing face recognition on the face image to obtain a person in the face image a person name of the face; the network resource information corresponding to the person name is obtained from the network according to the name of the person; and the network resource information corresponding to the person name is displayed to the user.
- a face image information acquiring apparatus comprising: a face image obtaining module, configured to acquire a face image specified by a user; and a person name obtaining module, configured to face the face The image is subjected to face recognition to obtain the name of the face in the face image; the information acquisition module is configured to acquire network resource information corresponding to the person name from the network according to the name of the person; and an information display module, configured to: The network resource information corresponding to the person name is displayed to the user.
- a computer program comprising computer readable code that, when executed on a computing device, causes the computing device to perform a similar face picture as described above Obtaining a method, or causing the computing device to perform the face image information acquisition method described above.
- a computer readable medium wherein the computer program described above is stored.
- the user is provided with a similar face image according to the face image specified by the user; the user specifies that the face in the picture has a great similarity with the face in the similar picture, and the user needs to specify the face.
- the technical solution of the present invention can satisfy the needs of the user.
- FIG. 1 shows a flow chart of a similar face image acquisition method in accordance with one embodiment of the present invention
- FIG. 2 shows a flow chart of a similar face image acquisition method in accordance with one embodiment of the present invention
- FIG. 3 is a flow chart showing a similar face image acquisition method according to an embodiment of the present invention.
- FIG. 4 shows a flow chart of a similar face image acquisition method in accordance with one embodiment of the present invention
- FIG. 5 is a flowchart showing a similar face image acquisition method according to an embodiment of the present invention.
- FIG. 6 is a block diagram showing a similar face image acquiring apparatus according to an embodiment of the present invention.
- FIG. 7 is a block diagram showing a similar face image acquiring apparatus according to an embodiment of the present invention.
- FIG. 8 is a block diagram showing a similar face image acquiring apparatus according to an embodiment of the present invention.
- FIG. 9 is a block diagram showing a similar face image acquiring apparatus according to an embodiment of the present invention.
- FIG. 10 is a block diagram showing a similar face image acquiring apparatus according to an embodiment of the present invention.
- FIG. 11 is a flowchart showing a method for acquiring face picture information according to an embodiment of the present invention.
- FIG. 12 is a flowchart showing a method for acquiring face picture information according to an embodiment of the present invention.
- FIG. 13 is a flowchart showing a method for acquiring face picture information according to an embodiment of the present invention.
- FIG. 14 is a flowchart showing a method for acquiring face picture information according to an embodiment of the present invention.
- FIG. 15 is a flowchart showing a method for acquiring face picture information according to an embodiment of the present invention.
- FIG. 16 is a flowchart showing a method of acquiring face picture information according to an embodiment of the present invention.
- FIG. 17 is a block diagram showing a face picture information acquiring apparatus according to an embodiment of the present invention.
- FIG. 18 is a block diagram showing a face picture information acquiring apparatus according to an embodiment of the present invention.
- FIG. 19 is a block diagram showing a face picture information acquiring apparatus according to an embodiment of the present invention.
- FIG. 20 is a block diagram showing a face picture information acquiring apparatus according to an embodiment of the present invention.
- Figure 21 schematically shows a block diagram of a computing device for performing a method in accordance with the present invention
- Fig. 22 schematically shows a storage unit for holding or carrying program code implementing the method according to the invention.
- an embodiment of the present invention provides a method for acquiring a similar face image, including:
- Step 110 Obtain a face image specified by the user.
- Step 120 Perform face recognition on the face image to identify a similar face image of the face image from the collected face image.
- the acquisition of the person's name can be completed.
- step 130 a similar face picture is displayed to the user.
- the user is provided with a similar face image according to the face image specified by the user; then the face specified by the user in the picture has a great similarity with the face in the similar picture, and the user needs to specify
- the technical solution of the embodiment can meet the needs of the user.
- the user inputs the photo of the star A, and needs to find a photo of other characters similar to the star A style; the star recognition A in the picture is recognized based on the face recognition technology, and is found from the preset face image library.
- the photo of the star C similar to the star A face is conducive to increasing the user's understanding of the star C.
- an embodiment of the present invention provides a method for acquiring a similar face image, which includes:
- Step 210 Acquire a face picture specified by the user.
- Step 220 Perform face recognition on the face image to identify a similar face image of the face image from the collected face image.
- Step 230 Obtain a person name of a face in a similar face picture.
- Step 240 Acquire a similarity between the similar face image and the face image, and accumulate the similarity corresponding to all similar face images having the same person name.
- performing the accumulation of the similarity is equivalent to calculating the similarity between one or more face images of each character and the original face image based on the person name, and then adding the similar images is advantageous for finding the similarity. The highest person.
- Step 250 When the accumulated maximum similarity is less than a predetermined threshold, select one or more facial images from the similar facial images to display to the user according to the similarity degree corresponding to the similar facial image.
- the maximum similarity after the accumulation exceeds a certain threshold, it indicates that the face image of the same person has been queried, and the face image of the same person is preferentially provided to the user to indicate that the face is accurately recognized. character.
- the user inputs the photo of star A, and obtains a similar face picture as follows: photo a is a photo of star C, the similarity is 80%; photo b is a photo of star D, the similarity is 70%; photo c is a star Photograph of C, the similarity is 70%; photo d is the photo of star D, the similarity is 55%; after the accumulation, the similarity of star C is 135%, the similarity of star D is 140%; the preset threshold is 90 % ⁇ n,n is the number of photos corresponding to each person's name, and the thresholds for the stars C and D are both 180%, indicating that there is no face image of the same person; at this time, the photo corresponding to the star D can be output, indicating Star D is a similar face image of star A.
- An embodiment of the present invention provides a similar face image acquisition method, including:
- Step 210 Acquire a face picture specified by the user.
- Step 220 Perform face recognition on the face image to identify a similar face image of the face image from the collected face image.
- Step 230 Obtain a person name of a face in a similar face picture.
- Step 240 Acquire a similarity between the similar face image and the face image, and accumulate the similarity corresponding to all similar face images having the same person name.
- Step 250 When the accumulated maximum similarity is less than a predetermined threshold, select one or more facial images from the similar facial images to display to the user according to the similarity degree corresponding to the similar facial image.
- Step 251 Display the similarity corresponding to one or more face pictures and/or the name of the person whose face is displayed to the user.
- the user prompts for similarity and person name, so that the user has more understanding of similar face images.
- the user inputs a photo of the star A
- the similar face image obtained according to the technical solution of the embodiment is a photo of the star C
- the similarity is 80% on the photo of the star C
- the person name is C. .
- An embodiment of the present invention provides a similar face image acquisition method, including:
- Step 210 Acquire a face picture specified by the user.
- Step 220 Perform face recognition on the face image to identify a similar face image of the face image from the collected face image.
- Step 230 Obtain a person name of a face in a similar face picture.
- Step 240 Acquire a similarity between the similar face image and the face image, and accumulate the similarity corresponding to all similar face images having the same person name.
- Step 250 When the accumulated maximum similarity is less than a predetermined threshold, select one or more facial images from the similar facial images to display to the user according to the similarity degree corresponding to the similar facial image.
- Step 252 Set a display order of one or more face pictures according to the similarity level of the one or more face pictures. In this embodiment, it is advantageous to provide the most similar photo first to the user.
- the user inputs a photo of the star A.
- the similar face image obtained according to the technical solution of the embodiment is a photo of the star C and a photo of the star D, and the similarities are 80% and 85%, respectively.
- a photo of the star D is displayed, and then a photo of the star C is displayed.
- An embodiment of the present invention provides a similar face image acquisition method, including:
- Step 210 Acquire a face picture specified by the user.
- Step 220 Perform face recognition on the face image to identify a similar face image of the face image from the collected face image.
- Step 230 Obtain a person name of a face in a similar face picture.
- Step 240 Acquire a similarity between the similar face image and the face image, and accumulate the similarity corresponding to all similar face images having the same person name.
- Step 250 When the accumulated maximum similarity is less than a predetermined threshold, select one or more facial images from the similar facial images to display to the user according to the similarity degree corresponding to the similar facial image.
- Step 253 Set corresponding evaluation information for one or more facial images according to the similarity level of the one or more facial images, and display the corresponding evaluation information to the user.
- the evaluation information can reflect the level of similarity in a manner that is easier for the user to understand.
- the user inputs a photo of the star A, and the similar face image obtained according to the technical solution of the embodiment is a photo of the star C, and the similarity is 80%, respectively, and the evaluation information is displayed on the photo as: sisters".
- an embodiment of the present invention provides a method for acquiring a similar face image, which includes:
- Step 310 Acquire a face image specified by the user.
- Step 320 Perform face recognition on the face image to identify a similar face image of the face image from the collected face image.
- Step 330 Obtain a person name of a face in a similar face picture.
- Step 340 Acquire the similarity between the similar face image and the face image, and accumulate the similarity corresponding to all similar face images having the same person name.
- Step 350 When the maximum similarity after the accumulation is less than a predetermined threshold, the face image is obtained in the order of similarity corresponding to the similar face image.
- Step 360 Determine whether the name of the face in the newly acquired face image is located in the name of the face in the other acquired face image.
- Step 370 When the determination result is no, the newly acquired face image is displayed to the user. According to the technical solution of the embodiment, it is possible to avoid occurrence of similar face pictures of duplicate characters, so as to ensure that the user sees similar face pictures of a plurality of characters as much as possible.
- the user inputs a photo of star A, and obtains a similar face picture as follows: photo a is a photo of star C, the similarity is 80%; photo b is a photo of star D, the similarity is 70%; photo c is a star C's photo, similarity is 70%; photo d is a photo of star D, with a similarity of 55%. Then, according to the similarity level, the photo a is first displayed; when the photo b is obtained, and the photo of the star D is not displayed, the photo b is displayed; when the photo c is obtained, and the photo of the star C is displayed, the display is discarded; If the photo of the star D has been displayed, the display is abandoned.
- an embodiment of the present invention provides a method for acquiring a similar face image, including:
- Step 410 Acquire a face image specified by the user.
- Step 420 extracting features of the face image, and extracting features of the collected face image.
- the face image specified by the user may be pre-processed and normalized in advance to facilitate feature extraction; in this embodiment, the sample face image may be collected and targeted to the skin color, eyes, nose, and mouth.
- the detected data can train the face model, and the face model can identify the position of the face in the user-specified picture and perform feature extraction.
- step 430 the features of the face image are compared with the features of the collected face image.
- Step 440 Select a similar face picture from the collected face picture according to the comparison result.
- Step 450 Obtain a person name of a face in a similar face picture.
- Step 460 Calculate the similarity between the similar face image and the face image according to the comparison result, and accumulate the similarity corresponding to all similar face images having the same person name.
- Step 470 When the accumulated maximum similarity is less than a predetermined threshold, select one or more facial images from the similar facial images to display to the user according to the similarity degree corresponding to the similar facial image.
- the user specifies a certain face belonging to the database in which the name has been established (the collected face image): First, a person with a known name is established by face detection, feature extraction, and person name extraction. Face database; for the new face image specified by the user, face detection is performed on the picture. If there is no face, it returns directly. If there is a face, the face feature is extracted and quantized into a high-dimensional vector. The vector of the input picture is compared with all the face feature high-dimensional vectors in the library, and the Euclidean distance is calculated, and the first N vectors closest to each other are taken. The faces represented by these vectors are the faces most similar to the input face.
- the face database is too large, it takes a long time to compare one by one. You can cluster the faces in the library in advance, and then compare them with the faces of the clusters. For the first N similar faces, the similarity is used as the weight. Calculate the weight of each name, and add the weights of the same name. And find the name with the highest weight. If the name is greater than a certain threshold, it is considered that the input face belongs to the face corresponding to the name, otherwise it is considered that the face cannot be accurately recognized.
- an embodiment of the present invention provides a similar face image acquisition method, including:
- Step 510 Acquire a face image specified by the user.
- Step 520 Perform face recognition on the face image to identify a similar face image of the face image from the collected face image.
- Step 530 Extract one or more person names from the text corresponding to the similar face image.
- the text type is not limited, and may be a title of the news in which the picture is located, a surrounding text, or the like.
- Step 540 Calculate a weight value for one or more person names according to attributes of one or more person names.
- the attribute is not limited, which may be the frequency, location, and the like of the name of the person, because the names of the people of different frequencies and positions are similar to the possibility of the person name corresponding to the face picture.
- Step 550 Select a person name of a face in a similar face picture from one or more person names according to the level of the weight value.
- Step 560 Acquire similarity between the similar face image and the face image, and accumulate the similarity corresponding to all similar face images having the same person name.
- Step 570 When the accumulated maximum similarity is less than a predetermined threshold, select one or more facial images from the similar facial images to display to the user according to the similarity degree corresponding to the similar facial image.
- an embodiment of the present invention provides a similar face image obtaining apparatus, including:
- the face image obtaining module 610 acquires a face image specified by the user.
- the face picture recognition module 620 performs face recognition on the face picture to identify a similar face picture of the face picture from the collected face picture. In this embodiment, based on the existing face recognition technology, the acquisition of the person's name can be completed.
- the face picture display module 630 displays similar face pictures to the user.
- the user is provided with a similar face image according to the face image specified by the user; then the face specified by the user in the picture has a great similarity with the face in the similar picture, and the user needs to specify
- the technical solution of the embodiment can meet the needs of the user.
- the user inputs a photo of the star A, and needs to find a photo of another character similar to the star A style; the star recognition A in the image is recognized based on the face recognition technology, and is found from the preset face image library.
- the photo of the star C similar to the star A face is conducive to increasing the user's understanding of the star C.
- an embodiment of the present invention provides a similar face image obtaining apparatus, including:
- the face image obtaining module 710 acquires a face image specified by the user.
- the face picture recognition module 720 performs face recognition on the face picture to identify a similar face picture of the face picture from the collected face picture.
- the person name obtaining module 730 acquires a person name of a face in a similar face picture.
- the similarity obtaining module 740 acquires the similarity between the similar face image and the face image, and accumulates the similarity corresponding to all similar face images having the same person name.
- performing the accumulation of the similarity is equivalent to calculating the similarity between one or more face images of each character and the original face image based on the person name, and then adding the similar images is advantageous for finding the similarity. The highest person.
- the face picture display module 750 selects one or more face pictures from the similar face pictures to display to the user according to the similarity degree corresponding to the similar face picture when the maximum similarity after the accumulation is less than the predetermined threshold.
