CN112732966A - Method for adding shooting place to photo - Google Patents
Method for adding shooting place to photo Download PDFInfo
- Publication number
- CN112732966A CN112732966A CN202011624068.9A CN202011624068A CN112732966A CN 112732966 A CN112732966 A CN 112732966A CN 202011624068 A CN202011624068 A CN 202011624068A CN 112732966 A CN112732966 A CN 112732966A
- Authority
- CN
- China
- Prior art keywords
- shooting
- library
- photo
- head portrait
- time
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 31
- 208000029152 Small face Diseases 0.000 description 2
- 230000001815 facial effect Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
- G06F16/5866—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using information manually generated, e.g. tags, keywords, comments, manually generated location and time information
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
- G06F16/55—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Databases & Information Systems (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Library & Information Science (AREA)
- Studio Devices (AREA)
Abstract
A method of adding a shooting location to a photograph, comprising the steps of: the first step is as follows: identifying the shooting time, the shooting place and the shooting equipment of the photos in the photo library, and determining a time axis according to the shooting time of the photos shot at the same shooting place; the second step is as follows: selecting a first person in the photos shot at the same shooting place, and marking the photos, which are shot at the time axis or close to the time axis, of the non-shooting places containing the first person to obtain a first marked photo; the third step: identifying shooting equipment of the first marked picture, and marking pictures which are shot by the shooting equipment in the picture library and have shooting time in a time axis or close to the time axis to obtain a second marked picture; the fourth step: according to the shooting position of the picture with the shooting position in the time axis, the shooting position is added to the first marked picture and the second marked picture.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a method for adding a shooting place to a photo.
Background
Along with the rapid development of science and technology and the improvement of people's quality of life, people often can utilize cell-phone, camera to shoot work scene, record life etc. then, through the internet transmission or share the photo of oneself shooting in work and life, along with mobile terminal or the storage space on the storage device is bigger and bigger, makes the user store a large amount of photos, convenient browsing and use at any time.
At present, when a user views pictures through a terminal gallery, a great number of pictures are faced, and the classification method is too general, so that the user cannot quickly find the pictures which the user wants to view, in the prior art, a method for classifying according to shooting time or shooting place exists, however, if the classification is carried out according to the method, the user probably does not remember the shooting time or shooting place of the pictures after a period of time, even if the pictures at a certain time or a certain place are put together, the user is difficult to put all the pictures of the same person at different times or different places in one picture group, so that the user needs to manually view the pictures one by one when viewing in the later period, and more time is spent, therefore, great inconvenience is brought to the user, and the searching efficiency is reduced; sometimes, the children, moving objects and the like are captured, focusing may not be achieved in time or the images are shaken by hands, so that the images are not clear, the quality of the images is poor, and the images need to be manually selected or deleted and sorted in the later period, so that a large amount of time is spent, the operation is increased, and the user experience is poor. Many photos may not have shooting places, so that the photos cannot be classified according to address information, and the photos need to be manually checked one by one when being checked in the later period of a user, thereby spending more time, bringing great inconvenience to the user and reducing the searching efficiency. When different devices are adopted for shooting, due to different default formats of various devices, photos of different devices cannot be integrated and divided according to the same standard through the existing classification method.
In order to solve the problem that some pictures have no shooting location and cannot be classified according to the shooting location, a method for adding a shooting location to a picture is urgently needed.
Disclosure of Invention
A method of adding a shooting location to a photograph, comprising the steps of: first step S11: identifying the shooting time, the shooting place and the shooting equipment of the photos in the photo library, and determining a time axis according to the shooting time of the photos shot at the same shooting place; second step S12: selecting a first person in the photos shot at the same shooting place, and marking the photos, which are shot at the time axis or close to the time axis and contain the shooting place of the first person, to obtain a first marked photo; third step S13: identifying shooting equipment of the first marked photo, and marking photos which are shot by the shooting equipment in the photo library and have shooting time in the time axis or close to the time axis to obtain a second marked photo; fourth step S14: and adding the shooting positions in the first marked picture and the second marked picture according to the shooting positions of the pictures with the shooting positions in the time axis.
According to an embodiment of the present invention, the second step S12 further includes, when other people are included in the photo taken at the same location and/or in the first tagged photo, selecting the other people, and repeating the second step S12 to expand the first tagged photo until all people in the photo taken at the same location and/or in the first tagged photo are selected.
According to an embodiment of the present invention, the third step S13 further includes, when other people are included in the photo taken at the same location and/or in the first tagged photo and/or in the second tagged photo, selecting the other people, and repeating the second step S12 and the third step S13 until all people in the photo taken at the same location and/or in the first tagged photo and/or in the second tagged photo are selected.
According to one embodiment of the present invention, in the first step S11, 0 point of the shooting date of the photo having the same shooting location and the earliest shooting time in the photo library is taken as the start time of the time axis; the 24 points of the shooting date of the photograph having the same shooting location and the latest shooting time are taken as the end time of the time axis.
According to one embodiment of the invention, the time axis is set to a time span of not more than 3 days.
According to an embodiment of the present invention, in the fourth step S14, when there are a plurality of persons in the picture with shooting location information shot on the time axis, the person who has shot the most pictures is selected as the first person.
According to one embodiment of the invention, the shooting device can be a single lens, a mini-lens, a card machine or a mobile phone.
In one embodiment of the present invention, in the second step S12, a first marked picture is obtained by marking a picture of a non-shooting location including the first person, the shooting time of which is within the time axis or within a day close to the time axis.