- the maximum similarity after the accumulation exceeds a certain threshold, it indicates that the face image of the identical person has been queried, and the face image of the same person is preferentially provided to the user to indicate the accurate recognition. The character is out.
- a photo of star C is a photo of star C, the similarity is 80%; photo b is a photo of star D, the similarity is 70%; photo c is a star Photograph of C, the similarity is 70%; photo d is the photo of star D, the similarity is 55%; after the accumulation, the similarity of star C is 135%, the similarity of star D is 140%; the preset threshold is 90 % ⁇ n,n is the number of photos corresponding to each person's name, and the thresholds for the stars C and D are both 180%, indicating that there is no face image of the same person; at this time, the photo corresponding to the star D can be output, indicating Star D is a similar face image of star A.
- An embodiment of the present invention provides a similar face image obtaining apparatus, including:
- the face image obtaining module 710 acquires a face image specified by the user.
- the face picture recognition module 720 performs face recognition on the face picture to identify a similar face picture of the face picture from the collected face picture.
- the person name obtaining module 730 acquires a person name of a face in a similar face picture.
- the similarity obtaining module 740 acquires the similarity between the similar face image and the face image, and accumulates the similarity corresponding to all similar face images having the same person name.
- the face picture display module 750 selects one or more face pictures from the similar face pictures to display to the user according to the similarity degree corresponding to the similar face picture when the maximum similarity after the accumulation is less than the predetermined threshold.
- the similarity/person name display module 751 displays the similarity corresponding to one or more face pictures and/or the name of the person whose face is displayed to the user. In this embodiment, the user is prompted for the similarity and the name of the person, so that the user has more understanding of the similar face picture.
- the user inputs a photo of the star A, and the similar face image acquired according to the technical solution of the embodiment is a photo of the star C, and the similarity is 80% on the photo of the star C, and the person name is C. .
- An embodiment of the present invention provides a similar face image obtaining apparatus, including:
- the face image obtaining module 710 acquires a face image specified by the user.
- the face picture recognition module 720 performs face recognition on the face picture to identify a similar face picture of the face picture from the collected face picture.
- the person name obtaining module 730 acquires a person name of a face in a similar face picture.
- the similarity obtaining module 740 acquires the similarity between the similar face image and the face image, and accumulates the similarity corresponding to all similar face images having the same person name.
- the face picture display module 750 selects one or more face pictures from the similar face pictures to display to the user according to the similarity degree corresponding to the similar face picture when the maximum similarity after the accumulation is less than the predetermined threshold.
- the display order setting module 752 sets the display order of one or more face pictures according to the similarity level corresponding to one or more face pictures. In this embodiment, it is advantageous to provide the most similar photo first to the user.
- the user inputs a photo of the star A.
- the similar face image obtained according to the technical solution of the embodiment is a photo of the star C and a photo of the star D, and the similarities are 80% and 85%, respectively.
- a photo of the star D is displayed, and then a photo of the star C is displayed.
- An embodiment of the present invention provides a similar face image obtaining apparatus, including:
- the face image obtaining module 710 acquires a face image specified by the user.
- the face picture recognition module 720 performs face recognition on the face picture to identify a similar face picture of the face picture from the collected face picture.
- the person name obtaining module 730 acquires a person name of a face in a similar face picture.
- the similarity obtaining module 740 acquires the similarity between the similar face image and the face image, and accumulates the similarity corresponding to all similar face images having the same person name.
- the face picture display module 750 selects one or more face pictures from the similar face pictures to display to the user according to the similarity degree corresponding to the similar face picture when the maximum similarity after the accumulation is less than the predetermined threshold.
- the evaluation information display module 753 sets corresponding evaluation information for one or more face images according to the degree of similarity corresponding to one or more face images, and displays the corresponding evaluation information to the user.
- the evaluation information can reflect the level of similarity in a manner that is easier for the user to understand.
- the user inputs a photo of the star A, and the similar face image obtained according to the technical solution of the embodiment is a photo of the star C, and the similarity is 80%, respectively, and the evaluation information is displayed on the photo as: sisters".
- an embodiment of the present invention provides a similar face image obtaining apparatus, including:
- the face picture obtaining module 810 acquires a face picture specified by the user.
- the face picture recognition module 820 performs face recognition on the face picture to identify a similar face picture of the face picture from the collected face picture.
- the person name obtaining module 830 acquires a person name of a face in a similar face picture.
- the similarity obtaining module 840 acquires the similarity between the similar face image and the face image, and accumulates the similarity corresponding to all similar face images having the same person name.
- the sequence obtaining module 850 acquires the face image in order of similarity corresponding to the similar face image when the maximum similarity after the accumulation is less than the predetermined threshold.
- the person name judging module 860 determines whether the person name of the face in the newly acquired face picture is located in the name of the face of the other acquired face picture.
- the face picture display module 870 displays the newly acquired face picture to the user when the determination result is no. According to the technical solution of the embodiment, it is possible to avoid occurrence of similar face pictures of duplicate characters, so as to ensure that the user sees similar face pictures of a plurality of characters as much as possible.
- the user inputs a photo of star A, and obtains a similar face picture as follows: photo a is a photo of star C, the similarity is 80%; photo b is a photo of star D, the similarity is 70%; photo c is a star C's photo, similarity is 70%; photo d is a photo of star D, with a similarity of 55%. Then, according to the similarity level, the photo a is first displayed; when the photo b is obtained, and the photo of the star D is not displayed, the photo b is displayed; when the photo c is obtained, and the photo of the star C is displayed, the display is discarded; If the photo of the star D has been displayed, the display is abandoned.
- an embodiment of the present invention provides a similar face image obtaining apparatus, including:
- the face image obtaining module 910 acquires a face image specified by the user.
- the feature extraction module 920 extracts features of the face image and extracts features of the collected face image.
- the face image specified by the user may be pre-processed and normalized in advance to facilitate feature extraction; in this embodiment, the sample face image may be collected and targeted to the skin color, eyes, nose, and mouth.
- the detected data can train the face model, and the face model can identify the position of the face in the user-specified picture and perform feature extraction.
- the feature comparison module 930 compares the features of the face image with the features of the collected face image.
- the face picture recognition module 940 selects a similar face picture from the collected face pictures according to the comparison result.
- the person name obtaining module 950 acquires a person name of a face in a similar face picture.
- the similarity obtaining module 960 calculates the similarity between the similar face image and the face image according to the comparison result, and accumulates the similarity corresponding to all similar face images having the same person name.
- the face picture display module 970 when the maximum similarity after the accumulation is less than a predetermined threshold, selects one or more face pictures from the similar face pictures to display to the user according to the similarity degree corresponding to the similar face picture.
- the user specifies a certain face belonging to the database in which the name has been established (the collected face image): First, a person with a known name is established by face detection, feature extraction, and person name extraction. Face database; for the new face image specified by the user, face detection is performed on the picture. If there is no face, it returns directly. If there is a face, the face feature is extracted and quantized into a high-dimensional vector. The vector of the input picture is compared with all the face feature high-dimensional vectors in the library, and the Euclidean distance is calculated, and the first N vectors closest to each other are taken. The faces represented by these vectors are the faces most similar to the input face.
- the face database is too large, it takes a long time to compare one by one. You can cluster the faces in the library in advance, and then compare them with the faces of the clusters. For the first N similar faces, the similarity is used as the weight. Calculate the weight of each name, and add the weights of the same name. And find the name with the highest weight. If the name is greater than a certain threshold, it is considered that the input face belongs to the face corresponding to the name, otherwise it is considered that the face cannot be accurately recognized.
- an embodiment of the present invention provides a similar face image obtaining apparatus, including:
- the face image obtaining module 1010 acquires a face image specified by the user.
- the face picture recognition module 1020 performs face recognition on the face picture to identify a similar face picture of the face picture from the collected face picture.
- the name acquisition module specifically includes:
- the person name extraction module 1030 extracts one or more person names from the text corresponding to the similar face picture.
- the text type is not limited, and may be a title of the news in which the picture is located, a surrounding text, or the like.
- the weight value calculation module 1040 calculates a weight value for one or more person names based on attributes of one or more person names.
- the attribute is not limited, which may be the frequency, location, and the like of the name of the person, because the names of the people of different frequencies and positions are similar to the possibility of the person name corresponding to the face picture.
- the person name selection module 1050 selects a person name of a face in a similar face picture from one or more person names according to the level of the weight value.
- the similarity obtaining module 1060 obtains the similarity between the similar face image and the face image, and accumulates the similarity corresponding to all similar face images having the same person name.
- the face picture display module 1070 selects one or more face pictures from the similar face pictures to display to the user according to the similarity degree corresponding to the similar face picture when the maximum similarity after the accumulation is less than the predetermined threshold.
- the present invention provides the following scheme
- an embodiment of the present invention provides a method for acquiring face image information, including:
- Step 1110 Obtain a face image specified by the user.
- Step 1120 Perform face recognition on the face image to obtain the name of the face in the face image.
- the acquisition of the person's name can be completed.
- Step 1130 Obtain network resource information corresponding to the person name from the network according to the name of the person.
- Step 1140 Display network resource information corresponding to the name of the person to the user.
- the method for face recognition accurately identifies the name of the person involved in the face picture, and associates the picture with the related information of the person name, thereby providing the user with more information.
- the person name A can be automatically recognized, and the search engine is used to search according to A as the keyword, and the related report of the star A is provided to the user, which helps to increase the user. Understanding of Star A.
- an embodiment of the present invention provides a method for acquiring face image information, including:
- Step 1210 Obtain a face image specified by the user.
- Step 1220 Perform face recognition on the face image to obtain the name of the face in the face image.
- the structured information corresponding to the person name is obtained as the network resource information corresponding to the person name from the site for recording the structured information in the network, wherein the structured information corresponding to the person name includes information corresponding to the plurality of items.
- the site for recording structured information is typically a Wikipedia website. Since the information in the Wikipedia website is structured, as long as the corresponding person name is found, it can be directly captured.
- Step 1240 Display network resource information corresponding to the name of the person to the user. According to the technical solution of the embodiment, since the distribution of multiple items in the structured information is relatively clear, it is advantageous for the user to understand the characters involved in the face image.
- the face is automatically detected and face recognition is performed, and the name of the person is recognized; then, according to the name of the person, the related structured information of the character is obtained from the Wikipedia website. , including personal profile, experience, height, weight, major works, and the latest developments on Weibo. This makes it easy for the user to know the star A, to know more information and dynamics of the star A, and to establish certain contact and interaction with the star A through Weibo.
- an embodiment of the present invention provides a method for acquiring face image information, including:
- step 1310 a face image specified by the user is obtained.
- Step 1320 Perform face recognition on the face image to obtain the name of the face in the face image.
- the acquisition of the person's name can be completed.
- Step 1330 Filter, according to the name of the person, the information corresponding to the plurality of items from the network resource information corresponding to the person name according to the preset multiple items.
- the type of the entry is not limited, for example, it may be height, weight, etc.; further, in order to timely pay attention to the character, the entry may be a social account entry; in order to timely related events of the character Understand, the entry can be an information message entry.
- Step 1340 processing the extracted information into structured information according to various entries.
- the social account entry may be used to search for the identification information of the social account corresponding to the name of the person and/or the content of the social account corresponding to the name of the person; or may be obtained from the network according to the news information item. News information corresponding to the name of the person.
- Step 1350 Display network resource information corresponding to the name of the person to the user.
- the information in the network can be organized into structured information and provided to the user; further, the social network dynamics of the related person and the latest news can be obtained according to the face image.
- the face is automatically detected and face recognition is performed, and the person name is identified; then, according to the name of the person, the microblog account of the relevant person is found, and the micro The blog account "starA" and the latest update of the account release are "very busy” available to the user; or the latest news of the disputed event of the relevant person can be found according to the name of the person, and the news title is provided to the user.
- an embodiment of the present invention provides a method for acquiring face image information, including:
- step 1410 a face image specified by the user is obtained.
- Step 1420 extracting features of the face image, and extracting features of the collected face image.
- the face image specified by the user may be pre-processed and normalized in advance to facilitate feature extraction; in this embodiment, the sample face image may be collected and targeted to the skin color, eyes, nose, and mouth.
- the detected data can train the face model, and the face model can identify the position of the face in the user-specified picture and perform feature extraction.
- step 1430 the features of the face image are compared with the features of the collected face image.
- Step 1440 Select a similar face image of the face image from the collected face images according to the comparison result.
- step 1450 the name of the face in the face image is determined according to the name of the face in the similar face picture.
- Step 1460 Obtain network resource information corresponding to the person name from the network according to the name of the person.
- Step 1470 Display network resource information corresponding to the name of the person to the user. According to the technical solution of the embodiment, it is advantageous to accurately identify the name of the person corresponding to the face picture by means of feature comparison.
- an embodiment of the present invention provides a method for acquiring face image information, including:
- Step 1510 Obtain a face image specified by the user.
- Step 1520 Extract features of the face image and extract features of the collected face image.
- step 1530 the features of the face image are compared with the features of the collected face image.
- Step 1540 Select a similar face image of the face image from the collected face images according to the comparison result.
- Step 1550 Extract one or more person names from the text corresponding to the similar face image.
- the text type is not limited, and may be a title of the news in which the picture is located, a surrounding text, or the like.
- Step 1560 Calculate weight values for one or more person names based on attributes of one or more person names.
- the attribute is not limited, which may be the frequency, location, and the like of the name of the person, because the names of the people of different frequencies and positions are similar to the possibility of the person name corresponding to the face picture.
- Step 1570 Select, according to the weight value, a person name corresponding to the face in the similar face image from one or more person names.
- Step 1580 Determine the name of the face of the face in the face image according to the name of the face in the similar face picture.
- Step 1590 Obtain network resource information corresponding to the person name from the network according to the name of the person.
- Step 15100 Display network resource information corresponding to the name of the person to the user.
- an embodiment of the present invention provides a method for acquiring face image information, including:
- Step 1610 Obtain a face image specified by the user.
- Step 1620 extract features of the face image, and extract features of the collected face image.
- step 1630 the features of the face image are compared with the features of the collected face image.
- Step 1640 Select a similar face image of the face image from the collected face images according to the comparison result.
- Step 1650 Calculate the similarity between the similar face image and the face image according to the comparison result.
- step 1660 the similarity of all similar face pictures corresponding to the same person name is accumulated.
- Step 1670 When the maximum similarity after the accumulation is greater than the preset threshold, determine the name of the face in the face image according to the name of the face in the similar face image corresponding to the maximum similarity after the accumulation. In the present embodiment, accurate recognition of a face can be achieved based on the accumulated similarity.
- Step 1680 Obtain network resource information corresponding to the person name from the network according to the name of the person.
- Step 1690 Display network resource information corresponding to the name of the person to the user.
- the user specifies a certain face belonging to the database in which the name has been established (the collected face image): First, a person with a known person name is established through face detection, feature extraction, and person name extraction. Face database; for the new face image specified by the user, face detection is performed on the picture. If there is no face, it returns directly. If there is a face, the face feature is extracted and quantized into a high-dimensional vector. The vector of the input picture is compared with all the face feature high-dimensional vectors in the library, and the Euclidean distance is calculated, and the first N vectors closest to each other are taken. The faces represented by these vectors are the faces most similar to the input face.
- the face database is too large, it takes a long time to compare one by one. You can cluster the faces in the library in advance, and then compare them with the faces of the clusters. For the first N similar faces, the similarity is used as the weight. Calculate the weight of each name, and add the weights of the same name. And find the name with the highest weight. If the name is greater than a certain threshold, it is considered that the input face belongs to the face corresponding to the name, otherwise it is considered that the face cannot be accurately recognized.
- an embodiment of the present invention provides a face image information acquiring apparatus, including:
- the face picture obtaining module 1710 acquires a face picture specified by the user.
- the person name obtaining module 1720 performs face recognition on the face image to obtain the name of the person face in the face picture. In this embodiment, based on the existing face recognition technology, the acquisition of the person's name can be completed.
- the information obtaining module 1730 obtains network resource information corresponding to the person name from the network according to the name of the person.
- the information display module 1740 displays the network resource information corresponding to the person name to the user. According to the technical solution of the embodiment, through face recognition The method accurately identifies the name of the person involved in the face picture, and associates the picture with the related information of the person name, thereby providing the user with more information.
- the person name A can be automatically recognized, and the search engine is used to search according to A as the keyword, and the related report of the star A is provided to the user, which helps to increase the user. Understanding of Star A.
- An embodiment of the present invention provides a face image information acquiring apparatus, including:
- the face picture obtaining module 1710 acquires a face picture specified by the user.
- the person name obtaining module 1720 performs face recognition on the face image to obtain the name of the person face in the face picture.
- the information obtaining module 1730 obtains the structured information corresponding to the person name as the network resource information corresponding to the person name from the site for recording the structured information in the network according to the name of the person, wherein the structured information corresponding to the person name includes multiple items corresponding to each other.
- Information In this embodiment, the site for recording structured information is typically a Wikipedia website. Since the information in the Wikipedia website is structured, as long as the corresponding person name is found, it can be directly captured.
- the information display module 1740 displays the network resource information corresponding to the person name to the user. According to the technical solution of the embodiment, since the distribution of multiple items in the structured information is relatively clear, it is advantageous for the user to understand the characters involved in the face image.
- the face is automatically detected and face recognition is performed, and the person name is identified; then, according to the name of the person, the related structured information of the character is obtained from the Wikipedia website. , including personal profile, experience, height, weight, major works, and the latest developments on Weibo. This makes it easy for the user to know the star A, to know more information and dynamics of the star A, and to establish certain contact and interaction with the star A through Weibo.
- An embodiment of the present invention provides a face image information acquiring apparatus, including:
- the face picture obtaining module 1710 acquires a face picture specified by the user.
- the person name obtaining module 1720 performs face recognition on the face image to obtain the name of the person face in the face picture. In this embodiment, based on the existing face recognition technology, the acquisition of the person's name can be completed.
- the information obtaining module 1730 filters out information corresponding to the plurality of items from the network resource information corresponding to the person name according to the name of the person.
- the type of the entry is not limited, for example, it may be height, weight, etc.; further, in order to timely pay attention to the character, the entry may be a social account entry; in order to timely related events of the character Understand, the entry can be an information message entry. And processing the extracted information into structured information according to various entries.
- the social account entry may be used to search for the identification information of the social account corresponding to the name of the person and/or the content of the social account corresponding to the name of the person; or may be obtained from the network according to the news information item. News information corresponding to the name of the person.
- the information display module 1740 displays the network resource information corresponding to the person name to the user.
- the information in the network can be organized into structured information and provided to the user; further, the social network dynamics of the related person and the latest news can be obtained according to the face image.
- the face is automatically detected and face recognition is performed, and the person name is identified; then, according to the name of the person, the microblog account of the relevant person is found, and the micro The blog account "starA" and the latest update of the account release are "very busy” available to the user; or the latest news of the disputed event of the relevant person can be found according to the name of the person, and the news title is provided to the user.
- an embodiment of the present invention provides a face image information acquiring apparatus, including:
- the face picture obtaining module 1810 acquires a face picture specified by the user.
- the name acquisition module specifically includes:
- the feature extraction module 1820 extracts features of the face image and extracts features of the collected face image.
- the face image specified by the user may be pre-processed and normalized in advance to facilitate feature extraction; in this embodiment, the sample face image may be collected and targeted to the skin color, eyes, nose, and mouth.
- the detected data can train the face model, and the face model can identify the position of the face in the user-specified picture and perform feature extraction.
- the feature comparison module 1830 compares the features of the face image with the features of the collected face image.
- the face picture selection module 1840 selects a similar face picture of the face picture from the collected face pictures according to the comparison result.
- the person name determining module 1850 determines the person name of the face in the face picture according to the name of the person face in the similar face picture.
- the information obtaining module 1860 acquires network resource information corresponding to the person name from the network according to the name of the person.
- the information display module 1870 displays the network resource information corresponding to the person name to the user. According to the technical solution of the embodiment, it is advantageous to accurately identify the name of the person corresponding to the face picture by means of feature comparison.
- an embodiment of the present invention provides a face image information acquiring apparatus, including:
- the face picture obtaining module 1910 acquires a face picture specified by the user.
- the name acquisition module specifically includes:
- the feature extraction module 1920 extracts features of the face image and extracts features of the collected face image.
- the feature comparison module 1930 compares the features of the face image with the features of the collected face image.
- the face picture selection module 1940 selects a similar face picture of the face picture from the collected face pictures according to the comparison result.
- the person name extraction module 1950 extracts one or more person names from the text corresponding to the similar face picture.
- the text type is not limited, and may be a title of the news in which the picture is located, a surrounding text, or the like.
- the weight value calculation module 1960 calculates a weight value for one or more person names based on attributes of one or more person names.
- the attribute is not limited, which may be the frequency, location, and the like of the name of the person, because the names of the people of different frequencies and positions are similar to the possibility of the person name corresponding to the face picture.
- the person name selection module 1970 selects a person name corresponding to the face in the similar face picture from one or more person names according to the level of the weight value.
- the person name determining module 1980 determines the person name of the face in the face picture according to the name of the person face in the similar face picture.
- the information obtaining module 1990 acquires network resource information corresponding to the person name from the network according to the name of the person.
- the information display module 19100 displays the network resource information corresponding to the person name to the user.
- an embodiment of the present invention provides a face image information acquiring apparatus, including:
- the face image obtaining module 2010 acquires a face image specified by the user.
- the name acquisition module specifically includes:
- the feature extraction module 2020 extracts features of the face image and extracts features of the collected face image.
- the feature comparison module 2030 compares the features of the face image with the features of the collected face image.
- the face picture selection module 2040 selects a similar face picture of the face picture from the collected face pictures according to the comparison result.
- the similarity calculation module 2050 calculates the similarity between the similar face image and the face image according to the comparison result.
- the similarity accumulation module 2060 accumulates the similarity of all similar face pictures corresponding to the same person name.
- the person name determining module 2070 determines the person name of the face in the face image according to the name of the face in the similar face image corresponding to the maximum similarity after the accumulation, when the maximum similarity is greater than the preset threshold. In the present embodiment, accurate recognition of a face can be achieved based on the accumulated similarity.
- the information obtaining module 2080 obtains network resource information corresponding to the person name from the network according to the name of the person.
- the information display module 2090 displays the network resource information corresponding to the person name to the user.
- a user-specified person belongs to a face of a database in which a person name has been established (the face image has been collected): First, a person with a known person name is established by face detection, feature extraction, and person name extraction. Face database; for the new face image specified by the user, face detection is performed on the picture. If there is no face, it returns directly. If there is a face, the face feature is extracted and quantized into a high-dimensional vector. The vector of the input picture is compared with all the face feature high-dimensional vectors in the library, and the Euclidean distance is calculated, and the first N vectors closest to each other are taken. The faces represented by these vectors are the faces most similar to the input face.
- the face database is too large, it takes a long time to compare one by one. You can cluster the faces in the library in advance, and then compare them with the faces of the clusters. For the first N similar faces, the similarity is used as the weight. Calculate the weight of each name, and add the weights of the same name. And find the name with the highest weight. If the name is greater than a certain threshold, it is considered that the input face belongs to the face corresponding to the name, otherwise it is considered that the face cannot be accurately recognized.
- modules in the devices of the embodiments can be adaptively changed and placed in one or more devices different from the embodiment.
- the modules or units or components of the embodiments may be combined into one module or unit or component, and further they may be divided into a plurality of sub-modules or sub-units or sub-components.
- any combination of the features disclosed in the specification, including the accompanying claims, the abstract and the drawings, and so on All processes or units of any method or device disclosed are combined.
- Each feature disclosed in this specification (including the accompanying claims, the abstract and the drawings) may be replaced by alternative features that provide the same, equivalent or similar purpose.
- the various component embodiments of the present invention may be implemented in hardware, or in a software module running on one or more processors, or in a combination thereof.
- a microprocessor or digital signal processor may be used in practice to implement some or all of the functionality of some or all of the components of a similar face image acquisition device in accordance with embodiments of the present invention.
- the invention can also be implemented as a device or device program (e.g., a computer program and a computer program product) for performing some or all of the methods described herein.
- a program implementing the invention may be stored on a computer readable medium or may be in the form of one or more signals. Such signals may be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.
- Figure 21 schematically illustrates a block diagram of a computing device for performing the method in accordance with the present invention.
- the computing device conventionally includes a processor 2110 and a computer program product or computer readable medium in the form of a memory 2120.
- the memory 2120 may be an electronic memory such as a flash memory, an EEPROM (Electrically Erasable Programmable Read Only Memory), an EPROM, a hard disk, or a ROM.
- Memory 2120 has a storage space 2130 for program code 2131 for performing any of the method steps described above.
- storage space 2130 for program code may include various program code 2131 for implementing various steps in the above methods, respectively.
- the program code can be read from or written to one or more computer program products.
- Such computer program products include program code carriers such as hard disks, compact disks (CDs), memory cards or floppy disks.
- Such a computer program product is typically a portable or fixed storage unit as described with reference to FIG.
- the storage unit may have a storage segment, a storage space, and the like that are similarly arranged to the storage 2120 in the computing device of FIG.
- the program code can be compressed, for example, in an appropriate form.
- the storage unit comprises computer readable code 2131' for performing the steps of the method according to the invention, ie code that can be read by a processor such as, for example, 2110, which when executed by the computing device causes the calculation The device performs the various steps in the methods described above.
- the present invention is applicable to computer systems/servers that can operate with numerous other general purpose or special purpose computing system environments or configurations.
- Examples of well-known computing systems, environments, and/or configurations suitable for use with computer systems/servers include, but are not limited to, personal computer systems, server computer systems, thin clients, thick clients, handheld or laptop devices, based on Microprocessor systems, set-top boxes, programmable consumer electronics, networked personal computers, small computer systems, mainframe computer systems, and distributed cloud computing technology environments including any of the above, and the like.
- the computer system/server can be described in the general context of computer system executable instructions (such as program modules) being executed by a computer system.
- program modules may include routines, programs, target programs, components, logic, data structures, and the like that perform particular tasks or implement particular abstract data types.
- the computer system/server can be implemented in a distributed cloud computing environment where tasks are performed by remote processing devices that are linked through a communication network.