According to an embodiment of the present invention, in the third step S13, the photos in the photo library that were captured by the capturing device and whose capturing time is within the time axis or within one day close to the time axis are marked, and a second marked photo is obtained
The invention can quickly and accurately add the shooting places to a large number of photos without shooting places in the photo library, can solve the problem that the photos cannot be classified according to the shooting places because the photo library contains a large number of photos without shooting places, and saves time for viewing the photos by people in a large amount.
Drawings
FIG. 1 is a flow chart of a method of adding a shooting location to a photograph;
FIG. 2 is a flowchart of a photo classification method.
Detailed Description
In the following detailed description of the preferred embodiments of the invention, reference is made to the accompanying drawings that form a part hereof, and in which is shown by way of illustration, specific features of the invention, such that the advantages and features of the invention may be more readily understood and appreciated. The following description is an embodiment of the claimed invention, and other embodiments related to the claims not specifically described also fall within the scope of the claims.
Fig. 1 shows a flow chart of a method for adding a shooting location to a photo.
As shown in fig. 1, a method for adding a shooting location to a photo includes the steps of: first step S11: identifying the shooting time, the shooting place and the shooting equipment of the photos in the photo library, and determining a time axis according to the shooting time of the photos shot at the same shooting place; second step S12: selecting a first person in the photos shot at the same shooting place, and marking the photos, which are shot at the time axis or close to the time axis and contain the shooting place of the first person, to obtain a first marked photo; third step S13: identifying shooting equipment of the first marked photo, and marking photos which are shot by the shooting equipment in the photo library and have shooting time in the time axis or close to the time axis to obtain a second marked photo; fourth step S14: and adding the shooting positions in the first marked picture and the second marked picture according to the shooting positions of the pictures with the shooting positions in the time axis.
Setting a time axis for selecting photos taken at the same shooting location, wherein the time axis can be set according to actual conditions, for example, the shooting date of the photo with the same shooting location and the earliest shooting time in a photo library can be used as the starting time of the time axis; the shooting date of the picture with the same shooting place and the latest shooting time is used as the end time of the time axis, the time range of the time axis is generally set to be 1-3 days, and if the time period for shooting the picture in the same shooting place exceeds 3 days, the shooting places can be added to the picture in batches according to the sequence of the shooting time.
When the first person is selected, if there are a plurality of persons in the picture with shooting location information shot in the time axis, the person who has shot the most pictures can be selected as the first person.
When the shooting devices for reading all marked photos and selecting and marking corresponding pictures according to all the shooting devices, the photos shot in a range beyond the time axis can be marked, but the time beyond the range can be set according to actual conditions, for example, the time within 2 days of the time axis can be set to be not beyond the time range, for example, the start time of the time axis is 10 months, 3 am, 10 am, and the end time of the time axis is 10 months, 5 am, 10 am, then the shooting devices can mark the photos shot between 10 months, 1 am, 10 am, and 7 am, 10 am, when the shooting devices mark photos without shooting places with similar time, most of the photos without shooting places shot at a certain place can be marked and selected, and thus shooting places are added.
When the other persons in the photos taken at the same location are used as the first person for photo marking, the photos taken in the range beyond the time axis can be marked, and the time beyond the range can be set according to the actual situation, for example, the time within 1 day of the time axis can be set to be not beyond the time range within 1 day of the time axis, for example, the start time of the time axis is 10 months, 1 am, 10 hours, and the end time is 10 months, 3 am, 10 hours, the shooting device can mark the photos without shooting location information taken at 9 am, 10 months, 5 pm, 4 hours, and can select most of the photos without shooting locations taken at a certain location by photo marking through the persons.
During working, recognizing the shooting time, the shooting location and the shooting equipment of the photos in the photo library, setting the starting time and the ending time of the photos with the shooting location information shot in the same shooting location as the starting time and the ending time of a time axis, selecting a first person in the photos shot in the same shooting location, and marking all the photos containing the first person, the shooting time of which is in the time axis or close to the time axis; reading the shooting equipment information of the photos, determining the shooting equipment used by the first marked photo, and marking all the photos shot by the shooting equipment with shooting time in the time axis or close to the time axis; and adding the shooting positions in the first marked picture and the second marked picture according to the shooting positions of the pictures with the shooting positions in the time axis.
The invention can define the photos shot at the same place through identifying the shooting time and the shooting place of the photos in the photo library and selecting the shooting time of the photos shot at the same shooting place to define the time axis, and can accurately select the photos shot at the same place; by selecting the first person, the photo containing the first person can be added to the shooting place, and the photo of the first person shot in the same place can be accurately identified; the method can quickly and accurately add the shooting places to a large number of photos without shooting places in the photo library, can solve the problem that the photos cannot be classified according to the shooting places because the photos without the shooting places are contained in the photo library, and saves a large amount of time for people to check the photos.
According to an embodiment of the present invention, the second step S12 further includes, when other people are included in the photo taken at the same location and/or in the first tagged photo, selecting the other people, and repeating the second step S12 to expand the first tagged photo until all people in the photo taken at the same location and/or in the first tagged photo are selected.
According to an embodiment of the present invention, the third step S13 further includes, when other people are included in the photo taken at the same location and/or in the first tagged photo and/or in the second tagged photo, selecting the other people, and repeating the second step S12 and the third step S13 until all people in the photo taken at the same location and/or in the first tagged photo and/or in the second tagged photo are selected.
According to one embodiment of the present invention, in the first step S11, 0 point of the shooting date of the photo having the same shooting location and the earliest shooting time in the photo library is taken as the start time of the time axis; the 24 points of the shooting date of the photograph having the same shooting location and the latest shooting time are taken as the end time of the time axis.