- program modules may be located on a local or remote computing system storage medium including storage devices.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Human Computer Interaction (AREA)
- General Health & Medical Sciences (AREA)
- Library & Information Science (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Geometry (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Processing Or Creating Images (AREA)
- Image Analysis (AREA)
- Collating Specific Patterns (AREA)
Abstract
本发明提供了一种相似人脸图片获取和人脸图片信息获取方法和装置,主要涉及互联网技术领域,主要目的在于在提供相似图片时为用户提供具有人物相似的相似人脸图片。方法其包括:获取用户指定的人脸图片;对人脸图片进行人脸识别,以从已收集人脸图片中识别出人脸图片的相似人脸图片;将相似人脸图片显示给用户。根据本发明,基于人脸识别技术,根据用户指定的人脸图片为用户提供相似人脸图片;则用户指定图片中的人脸与相似图片中的人脸具有很大的相似度,在用户需求对指定人脸图片中人物相似的其他人物进行了解的情况下,本发明的技术方案可以满足用户的需求。
Description
本发明涉及互联网技术领域,具体而言,涉及一种相似人脸图片获取方法和装置以及一种人脸图片信息获取方法和装置。
在互联网技术领域,图片浏览占据了用户访问量的很大一部分。
目前,互联网图片在展现时,往往会提供一些相似的图片供用户参考。由于受到技术的限制,所提供的相似图片大部分与原始图片往往只是整体近似,而核心内容往往完全不同,相似度较低,从而相似图片对用户的价值也较小。例如,用户在网络上发现明星A在海边场景的照片,并需求与明星A类似长相风格的其他明星照片,而根据现有技术方案为人物B在海边场景的照片,其中明星A和人物B长相风格完全不同,则得到的照片不能满足用户需求。
发明内容
鉴于上述问题,提出了本发明以便提供一种克服上述问题或者至少部分地解决上述问题的相似人脸图片获取方法和装置以及一种人脸图片信息获取方法和装置。
依据本发明的一个方面,提供了一种相似人脸图片获取方法,其包括:获取用户指定的人脸图片;对所述人脸图片进行人脸识别,以从已收集人脸图片中识别出所述人脸图片的相似人脸图片;将所述相似人脸图片显示给所述用户。
依据本发明的另一方面,还提供了一种相似人脸图片获取装置,其包括:人脸图片获取模块,用于获取用户指定的人脸图片;人脸图片识别模块,用于对所述人脸图片进行人脸识别,以从已收集人脸图片中识别出所述人脸图片的相似人脸图片;人脸图片显示模块,用于将所述相似人脸图片显示给所述用户。
依据本发明的又一个方面,提供了一种人脸图片信息获取方法,其包括:获取用户指定的人脸图片;对所述人脸图片进行人脸识别,以得到所述人脸图片中人脸的人名;根据所述人名,从网络中获取所述人名对应的网络资源信息;将所述人名对应的网络资源信息显示给所述用户。
依据本发明的再一方面,还提供了一种人脸图片信息获取装置,其包括:人脸图片获取模块,用于获取用户指定的人脸图片;人名获取模块,用于对所述人脸图片进行人脸识别,以得到所述人脸图片中人脸的人名;信息获取模块,用于根据所述人名,从网络中获取所述人名对应的网络资源信息;信息显示模块,用于将所述人名对应的网络资源信息显示给所述用户。
根据本发明的又一个方面,提出了一种计算机程序,包括计算机可读代码,当所述计算机可读代码在计算设备上运行时,导致所述计算设备执行上文所述的相似人脸图片获取方法,或者导致所述计算设备执行上文所述的人脸图片信息获取方法。
根据本发明的再一个方面,提出了一种计算机可读介质,其中存储了上述的计算机程序。
根据以上技术方案,可知本发明的相似人脸图片获取方法和装置至少具有以下优点:
基于人脸识别技术,根据用户指定的人脸图片为用户提供相似人脸图片;则用户指定图片中的人脸与相似图片中的人脸具有很大的相似度,在用户需求对指定人脸图片中人物相似的其他人物进行了解的情况下,本发明的技术方案可以满足用户的需求。
上述说明仅是本发明技术方案的概述,为了能够更清楚了解本发明的技术手段,而可依照说明书的内容予以实施,并且为了让本发明的上述和其它目的、特征和优点能够更明显易懂,以下特举本发明的具体实施方式。
通过阅读下文优选实施方式的详细描述,各种其他的优点和益处对于本领域普通技术人员将变得清楚明了。附图仅用于示出优选实施方式的目的,而并不认为是对本发明的限制。而且在整个附图中,用相同的参考符号表示相同的部件。在附图中:
图1示出了根据本发明的一个实施例的相似人脸图片获取方法的流程图;
图2示出了根据本发明的一个实施例的相似人脸图片获取方法的流程图;
图3示出了根据本发明的一个实施例的相似人脸图片获取方法的流程图;
图4示出了根据本发明的一个实施例的相似人脸图片获取方法的流程图;
图5示出了根据本发明的一个实施例的相似人脸图片获取方法的流程图;
图6示出了根据本发明的一个实施例的相似人脸图片获取装置的框图;
图7示出了根据本发明的一个实施例的相似人脸图片获取装置的框图;
图8示出了根据本发明的一个实施例的相似人脸图片获取装置的框图;
图9示出了根据本发明的一个实施例的相似人脸图片获取装置的框图;
图10示出了根据本发明的一个实施例的相似人脸图片获取装置的框图;
图11示出了根据本发明的一个实施例的人脸图片信息获取方法的流程图;
图12示出了根据本发明的一个实施例的人脸图片信息获取方法的流程图;
图13示出了根据本发明的一个实施例的人脸图片信息获取方法的流程图;
图14示出了根据本发明的一个实施例的人脸图片信息获取方法的流程图;
图15示出了根据本发明的一个实施例的人脸图片信息获取方法的流程图;
图16示出了根据本发明的一个实施例的人脸图片信息获取方法的流程图;
图17示出了根据本发明的一个实施例的人脸图片信息获取装置的框图;
图18示出了根据本发明的一个实施例的人脸图片信息获取装置的框图;
图19示出了根据本发明的一个实施例的人脸图片信息获取装置的框图;
图20示出了根据本发明的一个实施例的人脸图片信息获取装置的框图。
图21示意性地示出了用于执行根据本发明的方法的计算设备的框图;以及
图22示意性地示出了用于保持或者携带实现根据本发明的方法的程序代码的存储单元。
下面将参照附图更详细地描述本公开的示例性实施例。虽然附图中显示了本公开的示例性实施例,然而应当理解,可以以各种形式实现本公开而不应被这里阐述的实施例所限制。相反,提供这些实施例是为了能够更透彻地理解本公开,并且能够将本公开的范围完整的传达给本领域的技术人员。
如图1所示,本发明的一个实施例提供了一种相似人脸图片获取方法,其包括:
步骤110,获取用户指定的人脸图片。
步骤120,对人脸图片进行人脸识别,以从已收集人脸图片中识别出人脸图片的相似人脸图片。在本实施例中,基于现有的人脸识别技术,即可以完成人名的获取。
步骤130,将相似人脸图片显示给用户。根据本实施例的技术方案,根据用户指定的人脸图片为用户提供相似人脸图片;则用户指定图片中的人脸与相似图片中的人脸具有很大的相似度,在用户需求对指定人脸图片中人物相似的其他人物进行了解的情况下,本实施例的技术方案可以满足用户的需求。
根据图1,用户输入了明星A的照片,需要找到与明星A长相风格类似的其他人物的照片;基于人脸识别技术对图片中的明星A进行识别,从预设的人脸图片库中找到与明星A人脸相似的明星C的照片,有利于增加用户对明星C的了解。
如图2所示,本发明的一个实施例提供了一种相似人脸图片获取方法,其包括:
步骤210,获取用户指定的人脸图片。
步骤220,对人脸图片进行人脸识别,以从已收集人脸图片中识别出人脸图片的相似人脸图片。
步骤230,获取相似人脸图片中人脸的人名。
步骤240,获取相似人脸图片与人脸图片的相似度,并累加具有相同人名的所有相似人脸图片对应的相似度。在本实施例中,进行相似度的累加,相当于基于人名综合计算了每个人物的一张或多张人脸图片与原始人脸图片的相似度,则按累加相似图片有利于找到相似度最高的人物。
步骤250,在累加后的最大相似度小于预定阈值时,根据相似人脸图片对应的相似度大小,从相似人脸图片中选择一张或多张人脸图片显示给用户。在本实施例中,在累加后的最大相似度超过一定阈值时,表明已经查询到了完全相同人物的人脸图片,此时优先为用户提供相同人物的人脸图片,以表明精确识别出了该人物。
根据图2,用户输入了明星A的照片,得到了相似人脸图片如下:照片a为明星C的照片,相似度80%;照片b为明星D的照片,相似度70%;照片c为明星C的照片,相似度70%;照片d为明星D的照片,相似度55%;经过累加后明星C对应的相似度为135%,明星D对应的相似度为140%;预设阈值为90%×n,n为每个人名对应的照片数,则对与明星C和D阈值都是180%,表明不存在相同人物的人脸图片;此时,可以将明星D对应的照片输出,表示明星D为明星A的相似人脸图片。
本发明的一个实施例提供了一种相似人脸图片获取方法,其包括:
步骤210,获取用户指定的人脸图片。
步骤220,对人脸图片进行人脸识别,以从已收集人脸图片中识别出人脸图片的相似人脸图片。
步骤230,获取相似人脸图片中人脸的人名。
步骤240,获取相似人脸图片与人脸图片的相似度,并累加具有相同人名的所有相似人脸图片对应的相似度。
步骤250,在累加后的最大相似度小于预定阈值时,根据相似人脸图片对应的相似度大小,从相似人脸图片中选择一张或多张人脸图片显示给用户。
步骤251,将一张或多张人脸图片对应的相似度和/或其中人脸的人名,显示给用户。在本实施例中,对
用户进行相似度、人名的提示,使得用户对相似人脸图片有更多了解。
根据图2,用户输入了明星A的照片,根据本实施例技术方案获取的相似人脸图片为明星C的照片,在明星C的照片上显示相似度为80%,并显其人名为C。
本发明的一个实施例提供了一种相似人脸图片获取方法,其包括:
步骤210,获取用户指定的人脸图片。
步骤220,对人脸图片进行人脸识别,以从已收集人脸图片中识别出人脸图片的相似人脸图片。
步骤230,获取相似人脸图片中人脸的人名。
步骤240,获取相似人脸图片与人脸图片的相似度,并累加具有相同人名的所有相似人脸图片对应的相似度。
步骤250,在累加后的最大相似度小于预定阈值时,根据相似人脸图片对应的相似度大小,从相似人脸图片中选择一张或多张人脸图片显示给用户。
步骤252,根据一张或多张人脸图片对应的相似度高低,设置一张或多张人脸图片的显示顺序。在本实施例中,有利于将最相似的照片首先提供给用户。
根据图2,用户输入了明星A的照片,根据本实施例技术方案获取的相似人脸图片为明星C的一张照片、明星D的一张照片,相似度分别为80%、85%,则首先显示明星D的照片,之后显示明星C的照片。
本发明的一个实施例提供了一种相似人脸图片获取方法,其包括:
步骤210,获取用户指定的人脸图片。
步骤220,对人脸图片进行人脸识别,以从已收集人脸图片中识别出人脸图片的相似人脸图片。
步骤230,获取相似人脸图片中人脸的人名。
步骤240,获取相似人脸图片与人脸图片的相似度,并累加具有相同人名的所有相似人脸图片对应的相似度。
步骤250,在累加后的最大相似度小于预定阈值时,根据相似人脸图片对应的相似度大小,从相似人脸图片中选择一张或多张人脸图片显示给用户。
步骤253,根据一张或多张人脸图片对应的相似度高低,为一张或多张人脸图片设置相应的评价信息并显示给用户。根据本实施例的技术方案,评价信息能够以用户更容易理解的方式,体现相似度的高低。
根据图2,用户输入了明星A的照片,根据本实施例技术方案获取的相似人脸图片为明星C的照片,相似度分别为80%,则在照片上显示评价信息为:“失散多年的姐妹”。
如图3所示,本发明的一个实施例提供了一种相似人脸图片获取方法,其包括:
步骤310,获取用户指定的人脸图片。
步骤320,对人脸图片进行人脸识别,以从已收集人脸图片中识别出人脸图片的相似人脸图片。
步骤330,获取相似人脸图片中人脸的人名。
步骤340,获取相似人脸图片与人脸图片的相似度,并累加具有相同人名的所有相似人脸图片对应的相似度。
步骤350,在累加后的最大相似度小于预定阈值时,按相似人脸图片对应的相似度高低顺序获取人脸图片。
步骤360,判断最新获取的人脸图片中人脸的人名,是否位于其他已获取的人脸图片中人脸的人名之中。
步骤370,在判断结果为否时,将最新获取的人脸图片显示给用户。根据本实施例的技术方案,可以避免出现重复人物的相似人脸图片,以保证用户尽可能看到多个人物的相似人脸图片。
根据图3,用户输入了明星A的照片,得到了相似人脸图片如下:照片a为明星C的照片,相似度80%;照片b为明星D的照片,相似度70%;照片c为明星C的照片,相似度70%;照片d为明星D的照片,相似度55%。则按相似度高低,首先显示照片a;获取照片b,发现明星D的照片并未被显示,则显示照片b;获取照片c,发现明星C的照片已经被显示,则放弃显示;获取照片d,发现明星D的照片已经被显示,则放弃显示。
如图4所示,本发明的一个实施例提供了一种相似人脸图片获取方法,其包括:
步骤410,获取用户指定的人脸图片。
步骤420,提取人脸图片的特征,以及提取已收集人脸图片的特征。在本实施例中,对用户指定的人脸图片可以提前进行预处理、归一化,以利于特征提取;在本实施例中,可以收集样本人脸图片,并针对肤色、眼睛、鼻子、嘴角等进行检测,检测到的数据可以训练人脸模型,通过该人脸模型可以识别出用户指定图片中人脸的位置,并进行特征提取。
步骤430,将人脸图片的特征与已收集人脸图片的特征进行比较。
步骤440,根据比较结果从已收集人脸图片中选出相似人脸图片。
步骤450,获取相似人脸图片中人脸的人名。
步骤460,根据比较结果计算相似人脸图片与人脸图片的相似度,并累加具有相同人名的所有相似人脸图片对应的相似度。
步骤470,在累加后的最大相似度小于预定阈值时,根据相似人脸图片对应的相似度大小,从相似人脸图片中选择一张或多张人脸图片显示给用户。
根据图4,判断用户指定的是否属于已经建立好人名的数据库的某张人脸(已收集人脸图片):首先,通过人脸检测、特征提取和人名提取,建立起一个已知人名的人脸数据库;对于用户指定的新人脸图片,对图片进行人脸检测,如果没有人脸,则直接返回,如果有人脸,则提取人脸特征,并且量化为一个高维向量。将该输入图片的向量和库内所有人脸特征高维向量进行比较,计算其欧式距离,并取距离最近的前N个向量。这些向量所表征的人脸就是与该输入人脸最相似的人脸。如果人脸数据库过于庞大,逐个比较费时很长,可以事先对库内人脸进行聚类,然后只与聚好类的人脸进行比较;对于前N个相似人脸,以相似度为权重,计算每个名字的权重,相同名字的权值相加。并求出权重最高的名字。如果该名字大于一定阈值,则认为输入人脸属于该名字对应的人脸,否则认为无法精确识别出该人脸。
如图5所示,本发明的一个实施例提供了一种相似人脸图片获取方法,其包括:
步骤510,获取用户指定的人脸图片。
步骤520,对人脸图片进行人脸识别,以从已收集人脸图片中识别出人脸图片的相似人脸图片。
步骤530,从相似人脸图片对应的文本中,提取一个或多个人名。在本实施例中,对文本类型不进行限制,可以是图片所在新闻的标题、环绕文本等。
步骤540,根据一个或多个人名的属性,为一个或多个人名计算权重值。在本实施例中,对属性不进行限制,其可以是人名出现的频次、位置等,因为不同频次、位置的人名就相似人脸图片对应的人名的可能性有所不同。
步骤550,根据权重值的高低,从一个或多个人名中选择相似人脸图片中人脸的人名。
步骤560,获取相似人脸图片与人脸图片的相似度,并累加具有相同人名的所有相似人脸图片对应的相似度。
步骤570,在累加后的最大相似度小于预定阈值时,根据相似人脸图片对应的相似度大小,从相似人脸图片中选择一张或多张人脸图片显示给用户。
根据图5,对于相似人脸图片所在的新闻页面,首先对该图片对应的新闻标题、环绕文本预处理,然后分词;从分词结果中提取出候选人名,并与人名词表比较,去除非人名;根据每个人名出现的频率、位置,以及与其他词语的关系计算一个权重;如果权重大于某个阈值,则选择权重最大的名字作为该图片中人脸的人名,否则认为无法提取可靠的人名。
如图6所示,本发明的一个实施例提供了一种相似人脸图片获取装置,其包括:
人脸图片获取模块610,获取用户指定的人脸图片。
人脸图片识别模块620,对人脸图片进行人脸识别,以从已收集人脸图片中识别出人脸图片的相似人脸图片。在本实施例中,基于现有的人脸识别技术,即可以完成人名的获取。
人脸图片显示模块630,将相似人脸图片显示给用户。根据本实施例的技术方案,根据用户指定的人脸图片为用户提供相似人脸图片;则用户指定图片中的人脸与相似图片中的人脸具有很大的相似度,在用户需求对指定人脸图片中人物相似的其他人物进行了解的情况下,本实施例的技术方案可以满足用户的需求。
根据图6,用户输入了明星A的照片,需要找到与明星A长相风格类似的其他人物的照片;基于人脸识别技术对图片中的明星A进行识别,从预设的人脸图片库中找到与明星A人脸相似的明星C的照片,有利于增加用户对明星C的了解。
如图7所示,本发明的一个实施例提供了一种相似人脸图片获取装置,其包括:
人脸图片获取模块710,获取用户指定的人脸图片。
人脸图片识别模块720,对人脸图片进行人脸识别,以从已收集人脸图片中识别出人脸图片的相似人脸图片。
人名获取模块730,获取相似人脸图片中人脸的人名。
相似度获取模块740,获取相似人脸图片与人脸图片的相似度,并累加具有相同人名的所有相似人脸图片对应的相似度。在本实施例中,进行相似度的累加,相当于基于人名综合计算了每个人物的一张或多张人脸图片与原始人脸图片的相似度,则按累加相似图片有利于找到相似度最高的人物。
人脸图片显示模块750,在累加后的最大相似度小于预定阈值时,根据相似人脸图片对应的相似度大小,从相似人脸图片中选择一张或多张人脸图片显示给用户。在本实施例中,在累加后的最大相似度超过一定阈值时,表明已经查询到了完全相同人物的人脸图片,此时优先为用户提供相同人物的人脸图片,以表明精确识别
出了该人物。
根据图7,用户输入了明星A的照片,得到了相似人脸图片如下:照片a为明星C的照片,相似度80%;照片b为明星D的照片,相似度70%;照片c为明星C的照片,相似度70%;照片d为明星D的照片,相似度55%;经过累加后明星C对应的相似度为135%,明星D对应的相似度为140%;预设阈值为90%×n,n为每个人名对应的照片数,则对与明星C和D阈值都是180%,表明不存在相同人物的人脸图片;此时,可以将明星D对应的照片输出,表示明星D为明星A的相似人脸图片。
本发明的一个实施例提供了一种相似人脸图片获取装置,其包括:
人脸图片获取模块710,获取用户指定的人脸图片。
人脸图片识别模块720,对人脸图片进行人脸识别,以从已收集人脸图片中识别出人脸图片的相似人脸图片。
人名获取模块730,获取相似人脸图片中人脸的人名。
相似度获取模块740,获取相似人脸图片与人脸图片的相似度,并累加具有相同人名的所有相似人脸图片对应的相似度。
人脸图片显示模块750,在累加后的最大相似度小于预定阈值时,根据相似人脸图片对应的相似度大小,从相似人脸图片中选择一张或多张人脸图片显示给用户。
相似度/人名显示模块751,将一张或多张人脸图片对应的相似度和/或其中人脸的人名,显示给用户。在本实施例中,对用户进行相似度、人名的提示,使得用户对相似人脸图片有更多了解。
根据图7,用户输入了明星A的照片,根据本实施例技术方案获取的相似人脸图片为明星C的照片,在明星C的照片上显示相似度为80%,并显其人名为C。
本发明的一个实施例提供了一种相似人脸图片获取装置,其包括:
人脸图片获取模块710,获取用户指定的人脸图片。
人脸图片识别模块720,对人脸图片进行人脸识别,以从已收集人脸图片中识别出人脸图片的相似人脸图片。
人名获取模块730,获取相似人脸图片中人脸的人名。
相似度获取模块740,获取相似人脸图片与人脸图片的相似度,并累加具有相同人名的所有相似人脸图片对应的相似度。
人脸图片显示模块750,在累加后的最大相似度小于预定阈值时,根据相似人脸图片对应的相似度大小,从相似人脸图片中选择一张或多张人脸图片显示给用户。
显示顺序设置模块752,根据一张或多张人脸图片对应的相似度高低,设置一张或多张人脸图片的显示顺序。在本实施例中,有利于将最相似的照片首先提供给用户。
根据图7,用户输入了明星A的照片,根据本实施例技术方案获取的相似人脸图片为明星C的一张照片、明星D的一张照片,相似度分别为80%、85%,则首先显示明星D的照片,之后显示明星C的照片。
本发明的一个实施例提供了一种相似人脸图片获取装置,其包括:
人脸图片获取模块710,获取用户指定的人脸图片。
人脸图片识别模块720,对人脸图片进行人脸识别,以从已收集人脸图片中识别出人脸图片的相似人脸图片。
人名获取模块730,获取相似人脸图片中人脸的人名。
相似度获取模块740,获取相似人脸图片与人脸图片的相似度,并累加具有相同人名的所有相似人脸图片对应的相似度。
人脸图片显示模块750,在累加后的最大相似度小于预定阈值时,根据相似人脸图片对应的相似度大小,从相似人脸图片中选择一张或多张人脸图片显示给用户。
评价信息显示模块753,根据一张或多张人脸图片对应的相似度高低,为一张或多张人脸图片设置相应的评价信息并显示给用户。根据本实施例的技术方案,评价信息能够以用户更容易理解的方式,体现相似度的高低。
根据图7,用户输入了明星A的照片,根据本实施例技术方案获取的相似人脸图片为明星C的照片,相似度分别为80%,则在照片上显示评价信息为:“失散多年的姐妹”。
如图8所示,本发明的一个实施例提供了一种相似人脸图片获取装置,其包括:
人脸图片获取模块810,获取用户指定的人脸图片。
人脸图片识别模块820,对人脸图片进行人脸识别,以从已收集人脸图片中识别出人脸图片的相似人脸图片。
人名获取模块830,获取相似人脸图片中人脸的人名。
相似度获取模块840,获取相似人脸图片与人脸图片的相似度,并累加具有相同人名的所有相似人脸图片对应的相似度。
顺序获取模块850,在累加后的最大相似度小于预定阈值时,按相似人脸图片对应的相似度高低顺序获取人脸图片。
人名判断模块860,判断最新获取的人脸图片中人脸的人名,是否位于其他已获取的人脸图片中人脸的人名之中。
人脸图片显示模块870,在判断结果为否时,将最新获取的人脸图片显示给用户。根据本实施例的技术方案,可以避免出现重复人物的相似人脸图片,以保证用户尽可能看到多个人物的相似人脸图片。
根据图8,用户输入了明星A的照片,得到了相似人脸图片如下:照片a为明星C的照片,相似度80%;照片b为明星D的照片,相似度70%;照片c为明星C的照片,相似度70%;照片d为明星D的照片,相似度55%。则按相似度高低,首先显示照片a;获取照片b,发现明星D的照片并未被显示,则显示照片b;获取照片c,发现明星C的照片已经被显示,则放弃显示;获取照片d,发现明星D的照片已经被显示,则放弃显示。
如图9所示,本发明的一个实施例提供了一种相似人脸图片获取装置,其包括:
人脸图片获取模块910,获取用户指定的人脸图片。
特征提取模块920,提取人脸图片的特征,以及提取已收集人脸图片的特征。在本实施例中,对用户指定的人脸图片可以提前进行预处理、归一化,以利于特征提取;在本实施例中,可以收集样本人脸图片,并针对肤色、眼睛、鼻子、嘴角等进行检测,检测到的数据可以训练人脸模型,通过该人脸模型可以识别出用户指定图片中人脸的位置,并进行特征提取。
特征比较模块930,将人脸图片的特征与已收集人脸图片的特征进行比较。
人脸图片识别模块940,根据比较结果从已收集人脸图片中选出相似人脸图片。
人名获取模块950,获取相似人脸图片中人脸的人名。
相似度获取模块960,根据比较结果计算相似人脸图片与人脸图片的相似度,并累加具有相同人名的所有相似人脸图片对应的相似度。
人脸图片显示模块970,在累加后的最大相似度小于预定阈值时,根据相似人脸图片对应的相似度大小,从相似人脸图片中选择一张或多张人脸图片显示给用户。
根据图9,判断用户指定的是否属于已经建立好人名的数据库的某张人脸(已收集人脸图片):首先,通过人脸检测、特征提取和人名提取,建立起一个已知人名的人脸数据库;对于用户指定的新人脸图片,对图片进行人脸检测,如果没有人脸,则直接返回,如果有人脸,则提取人脸特征,并且量化为一个高维向量。将该输入图片的向量和库内所有人脸特征高维向量进行比较,计算其欧式距离,并取距离最近的前N个向量。这些向量所表征的人脸就是与该输入人脸最相似的人脸。如果人脸数据库过于庞大,逐个比较费时很长,可以事先对库内人脸进行聚类,然后只与聚好类的人脸进行比较;对于前N个相似人脸,以相似度为权重,计算每个名字的权重,相同名字的权值相加。并求出权重最高的名字。如果该名字大于一定阈值,则认为输入人脸属于该名字对应的人脸,否则认为无法精确识别出该人脸。
如图10所示,本发明的一个实施例提供了一种相似人脸图片获取装置,其包括:
人脸图片获取模块1010,获取用户指定的人脸图片。
人脸图片识别模块1020,对人脸图片进行人脸识别,以从已收集人脸图片中识别出人脸图片的相似人脸图片。
人名获取模块,具体包括:
人名提取模块1030,从相似人脸图片对应的文本中,提取一个或多个人名。在本实施例中,对文本类型不进行限制,可以是图片所在新闻的标题、环绕文本等。
权重值计算模块1040,根据一个或多个人名的属性,为一个或多个人名计算权重值。在本实施例中,对属性不进行限制,其可以是人名出现的频次、位置等,因为不同频次、位置的人名就相似人脸图片对应的人名的可能性有所不同。
人名选择模块1050,根据权重值的高低,从一个或多个人名中选择相似人脸图片中人脸的人名。
相似度获取模块1060,获取相似人脸图片与人脸图片的相似度,并累加具有相同人名的所有相似人脸图片对应的相似度。
人脸图片显示模块1070,在累加后的最大相似度小于预定阈值时,根据相似人脸图片对应的相似度大小,从相似人脸图片中选择一张或多张人脸图片显示给用户。
根据图10,对于相似人脸图片所在的新闻页面,首先对该图片对应的新闻标题、环绕文本预处理,然后分词;从分词结果中提取出候选人名,并与人名词表比较,去除非人名;根据每个人名出现的频率、位置,
以及与其他词语的关系计算一个权重;如果权重大于某个阈值,则选择权重最大的名字作为该图片中人脸的人名,否则认为无法提取可靠的人名。
目前,传统的互联网图片在展现时,除了图片本身之外,只能展现一些简单的文本或者标题信息,用户除了能看到该简单的文本或者标题信息之外,很难获得和图片相关的其他信息,造成用户获得的信息十分匮乏。即使通过基于图片内容的搜索引擎来搜索,由于现有的图片搜索技术的局限性,也很难获得更多的该图片的精确信息。另外,对于包含人物的图片,用户往往希望知道更多有关该人物的信息,但缺不知如何组织搜索词进行查询,使得搜索引擎解决此方面的效率不高,用户体验也较差。对此,本发明提供了如下方案
如图11所示,本发明的一个实施例中提供了一种人脸图片信息获取方法,其包括:
步骤1110,获取用户指定的人脸图片。
步骤1120,对人脸图片进行人脸识别,以得到人脸图片中人脸的人名。在本实施例中,基于现有的人脸识别技术,即可以完成人名的获取。
步骤1130,根据人名,从网络中获取人名对应的网络资源信息。
步骤1140,将人名对应的网络资源信息显示给用户。根据本实施例的技术方案,通过人脸识别的方法,精确的识别出人脸图片涉及的人名,将图片和该人名的相关信息联系起来,从而给用户提供更多的信息。
根据图11,对于用户输入的明星A的照片,可以自动识别出其人名为A,并按照A作为关键词使用搜索引擎进行搜索,得到明星A的相关报道提供给用户,有助于增加用户对明星A的了解。
如图12所示,本发明的一个实施例中提供了一种人脸图片信息获取方法,其包括:
步骤1210,获取用户指定的人脸图片。
步骤1220,对人脸图片进行人脸识别,以得到人脸图片中人脸的人名。
步骤1230,根据人名,从网络中的用于记录结构化信息的站点上,获取人名对应的结构化信息作为人名对应的网络资源信息,其中人名对应的结构化信息中包含多种条目对应的信息。在本实施例中,记录结构化信息的站点比较典型的是百科网站,由于百科网站中的信息都是结构化的,所以只要找到相应人名,便可直接抓取。
步骤1240,将人名对应的网络资源信息显示给用户。根据本实施例的技术方案,由于结构化信息中多种条目分布较为明确,有利于用户对人脸图片中涉及人物进行了解。
根据图12,对于用户在网页上选定的明星A的图片,自动检测到人脸并进行人脸识别,识别出人名为A;然后根据人名,从百科网站获取该人物的相关结构化信息,包括个人简介,经历,身高,体重,主要作品,以及微博最新动态。这样可以方便用户认识该明星A,了解该明星A的更多信息和动态,以及通过微博与该明星A建立一定的联系和互动。
如图13所示,本发明的一个实施例中提供了一种人脸图片信息获取方法,其包括:
步骤1310,获取用户指定的人脸图片。
步骤1320,对人脸图片进行人脸识别,以得到人脸图片中人脸的人名。在本实施例中,基于现有的人脸识别技术,即可以完成人名的获取。
步骤1330,根据人名,按预设的多种条目,从人名对应的网络资源信息中筛选出多种条目对应的信息。在本实施例中,对于条目的类型不进行限制,例如,可以是身高、体重等;进一步地,为了对人物动态进行及时关注,则该条目可以是社交账号条目;为了对人物的相关事件及时了解,则该条目可以是信息消息条目。
步骤1340,根据多种条目,将提取的信息处理为结构化信息。在本实施例中,结合前述的内容,可以按社交账号条目,查找人名对应的社交账号的标识信息和/或人名对应的社交账号所发布的内容;也可以按新闻信息条目,从网络中获取人名对应的新闻信息。
步骤1350,将人名对应的网络资源信息显示给用户。根据本实施例的技术方案,能够将网络中的信息整理为结构化信息提供给用户;进一步,可以根据人脸图片来获取相关人物的社交网络动态以及最新新闻。
根据图13,对于用户在网页上选定的明星A的图片,自动检测到人脸并进行人脸识别,识别出人名为A;然后根据人名,找到相关人物的微博账号,将该微博账号“starA”以及账号发布的最新动态“很忙”提供给用户;也可以根据人名,找到相关人物的争议事件的最新新闻,并将新闻标题提供给用户。
如图14所示,本发明的一个实施例中提供了一种人脸图片信息获取方法,其包括:
步骤1410,获取用户指定的人脸图片。
步骤1420,提取人脸图片的特征,以及提取已收集人脸图片的特征。在本实施例中,对用户指定的人脸图片可以提前进行预处理、归一化,以利于特征提取;在本实施例中,可以收集样本人脸图片,并针对肤色、眼睛、鼻子、嘴角等进行检测,检测到的数据可以训练人脸模型,通过该人脸模型可以识别出用户指定图片中人脸的位置,并进行特征提取。
步骤1430,将人脸图片的特征与已收集人脸图片的特征进行比较。
步骤1440,根据比较结果从已收集人脸图片中选出人脸图片的相似人脸图片。
步骤1450,根据相似人脸图片中人脸的人名,确定人脸图片中人脸的人名。
步骤1460,根据人名,从网络中获取人名对应的网络资源信息。
步骤1470,将人名对应的网络资源信息显示给用户。根据本实施例的技术方案,通过特征比较的方式,有利于准确识别人脸图片对应的人名。
如图15所示,本发明的一个实施例中提供了一种人脸图片信息获取方法,其包括:
步骤1510,获取用户指定的人脸图片。
步骤1520,提取人脸图片的特征,以及提取已收集人脸图片的特征。