According to one embodiment of the invention, the time axis is set to a time span of not more than 3 days.
According to an embodiment of the present invention, in the fourth step S14, when there are a plurality of persons in the picture with shooting location information shot on the time axis, the person who has shot the most pictures is selected as the first person.
According to one embodiment of the invention, the shooting device can be a single lens, a mini-lens, a card machine or a mobile phone.
In one embodiment of the present invention, in the second step S12, a first marked picture is obtained by marking a picture of a non-shooting location including the first person, the shooting time of which is within the time axis or within a day close to the time axis.
According to an embodiment of the present invention, in the third step S13, the photos in the photo library that were captured by the capturing device and whose capturing time is within the time axis or within one day close to the time axis are marked, and a second marked photo is obtained
The invention can quickly and accurately add the shooting places to a large number of photos without shooting places in the photo library, can solve the problem that the photos cannot be classified according to the shooting places because the photo library contains a large number of photos without shooting places, and saves time for viewing the photos by people in a large amount.
Fig. 2 shows a flow chart of a photo classification method.
As shown in fig. 2, a photo classification method includes a first step S21 of reading photo information in a photo library, cutting a face in a photo into an independent avatar picture, and establishing a correspondence between the avatar picture and the photo; the second step S22, identifying the quality of the head portrait picture, and classifying according to the quality of the head portrait picture to form a high-quality library and a low-quality library; a third step S23, comparing the similarity of the head portrait pictures in the high-quality library, and dividing the high-quality library into one or more face libraries and similar libraries corresponding to the face libraries one by one according to the similarity of the head portrait pictures; a fourth step S24, comparing the similarity of the head portrait pictures with the similar shooting time in the face library and the corresponding similar library, merging the similar library and the corresponding face library if the head portrait pictures are judged to be the same person, and constructing the similar library as a new face library if the head portrait pictures are not judged to be the same person; a fifth step S25, comparing the similarity between the feature of the head portrait picture in the low-quality library and the feature union of the head portrait pictures in the face library, and if the head portrait picture in the low-quality library is identified as the same person, moving the head portrait picture in the low-quality library into the corresponding face library; a sixth step S26, classifying the photos according to the face library.
In the first step S21, the information of the photos in the photo library is read, so as to perform face position recognition, determine the position of the face in the avatar, recognize the face contour in the photos to distinguish from the photo background, thereby cutting the avatar in the photos to form individual face avatar pictures, making each avatar picture as an individual picture for comparison, and associating the avatar picture with the original photo, so that the face in the photos is not repeatedly recognized when performing the similarity comparison of the avatar pictures, different avatar pictures can be directly compared without comparing the image contents irrelevant to the face avatar, the workload of the similarity comparison of the avatar pictures is reduced, the work efficiency and accuracy of the contrast of the avatar pictures are improved, the avatar determined as the same person is moved into the same face library without moving the original photo, the similarity comparison is convenient to continue in the later stage.
In the second step S22, the image quality may be identified by dividing the image into a high quality library and a low quality library according to the quality of the image, where the quality of the head portrait image mainly includes the definition of the face, the face inclination, and the like in the head portrait image, and the high quality library is put into the head portrait image with good face definition and small face inclination, or the low quality library is put into the head portrait image with good face definition and small face inclination. The purpose of setting the high-quality library and the low-quality library is to separately compare the head portrait picture in the high-quality library with the head portrait picture in the low-quality library, compare the head portrait picture in the high-quality library with the head portrait picture in the low-quality library for similarity judgment, and set different judgment standards when the head portrait picture is the same person, the head portrait picture in the high-quality library can be compared for similarity, then the head portrait picture in the low-quality library is compared for similarity, the head portrait picture in the high-quality library is set for the same person for similarity judgment standard higher, the low-quality library is set for the same person for similarity judgment standard lower, because the head portrait picture in the low-quality library has lower definition or inclination of the face angle, when the head portrait picture in the low-quality library is compared for similarity judgment, the lower judgment standard is set, and the accuracy of head portrait picture judgment can be improved, the head portrait pictures in the low-quality library can be prevented from generating wrong judgment due to low definition or large face gradient.
When similarity comparison is performed on head portrait pictures in a high-quality face library, there may be two methods, that is, the first method: a face library may be created first, and whether the same person is determined by comparing the feature of the head portrait picture in the high-quality library with the feature combination of the head portrait pictures in the face library, for example, a first face library may be created first, and the first head portrait picture may be moved into the first face library, the feature of the second head portrait picture may be compared with the feature combination of the head portrait picture in the first face library in similarity, if the same person is determined, the second head portrait picture may be moved into the first face library, if the person is determined to be similar, a similarity library corresponding to the current face library may be created newly, and the second head portrait picture may be moved into the similarity library corresponding to the first face library, if the person is not determined, the second face library may be created newly, and the second head portrait picture may be moved into the second face library, and then the feature of the nth head portrait picture of the third head portrait picture … … may be compared with the feature union of the head portrait pictures in each face library continuously, if the Nth head portrait picture is judged to be the same person with the face libraries, the Nth head portrait picture is moved into the face library with the highest similarity, information of other face libraries judged to be the same person with the head portrait picture is recorded, for example, when the Nth head portrait picture is judged to be the same person with the 2-person face library, the Nth head portrait picture is moved into the face library with the highest similarity, information of another face library judged to be the same person with the N-th head portrait picture is recorded, when the Nth head portrait picture is judged to be the same person with the 4-person face library, the Nth head portrait picture is moved into the face library with the highest similarity, and information of other 3 face libraries judged to be the same person with the N-th head portrait picture is recorded. The criterion of whether the same person is determined may be set according to circumstances, for example, the same person may be determined when the similarity between the feature of the avatar picture and the feature set of the avatar picture in the face library is set to be greater than or equal to 70%, the same person may be determined when the similarity between the feature of the avatar picture and the feature set of the avatar picture in the face library is set to be less than 60%, and the same person may be determined when the similarity between the feature of the avatar picture and the feature set of the avatar picture in the face library is set to be greater than or equal to 60% and less than 70%.
The union of the features in the face library refers to useful features in all the head portrait pictures selected in the face library, and useless features in the head portrait pictures are discarded, so that a combination closest to the real face features is formed. When the method is in operation, when a new head image enters the face library, comparing the image characteristics of the new head image with the combination of the original head image characteristics in the face library, and selecting useful characteristics in the new head image to continuously perfect the original characteristic combination, for example, 100 characteristics exist in the combination of the original head image characteristics, 50 characteristics exist in the newly moved head image, 20 characteristics of the 50 characteristics repeat with 100 characteristics in the combination of the original head image characteristics, and 10 characteristics are useless characteristics, then the remaining 20 characteristics in the new head image are incorporated into the combination of the head image characteristics, and when the head image moves, the characteristic combination of the head image in the face library is continuously perfect.
When the head portrait picture is compared with the face library, the comparison result is more accurate by comparing the feature of the head portrait picture with the feature union of the head portrait pictures in the face library, because the feature union of the head portrait pictures in the face library is closer to the real face feature relative to the feature of a single head portrait picture in the face library, the comparison is carried out by taking the feature union of the head portrait pictures in the face library as a standard, and compared with the comparison carried out by taking a certain picture in the face library as a standard, the comparison result is higher in accuracy.
When the feature of the head portrait picture is compared with the feature union of the head portrait pictures in the face library in a similarity way, the feature union of a plurality of head portrait pictures in the face library is formed by combining the features of the head portrait pictures, when a new head portrait picture is moved in the face library, the face library can adjust the feature point union of the face library according to the head portrait feature point in the newly moved head portrait picture, and the similarity comparison is carried out according to the feature union of the head portrait picture and the feature union of the head portrait pictures in the face library, so that whether the head portrait pictures are the same person or not is judged.
And the second method comprises the steps of comparing every two head portrait pictures in a high-quality library in sequence, after the comparison of all head portrait pictures in the high-quality library is finished, moving the head portrait picture which is judged to be the same person into the same face library, and moving the head portrait picture which is judged to be similar to the head portrait picture in a certain face library into a similar library corresponding to the face library, so that one or more face libraries and similar libraries corresponding to the face libraries one to one are obtained. The criterion of whether the two avatar pictures are determined to be the same person or not may be set according to circumstances, for example, the same person may be determined if the similarity between the two avatar pictures is greater than or equal to 70%, the same person may be determined if the similarity between the two avatar pictures is less than 60%, and the same person may be determined if the similarity between the two avatar pictures is greater than or equal to 60% and less than 70%.
When comparing the avatar pictures, similarity comparison can be performed in sequence according to the sequence of the avatar pictures, such as shooting time sequence, size sequence and the like of the avatar pictures, and the avatar pictures can also be randomly selected from the avatar picture library for comparison.
By the method, the high-quality library is divided into the face library and the similar libraries, wherein the similar libraries correspond to the face library one by one, each similar library corresponds to one face library, and the similarity between the face pictures in the similar libraries and the contrast standard is smaller than the similarity between the face pictures in the face library and the contrast standard.
When a plurality of head portrait pictures with similar shooting time are selected, the selection can be performed according to the number of the head portrait pictures shot in the time period, for example, when the number of the head portrait pictures shot in the time period is small, the head portrait pictures can be simultaneously selected, for example, when the number of the head portrait pictures shot in the time period is large, a plurality of head portrait pictures, for example, 3-10 head portrait pictures can be selected.
When the face library is compared with the corresponding similar library, after the head portrait picture in the face library is compared with the head portrait picture in the corresponding similar library, if the head portrait picture is judged to be the same person, the similar library and the corresponding face library are merged, and if the head portrait picture is judged not to be the same person, the similar library is constructed into a new face library. When the face libraries are compared with the corresponding similar libraries, the standard for judging whether the face libraries are the same person can be set according to the situation, for example, the similarity of a plurality of head portrait pictures with similar shooting time in the two face libraries can be set to be more than or equal to 75 percent, the face libraries are determined as the same person, and otherwise, the face libraries are determined as not the same person.
When the face library is compared with the corresponding similar libraries, head portrait pictures with similar shooting time and the same quantity are respectively selected from the face library and the corresponding similar libraries for comparison, when the quantity of the head portrait pictures in at least one of the two libraries is less than three, all the head portrait pictures in the library with the less quantity of the head portrait pictures are selected, and the head portrait pictures with the similar shooting time and the same quantity are selected from the other library for comparison; when the number of the head portrait pictures in the two libraries is more than three, 3-10 head portrait pictures with similar shooting time and the same number are selected from the two libraries for comparison, when the number of the head portrait pictures in one library in the two libraries is less than a preset value, all the head portrait pictures in the library with the smaller number of the head portrait pictures are selected from the other library for comparison, for example, 1 head portrait picture in the similar library and 2 head portrait pictures in the face library are selected, 1 head portrait picture in the similar library is selected, and 1 head portrait picture in the face library and the head portrait pictures in the similar library in the shooting time are selected for comparison, for example, 15 head portrait pictures in the similar library and 30 head portrait pictures in the face library corresponding to the head portrait pictures in the similar library are selected, 3-10 head portrait pictures with the similar shooting time can be selected from the similar library, and head portrait pictures which are similar to the head portrait pictures selected in the similar library in shooting time and are the same in number are selected from the face library corresponding to the similar library for comparison.