步骤1530,将人脸图片的特征与已收集人脸图片的特征进行比较。
步骤1540,根据比较结果从已收集人脸图片中选出人脸图片的相似人脸图片。
步骤1550,从相似人脸图片对应的文本中,提取一个或多个人名。在本实施例中,对文本类型不进行限制,可以是图片所在新闻的标题、环绕文本等。
步骤1560,根据一个或多个人名的属性,为一个或多个人名计算权重值。在本实施例中,对属性不进行限制,其可以是人名出现的频次、位置等,因为不同频次、位置的人名就相似人脸图片对应的人名的可能性有所不同。
步骤1570,根据权重值的高低,从一个或多个人名中选择相似人脸图片中人脸对应的人名。
步骤1580,根据相似人脸图片中人脸的人名,确定人脸图片中人脸的人名。
步骤1590,根据人名,从网络中获取人名对应的网络资源信息。
步骤15100,将人名对应的网络资源信息显示给用户。
根据图5,对于相似人脸图片所在的新闻页面,首先对该图片对应的新闻标题、环绕文本预处理,然后分词;从分词结果中提取出候选人名,并与人名词表比较,去除非人名;根据每个人名出现的频率、位置,以及与其他词语的关系计算一个权重;如果权重大于某个阈值,则选择权重最大的名字作为该图片中人脸的人名,否则认为无法提取可靠的人名。
如图16所示,本发明的一个实施例中提供了一种人脸图片信息获取方法,其包括:
步骤1610,获取用户指定的人脸图片。
步骤1620,提取人脸图片的特征,以及提取已收集人脸图片的特征。
步骤1630,将人脸图片的特征与已收集人脸图片的特征进行比较。
步骤1640,根据比较结果从已收集人脸图片中选出人脸图片的相似人脸图片。
步骤1650,根据比较结果,计算相似人脸图片与人脸图片之间的相似度。
步骤1660,累加相同人名对应的所有相似人脸图片的相似度。
步骤1670,在累加后最大的相似度大于预设阈值时,按累加后的最大相似度对应的相似人脸图片中人脸的人名,确定人脸图片中人脸的人名。在本实施例中,基于累加后的相似度可以实现人脸的精确识别。
步骤1680,根据人名,从网络中获取人名对应的网络资源信息。
步骤1690,将人名对应的网络资源信息显示给用户。
根据图16,判断用户指定的是否属于已经建立好人名的数据库的某张人脸(已收集人脸图片):首先,通过人脸检测、特征提取和人名提取,建立起一个已知人名的人脸数据库;对于用户指定的新人脸图片,对图片进行人脸检测,如果没有人脸,则直接返回,如果有人脸,则提取人脸特征,并且量化为一个高维向量。将该输入图片的向量和库内所有人脸特征高维向量进行比较,计算其欧式距离,并取距离最近的前N个向量。这些向量所表征的人脸就是与该输入人脸最相似的人脸。如果人脸数据库过于庞大,逐个比较费时很长,可以事先对库内人脸进行聚类,然后只与聚好类的人脸进行比较;对于前N个相似人脸,以相似度为权重,计算每个名字的权重,相同名字的权值相加。并求出权重最高的名字。如果该名字大于一定阈值,则认为输入人脸属于该名字对应的人脸,否则认为无法精确识别出该人脸。
如图17所示,本发明的一个实施例中提供了一种人脸图片信息获取装置,其包括:
人脸图片获取模块1710,获取用户指定的人脸图片。
人名获取模块1720,对人脸图片进行人脸识别,以得到人脸图片中人脸的人名。在本实施例中,基于现有的人脸识别技术,即可以完成人名的获取。
信息获取模块1730,根据人名,从网络中获取人名对应的网络资源信息。
信息显示模块1740,将人名对应的网络资源信息显示给用户。根据本实施例的技术方案,通过人脸识别
的方法,精确的识别出人脸图片涉及的人名,将图片和该人名的相关信息联系起来,从而给用户提供更多的信息。
根据图17,对于用户输入的明星A的照片,可以自动识别出其人名为A,并按照A作为关键词使用搜索引擎进行搜索,得到明星A的相关报道提供给用户,有助于增加用户对明星A的了解。
本发明的一个实施例中提供了一种人脸图片信息获取装置,其包括:
人脸图片获取模块1710,获取用户指定的人脸图片。
人名获取模块1720,对人脸图片进行人脸识别,以得到人脸图片中人脸的人名。
信息获取模块1730,根据人名,从网络中的用于记录结构化信息的站点上,获取人名对应的结构化信息作为人名对应的网络资源信息,其中人名对应的结构化信息中包含多种条目对应的信息。在本实施例中,记录结构化信息的站点比较典型的是百科网站,由于百科网站中的信息都是结构化的,所以只要找到相应人名,便可直接抓取。
信息显示模块1740,将人名对应的网络资源信息显示给用户。根据本实施例的技术方案,由于结构化信息中多种条目分布较为明确,有利于用户对人脸图片中涉及人物进行了解。
根据图17,对于用户在网页上选定的明星A的图片,自动检测到人脸并进行人脸识别,识别出人名为A;然后根据人名,从百科网站获取该人物的相关结构化信息,包括个人简介,经历,身高,体重,主要作品,以及微博最新动态。这样可以方便用户认识该明星A,了解该明星A的更多信息和动态,以及通过微博与该明星A建立一定的联系和互动。
本发明的一个实施例中提供了一种人脸图片信息获取装置,其包括:
人脸图片获取模块1710,获取用户指定的人脸图片。
人名获取模块1720,对人脸图片进行人脸识别,以得到人脸图片中人脸的人名。在本实施例中,基于现有的人脸识别技术,即可以完成人名的获取。
信息获取模块1730,根据人名,按预设的多种条目,从人名对应的网络资源信息中筛选出多种条目对应的信息。在本实施例中,对于条目的类型不进行限制,例如,可以是身高、体重等;进一步地,为了对人物动态进行及时关注,则该条目可以是社交账号条目;为了对人物的相关事件及时了解,则该条目可以是信息消息条目。以及根据多种条目,将提取的信息处理为结构化信息。在本实施例中,结合前述的内容,可以按社交账号条目,查找人名对应的社交账号的标识信息和/或人名对应的社交账号所发布的内容;也可以按新闻信息条目,从网络中获取人名对应的新闻信息。
信息显示模块1740,将人名对应的网络资源信息显示给用户。根据本实施例的技术方案,能够将网络中的信息整理为结构化信息提供给用户;进一步,可以根据人脸图片来获取相关人物的社交网络动态以及最新新闻。
根据图17,对于用户在网页上选定的明星A的图片,自动检测到人脸并进行人脸识别,识别出人名为A;然后根据人名,找到相关人物的微博账号,将该微博账号“starA”以及账号发布的最新动态“很忙”提供给用户;也可以根据人名,找到相关人物的争议事件的最新新闻,并将新闻标题提供给用户。
如图18所示,本发明的一个实施例中提供了一种人脸图片信息获取装置,其包括:
人脸图片获取模块1810,获取用户指定的人脸图片。
人名获取模块,具体包括:
特征提取模块1820,提取人脸图片的特征,以及提取已收集人脸图片的特征。在本实施例中,对用户指定的人脸图片可以提前进行预处理、归一化,以利于特征提取;在本实施例中,可以收集样本人脸图片,并针对肤色、眼睛、鼻子、嘴角等进行检测,检测到的数据可以训练人脸模型,通过该人脸模型可以识别出用户指定图片中人脸的位置,并进行特征提取。
特征比较模块1830,将人脸图片的特征与已收集人脸图片的特征进行比较。
人脸图片选择模块1840,根据比较结果从已收集人脸图片中选出人脸图片的相似人脸图片。
人名确定模块1850,根据相似人脸图片中人脸的人名,确定人脸图片中人脸的人名。
信息获取模块1860,根据人名,从网络中获取人名对应的网络资源信息。
信息显示模块1870,将人名对应的网络资源信息显示给用户。根据本实施例的技术方案,通过特征比较的方式,有利于准确识别人脸图片对应的人名。
如图19所示,本发明的一个实施例中提供了一种人脸图片信息获取装置,其包括:
人脸图片获取模块1910,获取用户指定的人脸图片。
人名获取模块,具体包括:
特征提取模块1920,提取人脸图片的特征,以及提取已收集人脸图片的特征。
特征比较模块1930,将人脸图片的特征与已收集人脸图片的特征进行比较。
人脸图片选择模块1940,根据比较结果从已收集人脸图片中选出人脸图片的相似人脸图片。
人名提取模块1950,从相似人脸图片对应的文本中,提取一个或多个人名。在本实施例中,对文本类型不进行限制,可以是图片所在新闻的标题、环绕文本等。
权重值计算模块1960,根据一个或多个人名的属性,为一个或多个人名计算权重值。在本实施例中,对属性不进行限制,其可以是人名出现的频次、位置等,因为不同频次、位置的人名就相似人脸图片对应的人名的可能性有所不同。
人名选择模块1970,根据权重值的高低,从一个或多个人名中选择相似人脸图片中人脸对应的人名。
人名确定模块1980,根据相似人脸图片中人脸的人名,确定人脸图片中人脸的人名。
信息获取模块1990,根据人名,从网络中获取人名对应的网络资源信息。
信息显示模块19100,将人名对应的网络资源信息显示给用户。
根据图19,对于相似人脸图片所在的新闻页面,首先对该图片对应的新闻标题、环绕文本预处理,然后分词;从分词结果中提取出候选人名,并与人名词表比较,去除非人名;根据每个人名出现的频率、位置,以及与其他词语的关系计算一个权重;如果权重大于某个阈值,则选择权重最大的名字作为该图片中人脸的人名,否则认为无法提取可靠的人名。
如图20所示,本发明的一个实施例中提供了一种人脸图片信息获取装置,其包括:
人脸图片获取模块2010,获取用户指定的人脸图片。
人名获取模块,具体包括:
特征提取模块2020,提取人脸图片的特征,以及提取已收集人脸图片的特征。
特征比较模块2030,将人脸图片的特征与已收集人脸图片的特征进行比较。
人脸图片选择模块2040,根据比较结果从已收集人脸图片中选出人脸图片的相似人脸图片。
相似度计算模块2050,根据比较结果,计算相似人脸图片与人脸图片之间的相似度。
相似度累加模块2060,累加相同人名对应的所有相似人脸图片的相似度。
人名确定模块2070,在累加后最大的相似度大于预设阈值时,按累加后的最大相似度对应的相似人脸图片中人脸的人名,确定人脸图片中人脸的人名。在本实施例中,基于累加后的相似度可以实现人脸的精确识别。
信息获取模块2080,根据人名,从网络中获取人名对应的网络资源信息。
信息显示模块2090,将人名对应的网络资源信息显示给用户。
根据图20,判断用户指定的是否属于已经建立好人名的数据库的某张人脸(已收集人脸图片):首先,通过人脸检测、特征提取和人名提取,建立起一个已知人名的人脸数据库;对于用户指定的新人脸图片,对图片进行人脸检测,如果没有人脸,则直接返回,如果有人脸,则提取人脸特征,并且量化为一个高维向量。将该输入图片的向量和库内所有人脸特征高维向量进行比较,计算其欧式距离,并取距离最近的前N个向量。这些向量所表征的人脸就是与该输入人脸最相似的人脸。如果人脸数据库过于庞大,逐个比较费时很长,可以事先对库内人脸进行聚类,然后只与聚好类的人脸进行比较;对于前N个相似人脸,以相似度为权重,计算每个名字的权重,相同名字的权值相加。并求出权重最高的名字。如果该名字大于一定阈值,则认为输入人脸属于该名字对应的人脸,否则认为无法精确识别出该人脸。
在此提供的算法和显示不与任何特定计算机、虚拟系统或者其它设备固有相关。各种通用系统也可以与基于在此的示教一起使用。根据上面的描述,构造这类系统所要求的结构是显而易见的。此外,本发明也不针对任何特定编程语言。应当明白,可以利用各种编程语言实现在此描述的本发明的内容,并且上面对特定语言所做的描述是为了披露本发明的最佳实施方式。
在此处所提供的说明书中,说明了大量具体细节。然而,能够理解,本发明的实施例可以在没有这些具体细节的情况下实践。在一些实例中,并未详细示出公知的方法、结构和技术,以便不模糊对本说明书的理解。
类似地,应当理解,为了精简本公开并帮助理解各个发明方面中的一个或多个,在上面对本发明的示例性实施例的描述中,本发明的各个特征有时被一起分组到单个实施例、图、或者对其的描述中。然而,并不应将该公开的方法解释成反映如下意图:即所要求保护的本发明要求比在每个权利要求中所明确记载的特征更多的特征。更确切地说,如下面的权利要求书所反映的那样,发明方面在于少于前面公开的单个实施例的所有特征。因此,遵循具体实施方式的权利要求书由此明确地并入该具体实施方式,其中每个权利要求本身都作为本发明的单独实施例。
本领域那些技术人员可以理解,可以对实施例中的设备中的模块进行自适应性地改变并且把它们设置在与该实施例不同的一个或多个设备中。可以把实施例中的模块或单元或组件组合成一个模块或单元或组件,以及此外可以把它们分成多个子模块或子单元或子组件。除了这样的特征和/或过程或者单元中的至少一些是相互排斥之外,可以采用任何组合对本说明书(包括伴随的权利要求、摘要和附图)中公开的所有特征以及如此
公开的任何方法或者设备的所有过程或单元进行组合。除非另外明确陈述,本说明书(包括伴随的权利要求、摘要和附图)中公开的每个特征可以由提供相同、等同或相似目的的替代特征来代替。
此外,本领域的技术人员能够理解,尽管在此所述的一些实施例包括其它实施例中所包括的某些特征而不是其它特征,但是不同实施例的特征的组合意味着处于本发明的范围之内并且形成不同的实施例。例如,在下面的权利要求书中,所要求保护的实施例的任意之一都可以以任意的组合方式来使用。
本发明的各个部件实施例可以以硬件实现,或者以在一个或者多个处理器上运行的软件模块实现,或者以它们的组合实现。本领域的技术人员应当理解,可以在实践中使用微处理器或者数字信号处理器(DSP)来实现根据本发明实施例的相似人脸图片获取装置中的一些或者全部部件的一些或者全部功能。本发明还可以实现为用于执行这里所描述的方法的一部分或者全部的设备或者装置程序(例如,计算机程序和计算机程序产品)。这样的实现本发明的程序可以存储在计算机可读介质上,或者可以具有一个或者多个信号的形式。这样的信号可以从因特网网站上下载得到,或者在载体信号上提供,或者以任何其他形式提供。
例如,图21示意性地示出了用于执行根据本发明的方法的计算设备的框图。该计算设备传统上包括处理器2110和以存储器2120形式的计算机程序产品或者计算机可读介质。存储器2120可以是诸如闪存、EEPROM(电可擦除可编程只读存储器)、EPROM、硬盘或者ROM之类的电子存储器。存储器2120具有用于执行上述方法中的任何方法步骤的程序代码2131的存储空间2130。例如,用于程序代码的存储空间2130可以包括分别用于实现上面的方法中的各种步骤的各个程序代码2131。这些程序代码可以从一个或者多个计算机程序产品中读出或者写入到这一个或者多个计算机程序产品中。这些计算机程序产品包括诸如硬盘,紧致盘(CD)、存储卡或者软盘之类的程序代码载体。这样的计算机程序产品通常为如参考图22所述的便携式或者固定存储单元。该存储单元可以具有与图21的计算设备中的存储器2120类似布置的存储段、存储空间等。程序代码可以例如以适当形式进行压缩。通常,存储单元包括用于执行根据本发明的方法步骤的计算机可读代码2131’,即可以由例如诸如2110之类的处理器读取的代码,这些代码当由计算设备运行时,导致该计算设备执行上面所描述的方法中的各个步骤。
应该注意的是上述实施例对本发明进行说明而不是对本发明进行限制,并且本领域技术人员在不脱离所附权利要求的范围的情况下可设计出替换实施例。在权利要求中,不应将位于括号之间的任何参考符号构造成对权利要求的限制。单词“包含”不排除存在未列在权利要求中的元件或步骤。位于元件之前的单词“一”或“一个”不排除存在多个这样的元件。本发明可以借助于包括有若干不同元件的硬件以及借助于适当编程的计算机来实现。在列举了若干装置的单元权利要求中,这些装置中的若干个可以是通过同一个硬件项来具体体现。单词第一、第二、以及第三等的使用不表示任何顺序。可将这些单词解释为名称。
此外,还应当注意,本说明书中使用的语言主要是为了可读性和教导的目的而选择的,而不是为了解释或者限定本发明的主题而选择的。因此,在不偏离所附权利要求书的范围和精神的情况下,对于本技术领域的普通技术人员来说许多修改和变更都是显而易见的。对于本发明的范围,对本发明所做的公开是说明性的,而非限制性的,本发明的范围由所附权利要求书限定。