When two face libraries are compared, head portrait pictures with similar shooting time and the same quantity are respectively selected from the two face libraries for comparison, when the quantity of the head portrait pictures in at least one of the two libraries is less than three, all the head portrait pictures in the library with the less head portrait pictures are selected, and head portrait pictures with similar shooting time and the same quantity are selected from the other library for comparison; when the number of the head portrait pictures in the two libraries is more than three, 3-10 head portrait pictures with similar shooting time and the same number are selected from the two libraries for comparison, when the number of the head portrait pictures in one library in the two libraries is less than a preset value, all the head portrait pictures in the library with the smaller number of the head portrait pictures are selected from the other library for comparison, for example, 1 head portrait picture in the similar library and 2 head portrait pictures in the face library are selected, 1 head portrait picture in the similar library is selected, and 1 head portrait picture in the face library and the head portrait pictures in the similar library in the shooting time are selected for comparison, for example, 15 head portrait pictures in the similar library and 30 head portrait pictures in the face library corresponding to the head portrait pictures in the similar library are selected, 3-10 head portrait pictures with the similar shooting time can be selected from the similar library, and head portrait pictures which are similar to the head portrait pictures selected in the similar library in shooting time and are the same in number are selected from the face library corresponding to the similar library for comparison.
When the face libraries are compared, other face libraries which are recorded by the head portrait pictures in the face libraries and judged as the same person as the head portrait pictures are preferentially selected for comparison, and after a plurality of face libraries which are possibly the same person are compared and combined, the face libraries are continuously compared with the other face libraries, so that the workload can be reduced, and the accuracy and the efficiency of judging the head portrait pictures are improved.
When the two face libraries are compared, the standard for judging whether the two face libraries are the same person can be set according to the situation, for example, the similarity of the head portrait pictures in the two face libraries can be set to be more than or equal to 75%, the two face libraries are determined to be the same person, otherwise, the two face libraries are determined to be the same person and are merged.
When the two face libraries are compared, or the face libraries and the corresponding similarity libraries are compared, if a plurality of head portrait pictures are selected from the two libraries for comparison, the comparison can be sequentially performed, and the average value of the similarity is taken as a judgment basis, for example, 3 head portrait pictures A1, A2 and A3 with similar shooting time are selected from the A face library, 3 head portrait pictures B1, B2 and B3 with similar shooting time in the B face library and the A face library in the same time period are selected, the A1 is sequentially compared with B1, B2 and B3, the A2 is sequentially compared with the B1, B2 and B3, the A3 is sequentially compared with the B1, B2 and B3, the similarity is calculated according to the similarity of all head portrait pictures after comparison, when the average value of the head portrait pictures is more than or equal to 75%, the two face libraries are judged as the same person, the two face libraries are merged, otherwise they are not merged. Because the facial feature change of the head portrait pictures of the same person with the similar shooting time is small, a plurality of head portrait pictures with the similar shooting time are selected for comparison, and the accuracy of analysis and judgment of the head portrait pictures can be improved.
When a plurality of head portrait pictures with similar shooting time are selected, the shooting time is set according to the span of the shooting time of the pictures in the picture library, if the time span of the picture shooting is long, such as more than ten years, when the head portrait pictures with the similar shooting time are selected, the shooting time period selected by the pictures is the head portrait picture shot in the same year or shorter time; if the time span of the photo shooting in the photo library is short, for example, the photos are shot in one year, when head portrait pictures with similar shooting time are selected, head portrait pictures shot in the same month or shorter time are selected; if the shooting time of the photos in the photo library is close, for example, the photos are shot in a month or a week, when the head portrait photos with the close shooting time are selected, the head portrait photos should be shot in the same day or shorter time, and the closer the shooting time of the selected head portrait photos is, the smaller the change of the human facial features is, and the higher the recognition accuracy is.
When the similarity of the features of the head portrait pictures in the low-quality library and the feature set of the head portrait pictures in the face library is compared, a new face library is not needed, and only the comparison with the face library already established in the high-quality library is needed, and the judgment standard of whether the head portrait pictures in the low-quality library are the same person can be reduced, for example, the similarity of the feature set of the head portrait pictures and the feature set of the head portrait pictures in the face library is set to be more than or equal to 60%, the head portrait pictures are determined to be the same person, and when the similarity is less than 60%, the head portrait pictures are determined to be different persons.
When the photo library works, firstly, the information of the photos in the photo library is read, the head portrait of the face in the photos is identified, the position of the face in the photos is obtained, each face in all the photos is cut into an independent head portrait picture, and the corresponding relation between each head portrait picture and the original photo is established; then analyzing the quality of the head portrait pictures, classifying the head portrait pictures according to the quality, screening out high-quality head portrait pictures and low-quality head portrait pictures, moving the high-quality head portrait pictures into a high-quality library, and moving the low-quality head portrait pictures into a low-quality library; then, sequentially carrying out similarity comparison on the head portrait pictures in all the high-quality libraries, and judging whether the head portrait pictures are the same person or not according to the similarity to obtain one or more face libraries and a similarity library corresponding to the face libraries one by one; then sequentially and respectively selecting a plurality of head portrait pictures with similar shooting time in each face library, selecting a plurality of head portrait pictures with similar shooting time in a similar library corresponding to the face library, carrying out similarity comparison on two groups of head portrait pictures, merging the similar library and the face library if the head portrait pictures are judged to be the same person, and constructing the similar library into a new face library if the head portrait pictures are judged not to be the same person; and finally, sequentially comparing the characteristics of the head portrait pictures in the low-quality library with the feature collection of the head portrait pictures in the established face library, if the head portrait pictures are judged to be the same person, moving the head portrait pictures into the face library, deleting the low-quality head portrait pictures which are not moved into the face library after all the head portrait pictures in the low-quality library are compared with all the established face libraries, and finally classifying the pictures according to the face library.
According to the invention, the head portrait of the person in the photo is cut to form an independent head portrait picture, and the head portrait picture is associated with the original photo, so that the head portrait picture can be directly compared with the similarity, compared with the direct comparison of the photo, the analysis and comparison of irrelevant images in the photo and the repeated recognition of human faces are reduced, and the working efficiency and the accuracy can be improved; by setting the similarity library, head portrait pictures which are possibly the same as the human face library are placed in the similarity library corresponding to the human face library, and a plurality of head portrait pictures with similar shooting time in the human face library are selected to be compared with a plurality of head portrait pictures with similar shooting time in the similarity library corresponding to the head portrait pictures, so that the head portrait pictures with similar shooting time of the same person can be compared at the same time, and the accuracy of human face recognition is improved; the photos are divided into a high-quality library and a low-quality library, so that different photo qualities can be classified and compared, the judgment of the similarity of the head portrait pictures due to the fact that the faces of the low-quality head portrait pictures are not clear or the faces of the low-quality head portrait pictures are inclined is prevented from being influenced, the judgment standard of the similarity of the low-quality head portrait pictures is reduced, and the accuracy of portrait recognition is improved; when the head portrait picture is compared with the face library, the head portrait picture characteristics are compared with the collection of head portrait picture characteristics in the face library, so that the judgment accuracy can be improved.
According to an embodiment of the present invention, in the fourth step S24, after all the face libraries are compared with the corresponding similarity libraries, if there are a plurality of face libraries, the face libraries are continuously compared with each other in sequence for similarity comparison, and the face libraries determined as the same person are merged.
According to an embodiment of the present invention, in the second step S22, the quality of the head portrait picture is identified, mainly the definition and the face angle of the face in the head portrait picture are identified.
According to an embodiment of the present invention, in the third step S23, the first avatar picture in the high quality library is moved into the first face library, the features of the second avatar picture are compared with the union of the features of the avatar pictures in the first face library, if the same person is determined, the second avatar picture is moved into the similar library corresponding to the first face library, if the same person is determined, the avatar pictures are moved into the second face library, until the determination of the avatar pictures in the high quality library is completed, and one or more face libraries and similar libraries corresponding to the face libraries one to one are obtained.
According to an embodiment of the present invention, in the third step S23, the similarity between the feature of the head portrait picture in the high quality library and the feature of the head portrait picture in the face library is greater than or equal to 80%, and the head portrait pictures are considered as the same person; the similarity of the feature of the head portrait picture in the high-quality library and the feature union of the head portrait pictures in the face library is less than 80 percent and is more than or equal to 50 percent, and the head portrait pictures are considered to be possibly the same person; and if the similarity of the feature of the head portrait picture in the high-quality library and the feature of the head portrait picture in the face library is less than 50%, the head portrait pictures are not the same person.
According to an embodiment of the present invention, in the third step S23, if the similarity between the feature of the avatar picture in the high quality library and the union of the features of the avatar pictures in the face libraries is greater than or equal to 80%, the avatar picture is moved into the face library with the highest similarity, and the information of the three face libraries with the highest similarity is recorded.
According to an embodiment of the present invention, in the fourth step S24, when the face library is compared with the corresponding similar libraries, the head portrait pictures with the same number and the same shooting time are respectively selected from the face library and the corresponding similar libraries for comparison, when the number of the head portrait pictures in at least one of the face library and the similar libraries is less than three, all the head portrait pictures in the library with the smaller number of head portrait pictures are selected, and the head portrait pictures with the same number and the same shooting time are selected from the other library for comparison; when the number of the head portrait pictures in the face library and the similar library is more than three, selecting 3-10 head portrait pictures with similar shooting time and the same number from the two libraries for comparison; and if the similarity of the head portrait picture characteristics in the face library and the corresponding similarity library is more than or equal to 80%, determining the head portrait picture characteristics as the same person, otherwise, determining the head portrait picture characteristics as not the same person.
According to an embodiment of the present invention, in the fourth step S24, when two face libraries are compared, head portrait pictures with similar shooting time and same number are respectively selected from the two face libraries for comparison, when the number of head portrait pictures in at least one of the two face libraries is less than three, all head portrait pictures in the face library with less head portrait pictures are selected, and head portrait pictures with similar shooting time and same number are selected from the other face library for comparison; and when the number of the head portrait pictures in the two face libraries is more than three, selecting 3-10 head portrait pictures with similar shooting time and the same number from the two face libraries for comparison.
According to an embodiment of the present invention, in the fourth step S24, when the similarity between the head portrait image features in the two face libraries is greater than or equal to 80%, the head portrait image features are determined as the same person, otherwise, the head portrait image features are determined as not the same person.