本发明可以应用于计算机系统/服务器,其可与众多其它通用或专用计算系统环境或配置一起操作。适于与计算机系统/服务器一起使用的众所周知的计算系统、环境和/或配置的例子包括但不限于:个人计算机系统、服务器计算机系统、瘦客户机、厚客户机、手持或膝上设备、基于微处理器的系统、机顶盒、可编程消费电子产品、网络个人电脑、小型计算机系统、大型计算机系统和包括上述任何系统的分布式云计算技术环境,等等。
计算机系统/服务器可以在由计算机系统执行的计算机系统可执行指令(诸如程序模块)的一般语境下描述。通常,程序模块可以包括例程、程序、目标程序、组件、逻辑、数据结构等等,它们执行特定的任务或者实现特定的抽象数据类型。计算机系统/服务器可以在分布式云计算环境中实施,分布式云计算环境中,任务是由通过通信网络链接的远程处理设备执行的。在分布式云计算环境中,程序模块可以位于包括存储设备的本地或远程计算系统存储介质上。
本文中所称的“一个实施例”、“实施例”或者“一个或者多个实施例”意味着,结合实施例描述的特定特征、结构或者特性包括在本发明的至少一个实施例中。此外,请注意,这里“在一个实施例中”的词语例子不一定全指同一个实施例。
Claims (34)
- 一种相似人脸图片获取方法,其包括:获取用户指定的人脸图片;对所述人脸图片进行人脸识别,以从已收集人脸图片中识别出所述人脸图片的相似人脸图片;将所述相似人脸图片显示给所述用户。
- 根据权利要求1所述的方法,其中,在将所述相似人脸图片显示给所述用户之前,还包括:获取所述相似人脸图片中人脸的人名;获取所述相似人脸图片与所述人脸图片的相似度,并累加具有相同人名的所有相似人脸图片对应的相似度;将所述相似人脸图片显示给所述用户,具体包括:在累加后的最大相似度小于预定阈值时,根据所述相似人脸图片对应的相似度大小,从所述相似人脸图片中选择一张或多张人脸图片显示给所述用户。
- 根据权利要求1-2任一项所述的方法,其中,还包括:将所述一张或多张人脸图片对应的相似度和/或其中人脸的人名,显示给所述用户。
- 根据权利要求1-3任一项所述的方法,其中,还包括:根据所述一张或多张人脸图片对应的相似度高低,设置所述一张或多张人脸图片的显示顺序。
- 根据权利要求1-4任一项所述的方法,其中,还包括:根据所述一张或多张人脸图片对应的相似度高低,为所述一张或多张人脸图片设置相应的评价信息并显示给用户。
- 根据权利要求1-5任一项所述的方法,其中,根据所述相似人脸图片对应的相似度大小,从所述相似人脸图片中选择一张或多张人脸图片显示给所述用户,具体包括:按所述相似人脸图片对应的相似度高低顺序获取人脸图片;判断最新获取的人脸图片中人脸的人名,是否位于其他已获取的人脸图片中人脸的人名之中;在判断结果为否时,将所述最新获取的人脸图片显示给所述用户。
- 根据权利要求1-6任一项所述的方法,其中,对所述人脸图片进行人脸识别,以从已收集人脸图片中识别出所述人脸图片的相似人脸图片,具体包括:提取所述人脸图片的特征,以及提取已收集人脸图片的特征;将所述人脸图片的特征与所述已收集人脸图片的特征进行比较;根据比较结果从所述已收集人脸图片中选出所述相似人脸图片;获取所述相似人脸图片与所述人脸图片的相似度,具体包括:根据所述比较结果计算所述相似人脸图片与所述人脸图片的相似度。
- 根据权利要求1-7中任一项所述的方法,其中,获取所述相似人脸图片中人脸的人名,具体包括:从所述相似人脸图片对应的丈本中,提取一个或多个人名;根据所述一个或多个人名的属性,为所述一个或多个人名计算权重值;根据所述权重值的高低,从所述一个或多个人名中选择所述相似人脸图片中人脸的人名。
- 一种相似人脸图片获取装置,其包括:人脸图片获取模块,用于获取用户指定的人脸图片;人脸图片识别模块,用于对所述人脸图片进行人脸识别,以从已收集人脸图片中识别出所述人脸图片的相似人脸图片;人脸图片显示模块,用于将所述相似人脸图片显示给所述用户。
- 根据权利要求9所述的装置,其中,还包括:人名获取模块,用于获取所述相似人脸图片中人脸的人名;相似度获取模块,用于获取所述相似人脸图片与所述人脸图片的相似度,并累加具有相同人名的所有相似人脸图片对应的相似度;所述人脸图片显示模块在累加后的最大相似度小于预定阈值时,根据所述相似人脸图片对应的相似度大小,从所述相似人脸图片中选择一张或多张人脸图片显示给所述用户。
- 根据权利要求9-10任一项所述的装置,其中,还包括:相似度/人名显示模块,用于将所述一张或多张人脸图片对应的相似度和/或其中人脸的人名,显示给所述用户。
- 根据权利要求9-11任一项所述的装置,其中,还包括:显示顺序设置模块,用于根据所述一张或多张人脸图片对应的相似度高低,设置所述一张或多张人脸图片的显示顺序。
- 根据权利要求9-12任一项所述的装置,其中,还包括:评价信息显示模块,用于根据所述一张或多张人脸图片对应的相似度高低,为所述一张或多张人脸图片设置相应的评价信息并显示给用户。
- 根据权利要求9-13任一项所述的装置,其中,还包括:顺序获取模块,用于按所述相似人脸图片对应的相似度高低顺序获取人脸图片;人名判断模块,用于判断最新获取的人脸图片中人脸的人名,是否位于其他已获取的人脸图片中人脸的人名之中;所述人脸图片显示模块在判断结果为否时,将所述最新获取的人脸图片显示给所述用户。
- 根据权利要求9-14任一项所述的装置,其中,还包括:特征提取模块,用于提取所述人脸图片的特征,以及提取已史集人脸图片的特征;特征比较模块,用于将所述人脸图片的特征与所述已收集人脸图片的特征进行比较;所述人脸图片识别模块根据比较结果从所述已收集人脸图片中选出所述相似人脸图片;所述相似度获取模块根据所述比较结果计算所述相似人脸图片与所述人脸图片的相似度。
- 根据权利要求9-15任一项所述的装置,其中,所述人名获取模块,具体包括:人名提取模块,用于从所述相似人脸图片对应的丈本中,提取一个或多个人名;权重值计算模块,用于根据所述一个或多个人名的属性,为所述一个或多个人名计算权重值;人名选择模块,用于根据所述权重值的高低,从所述一个或多个人名中选择所述相似人脸图片中人脸的人名。
- 一种人脸图片信息获取方法,其包括:获取用户指定的人脸图片;对所述人脸图片进行人脸识别,以得到所述人脸图片中人脸的人名;根据所述人名,从网络中获取所述人名对应的网络资源信息;将所述人名对应的网络资源信息显示给所述用户。
- 根据权利要求17所述的方法,其中,根据所述人名,从网络中获取所述人名对应的网络资源信息,具体包括:从所述网络中的用于记录结构化信息的站点上,获取所述人名对应的结构化信息作为所述人名对应的网络资源信息,其中所述人名对应的结构化信息中包含多种条目对应的信息。
- 根据权利要求17-18任一项所述的方法,其中,还包括:按预设的多种条目,从所述人名对应的网络资源信息中筛选出所述多种条目对应的信息;根据所述多种条目,将所述提取的信息处理为结构化信息。
- 根据权利要求17-19任一项所述的方法,其中,所述预设的多种条目中包括社交账号条目;按预设的多种条目,从所述人名的网络资源信息中筛选出所述多种条目对应的信息,具体包括;按所述社交账号条目,查找所述人名对应的社交账号的标识信息和/或所述人名对应的社交账号所发布的内容。
- 根据权利要求17-20任一项所述的方法,其中,所述预设的多种条目中包括新闻消息条目;按预设的多种条目,从所述人名的网络资源信息中筛选出所述多种条目对应的信息,具体包括:按所述新闻信息条目,从所述网络中获取所述人名对应的新闻信息。
- 根据权利要求17-21中任一项所述的方法,其中,对所述人脸图片进行人脸识别,以得到所述人脸图片中人脸的人名,具体包括:提取所述人脸图片的特征,以及提取已收集人脸图片的特征;将所述人脸图片的特征与所述已收集人脸图片的特征进行比较;根据比较结果从所述已收集人脸图片中选出所述人脸图片的相似人脸图片;根据所述相似人脸图片中人脸的人名,确定所述人脸图片中人脸的人名。
- 根据权利要求17-22任一项所述的方法,其中,在根据所述相似人脸图片中人脸的人名,确定所述人脸图片中人脸的人之前,还包括:从所述相似人脸图片对应的丈本中,提取一个或多个人名;根据所述一个或多个人名的属性,为所述一个或多个人名计算权重值;根据所述权重值的高低,从所述一个或多个人名中选择所述相似人脸图片中人脸对应的人名。
- 根据权利要求17-23任一项所述的方法,其中,根据所述相似人脸图片中人脸的人名,确定所述人 脸图片中人脸的人名,具体包括:根据所述比较结果,计算所述相似人脸图片与所述人脸图片之间的相似度;累加相同人名对应的所有相似人脸图片的相似度;在累加后最大的相似度大于预设阈值时,按所述累加后的最大相似度对应的相似人脸图片中人脸的人名,确定所述人脸图片中人脸的人名。
- 一种人脸图片信息获取装置,其包括:人脸图片获取模块,用于获取用户指定的人脸图片;人名获取模块,用于对所述人脸图片进行人脸识别,以得到所述人脸图片中人脸的人名;信息获取模块,用于根据所述人名,从网络中获取所述人名对应的网络资源信息;信息显示模块,用于将所述人名对应的网络资源信息显示给所述用户。
- 根据权利要求25所述的装置,其中,所述信息获取模块从所述网络中的用于记录结构化信息的站点上,获取所述人名对应的结构化信息作为所述人名对应的网络资源信息,其中所述人名对应的结构化信息中包含多种条目对应的信息。
- 根据权利要求25-26任一项所述的装置,其中,还包括:所述信息获取模块按预设的多种条目,从所述人名对应的网络资源信息中筛选出所述多种条目对应的信息;以及根据所述多种条目,将所述提取的信息处理为结构化信息。
- 根据权利要求25-27任一项所述的装置,其中,所述预设的多种条目中包括社交账号条目;所述信息获取模块按所述社交账号条目,查找所述人名对应的社交账号的标识信息和/或所述人名对应的社交账号所发布的内容。
- 根据权利要求25-28任一项所述的装置,其中,所述预设的多种条目中包括新闻消息条目;所述信息获取模块按所述新闻信息条目,从所述网络中获取所述人名对应的新闻信息。
- 根据权利要求25-29中任一项所述的装置,其中,所述人名获取模块具体包括:特征提取模块,用于提取所述人脸图片的特征,以及提取已史集人脸图片的特征;特征比较模块,用于将所述人脸图片的特征与所述已收集人脸图片的特征进行比较;人脸图片选择模块,用于根据比较结果从所述已收集人脸图片中选出所述人脸图片的相似人脸图片;人名确定模块,用于根据所述相似人脸图片中人脸的人名,确定所述人脸图片中人脸的人名。
- 根据权利要求25-30任一项所述的装置,其中,所述人名获取模块还包括:人名提取模块,用于从所述相似人脸图片对应的丈本中,提取一个或多个人名;权重值计算模块,用于根据所述一个或多个人名的属性,为所述一个或多个人名计算权重值;人名选择模块,根据所述权重值的高低,从所述一个或多个人名中选择所述相似人脸图片中人脸对应的人名。
- 根据权利要求25-31任一项所述的装置,其中,所述人名获取模块还包括:相似度计算模块,用于根据所述比较结果,计算所述相似人脸图片与所述人脸图片之间的相似度;相似度累加模块,用于累加相同人名对应的所有相似人脸图片的相似度;所述人名确定模块在累加后最大的相似度大于预设阈值时,按所述累加后的最大相似度对应的相似人脸图片中人脸的人名,确定所述人脸图片中人脸的人名。
- 一种计算机程序,包括计算机可读代码,当所述计算机可读代码在计算设备上运行时,导致所述计算设备执行根据权利要求1-8中的任一项所述的相似人脸图片获取方法,或者导致所述计算设备执行根据权利要求17-24中的任一项所述的人脸图片信息获取方法。
- 一种计算机可读介质,其中存储了如权利要求33所述的计算机程序。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US15/539,644 US10489637B2 (en) | 2014-12-23 | 2015-11-27 | Method and device for obtaining similar face images and face image information |
Applications Claiming Priority (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410816768.6 | 2014-12-23 | ||
| CN201410816768.6A CN104463177A (zh) | 2014-12-23 | 2014-12-23 | 相似人脸图片获取方法和装置 |
| CN201410812175.2 | 2014-12-23 | ||
| CN201410812175.2A CN104537341B (zh) | 2014-12-23 | 2014-12-23 | 人脸图片信息获取方法和装置 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2016101766A1 true WO2016101766A1 (zh) | 2016-06-30 |
Family
ID=56149211
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2015/095861 Ceased WO2016101766A1 (zh) | 2014-12-23 | 2015-11-27 | 相似人脸图片获取和人脸图片信息获取方法和装置 |
Country Status (2)
| Country | Link |
|---|---|
| US (1) | US10489637B2 (zh) |
| WO (1) | WO2016101766A1 (zh) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114360029A (zh) * | 2022-01-12 | 2022-04-15 | 深圳市百川数安科技有限公司 | 针对内容社区的用户自拍识别方法及装置 |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108875989B (zh) * | 2018-06-29 | 2022-12-27 | 北京金山安全软件有限公司 | 基于人脸识别的预定方法、装置、计算机设备及存储介质 |
| CN109191184A (zh) * | 2018-08-14 | 2019-01-11 | 微梦创科网络科技(中国)有限公司 | 基于图像识别的广告投放方法及系统 |
| US11030487B2 (en) * | 2018-09-05 | 2021-06-08 | Vanderbilt University | Noise-robust neural networks and methods thereof |
| CN109858328B (zh) * | 2018-12-14 | 2023-06-02 | 上海集成电路研发中心有限公司 | 一种基于视频的人脸识别的方法及装置 |
| CN110942014B (zh) * | 2019-11-22 | 2023-04-07 | 浙江大华技术股份有限公司 | 人脸识别快速检索方法、装置、服务器及存储装置 |
| CN111368772B (zh) * | 2020-03-11 | 2023-08-22 | 杭州海康威视系统技术有限公司 | 身份识别方法、装置、设备及存储介质 |
| US11783628B2 (en) * | 2021-05-18 | 2023-10-10 | ID R&D Inc. | Method and computing device for performing a crowdsourcing task |