According to an embodiment of the present invention, in the fifth step S25, the similarity between the feature of the head portrait picture in the low quality library and the feature of the head portrait picture in the face library is greater than or equal to 70%, and the same person is determined, and if the similarity is less than 70%, the same person is determined.
Example 1: putting 1000 photos into a photo library, setting a 0 point of a shooting date of the photo with the same shooting place and the earliest shooting time in the photo library as a starting time of a time axis, setting a 24 point of the shooting date of the photo with the same shooting place and the latest shooting time as an ending time of the time axis, and setting a time span set by the time axis to be not more than 3 days; setting a photo with shooting time within the time axis or within one day close to the time axis and including the shooting-free place of the first person for marking to obtain a first marked photo; setting pictures shot by the shooting equipment in a picture library and shot within one day with the shooting time within or close to the time axis for marking to obtain second marked pictures; and when other people are also included in the photos shot at the same shooting position and/or the first marked photo and/or the second marked photo, selecting the other people as the first people, repeating the second step and the third step until all people in the photos shot at the same shooting position and/or the first marked photo and/or the second marked photo are completely selected, and adding shooting positions in the first marked photo and the second marked photo according to the shooting positions of the photos with the shooting positions in the time axis.
And (4) counting results: among the 1000 photographs, there are 426 photographs with shooting location information and 574 photographs without shooting location information;
570 photos of the shooting place are added to the photos of the non-shooting place, and the adding rate is 99.3%;
wherein 4 pictures of the shooting positions are not added to the pictures of the non-shooting positions, and the non-adding rate is 0.7 percent;
adding 569 correct pictures to all the pictures of the shooting place, wherein the accuracy rate is 99.82%;
wherein, adding 1 wrong photo in all photos of the shooting location, the error rate is 0.18%.
Example 2: putting 1000 photos into a photo library, setting a starting time of a time axis as a 0 point of a shooting date of a photo with the same shooting place and the earliest shooting time in the photo library, setting a finishing time of the time axis as a 24 point of a shooting date of a photo with the same shooting place and the latest shooting time, and setting a time span of the time axis to be not more than 3 days; marking photos of non-shooting places including the first person, the shooting time of which is within the time axis or within one day close to the time axis, so as to obtain first marked photos; when other people are set in the pictures shot at the same shooting place and/or the first marked picture, selecting the other people as the first people, and expanding the first marked picture until all the people in the pictures shot at the same shooting place and/or in the first marked picture are completely selected; and marking the photos which are shot by the shooting equipment in the photo library and have the shooting time in the time axis or in one day close to the time axis to obtain a second marked photo, and adding the shooting places in the first marked photo and the second marked photo according to the shooting places with the photos of the shooting places in the time axis.
And (4) counting results: in 1000 photographs, 355 photographs with shooting location information exist, and 645 photographs without shooting location information exist;
wherein 637 photos of shooting places are added to the photos of the non-shooting places, and the adding rate is 98.75 percent;
8 pictures of shooting places are not added to the pictures of the non-shooting places, and the non-adding rate is 1.25%;
wherein, adding 635 correct pictures to all the pictures of the shooting place, with 99.68% accuracy;
wherein, in all the photos of the shooting location, 2 wrong photos are added to the shooting location, and the error rate is 0.32%.
According to the embodiment, the shooting places can be quickly and accurately added to a large number of photos without shooting places in the photo library, the problem that the photos cannot be classified according to the shooting places because the photo library contains a large number of photos without shooting places can be solved, and time for viewing the photos by people is saved greatly.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and that those skilled in the art will be able to design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim.
Claims (9)
1. A method of adding a shooting location to a photograph, comprising the steps of:
first step (S11): identifying the shooting time, the shooting place and the shooting equipment of the photos in the photo library, and determining a time axis according to the shooting time of the photos shot at the same shooting place;
second step (S12): selecting a first person in the photos shot at the same shooting place, and marking the photos, which are shot at the time axis or close to the time axis and contain the shooting place of the first person, to obtain a first marked photo;
third step (S13): identifying shooting equipment of the first marked photo, and marking photos which are shot by the shooting equipment in the photo library and have shooting time in the time axis or close to the time axis to obtain a second marked photo;
fourth step (S14): and adding the shooting positions in the first marked picture and the second marked picture according to the shooting positions of the pictures with the shooting positions in the time axis.
2. The method according to claim 1, the second step (S12) further comprising,
when other persons are included in the photo taken at the same shooting location and/or the first marked photo, the other persons are selected, the second step (S12) is repeated, and the first marked photo is expanded until all the persons in the photo taken at the same shooting location and/or the first marked photo are selected.
3. The method according to claim 1, the third step (S13) further comprising,
when other persons are included in the photos taken at the same place and/or in the first tagged photo and/or the second tagged photo, the other persons are selected, and the second step S12 and the third step S13 are repeated until all the persons in the photos taken at the same place and/or in the first tagged photo and/or the second tagged photo are selected.
4. The method according to claim 1, wherein in the first step (S11), 0 point of a shooting date of a photograph having the same shooting location and the earliest shooting time in a photograph library is taken as a start time of a time axis; the 24 points of the shooting date of the photograph having the same shooting location and the latest shooting time are taken as the end time of the time axis.
5. The method of claim 1, wherein the timeline is configured to span no more than 3 days.
6. The method of adding a shooting location to a photo of claim 1, wherein in the fourth step (S14), when there are a plurality of people in the photo with shooting location information shot in the time axis, the person who shot the most photos is selected as the first person.