| CN115880761B (zh) * | 2023-02-09 | 2023-05-05 | 数据空间研究院 | 基于策略优化的人脸识别方法、系统、存储介质及应用 |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070172155A1 (en) * | 2006-01-21 | 2007-07-26 | Elizabeth Guckenberger | Photo Automatic Linking System and method for accessing, linking, and visualizing "key-face" and/or multiple similar facial images along with associated electronic data via a facial image recognition search engine |
| US20080052312A1 (en) * | 2006-08-23 | 2008-02-28 | Microsoft Corporation | Image-Based Face Search |
| CN102682091A (zh) * | 2012-04-25 | 2012-09-19 | 腾讯科技(深圳)有限公司 | 基于云服务的视觉搜索方法和系统 |
| CN102880729A (zh) * | 2012-11-02 | 2013-01-16 | 深圳市宜搜科技发展有限公司 | 基于人脸检测与识别的人物图片索引方法及装置 |
| CN103425709A (zh) * | 2012-05-25 | 2013-12-04 | 致伸科技股份有限公司 | 相片影像管理方法及相片影像管理系统 |
| CN104091164A (zh) * | 2014-07-28 | 2014-10-08 | 北京奇虎科技有限公司 | 人脸图片人名识别方法和系统 |
| CN104463177A (zh) * | 2014-12-23 | 2015-03-25 | 北京奇虎科技有限公司 | 相似人脸图片获取方法和装置 |
| CN104537341A (zh) * | 2014-12-23 | 2015-04-22 | 北京奇虎科技有限公司 | 人脸图片信息获取方法和装置 |
Family Cites Families (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR100442834B1 (ko) * | 2002-07-19 | 2004-08-02 | 삼성전자주식회사 | 얼굴/유사얼굴 영상으로 학습된 패턴 분류기를 이용한얼굴 검출 방법 및 시스템 |
| US20090091629A1 (en) | 2007-10-03 | 2009-04-09 | Casey Robert J | TV/Movie actor identifcation device |
| US8385971B2 (en) * | 2008-08-19 | 2013-02-26 | Digimarc Corporation | Methods and systems for content processing |
| US8670597B2 (en) | 2009-08-07 | 2014-03-11 | Google Inc. | Facial recognition with social network aiding |
| US8768313B2 (en) * | 2009-08-17 | 2014-07-01 | Digimarc Corporation | Methods and systems for image or audio recognition processing |
| EP2537124A4 (en) * | 2010-02-17 | 2017-08-16 | Photoccino Ltd | System and method for creating a collection of images |
| US9465993B2 (en) * | 2010-03-01 | 2016-10-11 | Microsoft Technology Licensing, Llc | Ranking clusters based on facial image analysis |
| US20120183270A1 (en) * | 2011-01-14 | 2012-07-19 | International Business Machines Corporation | Webcam captcha |
| US20120222100A1 (en) * | 2011-02-24 | 2012-08-30 | International Business Machines Corporation | Advanced captcha using integrated images |
| US8380711B2 (en) * | 2011-03-10 | 2013-02-19 | International Business Machines Corporation | Hierarchical ranking of facial attributes |
| EP2783304B1 (en) * | 2011-11-24 | 2017-12-20 | Microsoft Technology Licensing, LLC | Reranking using confident image samples |
| US8984622B1 (en) * | 2012-01-17 | 2015-03-17 | Amazon Technologies, Inc. | User authentication through video analysis |
| TWI488128B (zh) | 2012-03-07 | 2015-06-11 | Altek Corp | 人臉辨識系統及其人臉辨識方法 |
| CN102629275B (zh) | 2012-03-21 | 2014-04-02 | 复旦大学 | 面向跨媒体新闻检索的人脸-人名对齐方法及系统 |
| JP5924114B2 (ja) * | 2012-05-15 | 2016-05-25 | ソニー株式会社 | 情報処理装置、情報処理方法、コンピュータプログラムおよび画像表示装置 |
| US9177194B2 (en) * | 2014-01-29 | 2015-11-03 | Sony Corporation | System and method for visually distinguishing faces in a digital image |
| CN103793697B (zh) | 2014-02-17 | 2018-05-01 | 北京旷视科技有限公司 | 一种人脸图像的身份标注方法及人脸身份识别方法 |
| US20170116465A1 (en) * | 2015-10-26 | 2017-04-27 | Le Holdings (Beijing) Co., Ltd. | Video processing method and system, video player and cloud server |
| US10331945B2 (en) * | 2015-12-22 | 2019-06-25 | Intel Corporation | Fair, secured, and efficient completely automated public Turing test to tell computers and humans apart (CAPTCHA) |
| CN106096373A (zh) * | 2016-06-27 | 2016-11-09 | 旗瀚科技股份有限公司 | 机器人与用户的交互方法及装置 |
| US10089521B2 (en) * | 2016-09-02 | 2018-10-02 | VeriHelp, Inc. | Identity verification via validated facial recognition and graph database |
| US20180069937A1 (en) * | 2016-09-02 | 2018-03-08 | VeriHelp, Inc. | Event correlation and association using a graph database |
-
2015
- 2015-11-27 WO PCT/CN2015/095861 patent/WO2016101766A1/zh not_active Ceased
- 2015-11-27 US US15/539,644 patent/US10489637B2/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070172155A1 (en) * | 2006-01-21 | 2007-07-26 | Elizabeth Guckenberger | Photo Automatic Linking System and method for accessing, linking, and visualizing "key-face" and/or multiple similar facial images along with associated electronic data via a facial image recognition search engine |
| US20080052312A1 (en) * | 2006-08-23 | 2008-02-28 | Microsoft Corporation | Image-Based Face Search |
| CN102682091A (zh) * | 2012-04-25 | 2012-09-19 | 腾讯科技(深圳)有限公司 | 基于云服务的视觉搜索方法和系统 |
| CN103425709A (zh) * | 2012-05-25 | 2013-12-04 | 致伸科技股份有限公司 | 相片影像管理方法及相片影像管理系统 |
| CN102880729A (zh) * | 2012-11-02 | 2013-01-16 | 深圳市宜搜科技发展有限公司 | 基于人脸检测与识别的人物图片索引方法及装置 |
| CN104091164A (zh) * | 2014-07-28 | 2014-10-08 | 北京奇虎科技有限公司 | 人脸图片人名识别方法和系统 |
| CN104463177A (zh) * | 2014-12-23 | 2015-03-25 | 北京奇虎科技有限公司 | 相似人脸图片获取方法和装置 |
| CN104537341A (zh) * | 2014-12-23 | 2015-04-22 | 北京奇虎科技有限公司 | 人脸图片信息获取方法和装置 |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114360029A (zh) * | 2022-01-12 | 2022-04-15 | 深圳市百川数安科技有限公司 | 针对内容社区的用户自拍识别方法及装置 |
Also Published As
| Publication number | Publication date |
|---|---|
| US20180005022A1 (en) | 2018-01-04 |
| US10489637B2 (en) | 2019-11-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104537341B (zh) | 人脸图片信息获取方法和装置 | |
| US10242250B2 (en) | Picture ranking method, and terminal | |
| JP6759844B2 (ja) | 画像を施設に対して関連付けるシステム、方法、プログラム及び装置 | |
| US8422747B1 (en) | Finding untagged images of a social network member | |
| CN106055574B (zh) | 一种识别非法统一资源标识符url的方法与装置 | |
| US8649602B2 (en) | Systems and methods for tagging photos | |
| EP4057163B1 (en) | Facilitating use of images as search queries | |
| CN107562742B (zh) | 一种图像数据处理方法及装置 | |
| US10489637B2 (en) | Method and device for obtaining similar face images and face image information | |
| CN104463177A (zh) | 相似人脸图片获取方法和装置 | |
| US20130243249A1 (en) | Electronic device and method for recognizing image and searching for concerning information | |
| US9251395B1 (en) | Providing resources to users in a social network system | |
| CN104091164A (zh) | 人脸图片人名识别方法和系统 | |
| US20240037142A1 (en) | Systems and methods for filtering of computer vision generated tags using natural language processing | |
| WO2015188719A1 (zh) | 结构化数据与图片的关联方法与关联装置 | |
| WO2020155877A1 (zh) | 信息推荐 | |
| US11423057B2 (en) | Impression tagging system for locations | |
| CN103617192B (zh) | 一种数据对象的聚类方法和装置 | |
| WO2015196964A1 (zh) | 搜索匹配图片的方法、图片搜索方法及装置 | |
| WO2016058520A1 (zh) | 人脸图片人名识别方法和装置 | |
| CN115098729A (zh) | 视频处理方法、样本生成方法、模型训练方法及装置 | |
| US20160351185A1 (en) | Voice recognition device and method | |
| CN108009251A (zh) | 一种图像文件搜索方法及装置 | |
| CN104268504B (zh) | 图片识别方法和装置 | |
| CN110765299A (zh) | 一种网络资产确定方法、装置、电子设备和存储介质 |
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: 15871835 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 15539644 Country of ref document: US |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 15871835 Country of ref document: EP Kind code of ref document: A1 |