7. The method of claim 1, wherein the camera is a single lens, a mini-lens, a card machine, a mobile phone, or the like.
8. The method of claim 1, wherein in the second step (S12), a first tagged picture is obtained by tagging a picture of a non-shooting location containing the first person, the shooting time of which is within the time axis or within a day close to the time axis.
9. The method according to claim 1, wherein in the third step (S13) photos of the photo library taken by the capturing device and having a capturing time within or close to the time axis within one day are tagged resulting in a second tagged photo.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011624068.9A CN112732966A (en) | 2020-12-31 | 2020-12-31 | Method for adding shooting place to photo |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011624068.9A CN112732966A (en) | 2020-12-31 | 2020-12-31 | Method for adding shooting place to photo |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112732966A true CN112732966A (en) | 2021-04-30 |
Family
ID=75609658
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011624068.9A Pending CN112732966A (en) | 2020-12-31 | 2020-12-31 | Method for adding shooting place to photo |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112732966A (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102917173A (en) * | 2012-10-31 | 2013-02-06 | 广东欧珀移动通信有限公司 | Method and device for automatically adding photographing location during photographing and terminal |
| CN104298694A (en) * | 2013-07-19 | 2015-01-21 | 深圳市康睿祥通讯有限公司 | Picture message adding method and device and mobile terminal |
| CN104933146A (en) * | 2015-06-19 | 2015-09-23 | 深圳天珑无线科技有限公司 | Photo management method, photo management device and mobile terminal |
| CN105488074A (en) * | 2014-09-28 | 2016-04-13 | 小米科技有限责任公司 | Photo clustering method and device |
| CN105512328A (en) * | 2015-12-23 | 2016-04-20 | 北京奇虎科技有限公司 | Method, system and device for realizing uploading of album photos |
| CN109597908A (en) * | 2018-12-14 | 2019-04-09 | 深圳壹账通智能科技有限公司 | Photo searching method, device, equipment and storage medium based on recognition of face |
| CN110929063A (en) * | 2019-11-21 | 2020-03-27 | 深圳传音控股股份有限公司 | Album generating method, terminal device and computer readable storage medium |
| CN111611428A (en) * | 2020-05-25 | 2020-09-01 | 重庆大学 | Photographing processing method and system based on geographic historical path |
-
2020
- 2020-12-31 CN CN202011624068.9A patent/CN112732966A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102917173A (en) * | 2012-10-31 | 2013-02-06 | 广东欧珀移动通信有限公司 | Method and device for automatically adding photographing location during photographing and terminal |
| CN104298694A (en) * | 2013-07-19 | 2015-01-21 | 深圳市康睿祥通讯有限公司 | Picture message adding method and device and mobile terminal |
| CN105488074A (en) * | 2014-09-28 | 2016-04-13 | 小米科技有限责任公司 | Photo clustering method and device |
| CN104933146A (en) * | 2015-06-19 | 2015-09-23 | 深圳天珑无线科技有限公司 | Photo management method, photo management device and mobile terminal |
| CN105512328A (en) * | 2015-12-23 | 2016-04-20 | 北京奇虎科技有限公司 | Method, system and device for realizing uploading of album photos |
| CN109597908A (en) * | 2018-12-14 | 2019-04-09 | 深圳壹账通智能科技有限公司 | Photo searching method, device, equipment and storage medium based on recognition of face |
| CN110929063A (en) * | 2019-11-21 | 2020-03-27 | 深圳传音控股股份有限公司 | Album generating method, terminal device and computer readable storage medium |
| CN111611428A (en) * | 2020-05-25 | 2020-09-01 | 重庆大学 | Photographing processing method and system based on geographic historical path |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102033958B (en) | Photo sort management system and method | |
| US7920760B2 (en) | Image organizing device and method, and computer-readable recording medium storing image organizing program | |
| US8917943B2 (en) | Determining image-based product from digital image collection | |
| US20130050747A1 (en) | Automated photo-product specification method | |
| US20180025215A1 (en) | Anonymous live image search | |
| JP2020140577A (en) | Device, method, and program | |
| US20080205771A1 (en) | Classifying complete and incomplete date-time information | |
| JP2020140576A (en) | Device, method, and program | |
| JP2020140567A (en) | Device, method, and program | |
| US11470209B2 (en) | Image processing apparatus, method of controlling image processing apparatus, and storage medium for editing an album | |
| JPWO2011001587A1 (en) | Content classification device, content classification method, and content classification program | |
| US11936821B2 (en) | Image processing apparatus, method of controlling image processing apparatus, and storage medium with determination as to whether image is already contained in album data for first subject and album data for second subject and specifying image as change target | |
| US8831360B2 (en) | Making image-based product from digital image collection | |
| US20220262057A1 (en) | Storage medium, image processing apparatus, and image processing method | |
| JP2020140568A (en) | Device, method, and program | |
| TWI556640B (en) | Media file management method and system, and computer-readable medium | |
| US11915351B2 (en) | Image processing apparatus, image processing method, and storage medium | |
| US20220262055A1 (en) | Image processing apparatus, image processing method, and storage medium | |
| JP2022123211A (en) | Program, image processing apparatus, and image processing method | |
| CN112765384B (en) | Photo classification method | |
| CN112732966A (en) | Method for adding shooting place to photo | |
| CN112733664B (en) | Photo classification method | |
| CN112712041B (en) | Photo classification method | |
| JP2005275847A (en) | Image storage method and image storage device | |
| CN110214337A (en) | Image presentation system, image presentation method, program and recording medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |