A kind of method and device of retrieving imageTechnical field
This application involves monitoring field, in particular to a kind of method and device of retrieving image.
Background technique
To being photo of the manual type by suspect mostly in the tracing process of suspect in traditional security systemSearch result in face retrieval system is gone, search result in human body searching system is gone by the photo of suspect, or pass through suspicionIt doubts people's photo of riding and goes search result in vehicle retrieval system.
In this case it needs to carry out multi-pass operation respectively to go in different systems to be retrieved, work effect in this caseRate is low and retrieval in some system is easy to cause to be missed.
Summary of the invention
In order to solve the problems, such as that working efficiency is low in the related technology and retrieval is missed, the embodiment of the present application provides oneThe method and device of kind retrieving image.The technical solution is as follows:
In a first aspect, providing a kind of method of retrieving image, which comprises
Object set is obtained, the object set includes at least one target retrieval object, and target retrieval object is to be checkedSubject image either characteristics of image in rope picture；
According to the object type of each target retrieval object in the object set, each target retrieval pair is determinedAs corresponding picture retrieval system；
According to each target retrieval object, respectively in the corresponding picture retrieval system of each target retrieval objectIt is middle to retrieve the corresponding at least picture of each target retrieval object.
Optionally, the target retrieval object is the subject image in the picture to be retrieved,
It is described according to each target retrieval object, respectively in the corresponding picture retrieval of each target retrieval objectBefore retrieving the corresponding at least picture of each target retrieval object in system, further includes:
According to the object type of each target retrieval object, it is corresponding that each target retrieval object is obtained respectivelyCharacteristics of image；
It is described according to each target retrieval object, respectively in the corresponding picture retrieval of each target retrieval objectThe corresponding at least picture of each target retrieval object is retrieved in system, comprising:
According to each target retrieval object and the corresponding characteristics of image of each target retrieval object, respectively in instituteState that each target retrieval object is retrieved in the corresponding picture retrieval system of each target retrieval object is at least one correspondingPicture.
Optionally, the acquisition object set, comprising:
Each retrieval object for including in picture to be retrieved is identified, object is retrieved in the picture to be retrieved and is located at its correspondenceTarget frame in, and the attribute of one target frame is set as pushing up when there are a target frame is located in another target frameLayer frame, the attribute of another target frame are set as bottom frame；
Obtain the weight of each retrieval object in the picture to be retrieved；
Weight is extracted in the picture to be retrieved to be greater than in the corresponding target frame of retrieval object of preset weights threshold valueImage is as target retrieval object.
Optionally, described according to each target retrieval object, it is corresponding in each target retrieval object respectivelyAfter retrieving the corresponding at least picture of each target retrieval object in picture retrieval system, further includes:
Calculate the first similarity between the every picture and the object set in picture set, the picture set packetInclude the corresponding at least picture of each target retrieval object；
Every picture is shown according to the first similarity of every picture.
Optionally, the first similarity between the every picture and the object set calculated in picture set, packetIt includes:
Determine n target retrieval object in the object set that Target Photo includes, the Target Photo is describedAny picture in picture set, n are the integer greater than 0；
According to the weight of each target retrieval object in the n target retrieval object and the Target Photo with it is describedThe second similarity between each target retrieval object calculates first similar between the Target Photo and the object setDegree.
Optionally, before the acquisition object set, further includes:
Select the maximum at least picture of the first similarity as figure to be retrieved from the picture set that the last time obtainsPiece.
Second aspect, provides a kind of device of retrieving image, and described device includes:
First obtains module, and for obtaining object set, the object set includes at least one target retrieval object, meshMark retrieval object is the subject image either characteristics of image in picture to be retrieved；
Determining module, for the object type according to each target retrieval object in the object set, determine described inThe corresponding picture retrieval system of each target retrieval object；
Retrieval module, for being corresponded in each target retrieval object respectively according to each target retrieval objectPicture retrieval system in retrieve the corresponding at least picture of each target retrieval object.
Optionally, the target retrieval object is the subject image in the picture to be retrieved, described device further include:
Second obtains module, for the object type according to each target retrieval object, obtains respectively described eachThe corresponding characteristics of image of target retrieval object；
The retrieval module, for corresponding according to each target retrieval object and each target retrieval objectCharacteristics of image retrieves each target retrieval pair in the corresponding picture retrieval system of each target retrieval object respectivelyAs a corresponding at least picture.
Optionally, described first module is obtained, comprising:
Mark unit is retrieved in the picture to be retrieved for identifying each retrieval object for including in picture to be retrievedObject is located in its corresponding target frame, and one target frame when being located in another target frame there are a target frameAttribute be set as top-level boxes, the attribute of another target frame is set as bottom frame；
Acquiring unit, for obtaining the weight of each retrieval object in the picture to be retrieved；
Extraction unit, the retrieval object that preset weights threshold value is greater than for extracting weight in the picture to be retrieved correspond toTarget frame in image as target retrieval object.
Optionally, described device further include:
Computing module, for calculating the first similarity between every picture and the object set in picture set,The picture set includes the corresponding at least picture of each target retrieval object；
Display module, for showing every picture according to the first similarity of every picture.
Optionally, the computing module includes:
Determination unit, for determining n target retrieval object in the object set that Target Photo includes, the meshPiece of marking on a map is any picture in the picture set, and n is the integer greater than 0；
Computing unit, for according to the weight of each target retrieval object in the n target retrieval object and describedThe second similarity between Target Photo and each target retrieval object, calculates the Target Photo and the object setBetween the first similarity.
Optionally, described device further include:
Selecting module, for selecting the maximum at least picture of the first similarity from the picture set that the last time obtainsAs picture to be retrieved.
Technical solution provided by the embodiments of the present application can include the following benefits:
Due to the object type according to each target retrieval object in object set, each target retrieval object pair is determinedThe picture retrieval system answered；According to each target retrieval object, respectively in the corresponding picture retrieval system of each target retrieval objectThe corresponding at least picture of each target retrieval object is retrieved in system, such terminal can be automatically using every in object setA target retrieval object retrieving image in picture retrieval system, so as to avoid the problem that manual retrieval omits, andRecall precision can be improved.
It should be understood that above general description and following detailed description be only it is exemplary and explanatory, notThe application can be limited.
Detailed description of the invention
The drawings herein are incorporated into the specification and forms part of this specification, and shows the implementation for meeting the applicationExample, and together with specification it is used to explain the principle of the application.
Fig. 1 is a kind of network architecture schematic diagram that the embodiment of the present application 1 provides；
Fig. 2 is a kind of method flow diagram for retrieving image that the embodiment of the present application 2 provides；
Fig. 3 is a kind of method flow diagram for retrieving image that the embodiment of the present application 3 provides；
Fig. 4 is a kind of the first schematic diagram of picture to be retrieved that the embodiment of the present application 3 provides；
Fig. 5 is a kind of the second schematic diagram of picture to be retrieved that the embodiment of the present application 3 provides；
Fig. 6 is a kind of third schematic diagram for picture to be retrieved that the embodiment of the present application 3 provides；
Fig. 7 is a kind of method flow diagram for retrieving image that the embodiment of the present application 4 provides；
Fig. 8 is a kind of apparatus structure schematic diagram for retrieving image that the embodiment of the present application 5 provides；
Fig. 9 is a kind of apparatus structure schematic diagram for retrieving image that the embodiment of the present application 6 provides.
Through the above attached drawings, it has been shown that the specific embodiment of the application will be hereinafter described in more detail.These attached drawingsIt is not intended to limit the range of the application design in any manner with verbal description, but is by referring to specific embodimentsThose skilled in the art illustrate the concept of the application.
Specific embodiment
Example embodiments are described in detail here, and the example is illustrated in the accompanying drawings.Following description is related toWhen attached drawing, unless otherwise indicated, the same numbers in different drawings indicate the same or similar elements.Following exemplary embodimentDescribed in embodiment do not represent all embodiments consistent with the application.On the contrary, they be only with it is such as appendedThe example of the consistent device and method of some aspects be described in detail in claims, the application.
Embodiment 1
It is that the embodiment of the present application provides a kind of network architecture referring to Fig. 1, Fig. 1, comprising:
Terminal, picture retrieval system and picture pick-up device, picture retrieval system may include face retrieval system, human body inspectionCable system and vehicle retrieval system.
Terminal can be connected with each picture retrieval system, i.e., terminal can be with face retrieval system, human body searching systemIt is connected with vehicle retrieval system.
Picture pick-up device can be connected with each picture retrieval system, i.e., picture pick-up device can be with face retrieval system, human bodySearching system is connected with vehicle retrieval system.
Picture pick-up device may be arranged at road, the places such as street, community, station, airport, mansion and/or market, for shootingPicture detects the picture, if detecting in the picture to include facial image, which is stored in face retrievalIn system；If detecting in the picture to include human body image, which is stored in human body searching system；If detectionInclude out vehicle image in the picture, then the picture is stored in vehicle retrieval system.
Picture pick-up device can also obtain the pictorial information of the picture, which may include the camera site of the pictureAnd/or the information such as shooting time, if in the picture include facial image, the pictorial information can also include the face whetherIt wears glasses, whether smile and/or gender etc.；If include human body image in the picture, which can also include beingIt is no cycle, whether the information such as knapsack and/or gender；If including vehicle image in the picture, which can also includeLicense board information, vehicle color and/or vehicle etc..And picture pick-up device is in picture retrieval system when storing the picture, also in the figureThe pictorial information of the picture is stored in piece searching system.
Terminal, for obtaining object set, object set includes at least one target retrieval object, and target retrieval object isSubject image either characteristics of image in picture to be retrieved；According to the object class of each target retrieval object in object setType determines the corresponding picture retrieval system of each target retrieval object；According to each target retrieval object, respectively in each targetThe corresponding at least picture of each target retrieval object is retrieved in the corresponding picture retrieval system of retrieval object.
Target retrieval object can be facial image, human body image or vehicle image etc., alternatively, target retrieval object is behavedThe characteristics of image of face image, the characteristics of image of human body image or characteristics of image of vehicle image etc., so target retrieval objectObject type can be face, human body and vehicle.The corresponding picture retrieval system of face can be face searching system, human body pairThe picture retrieval system answered can be people's physical examination cable system and the corresponding picture retrieval system of vehicle can be vehicle retrieval systemSystem.
Wherein, above-mentioned terminal can be the equipment such as computer, laptop, tablet computer or mobile phone.Any of the above-describedPicture retrieval system can be a server or the server cluster etc. being made of multiple servers.
Embodiment 2
Referring to fig. 2, the embodiment of the present application provides a kind of method of retrieving image, and this method can be applied to embodiment 1The network architecture of offer, and the executing subject of this method can be the terminal in the network architecture.This method comprises:
Step 201: obtaining object set, which includes at least one target retrieval object, target retrieval objectIt is the subject image either characteristics of image in picture to be retrieved.
Step 202: according to the object type of each target retrieval object in the object set, determining each target retrievalThe corresponding picture retrieval system of object.
Step 203: according to each target retrieval object, respectively in the corresponding picture retrieval system of each target retrieval objectIt is middle to retrieve the corresponding at least picture of each target retrieval object.
In the embodiment of the present application, due to the object type according to each target retrieval object in the object set, reallyThe corresponding picture retrieval system of each target retrieval object is determined, according to each target retrieval object, respectively in each target retrievalThe corresponding at least picture of each target retrieval object is retrieved in the corresponding picture retrieval system of object, such terminal can be certainlyDynamic each target retrieval object retrieving image in picture retrieval system using in object set, so as to avoid manually examiningThe problem of rope is omitted, and recall precision can be improved.
Embodiment 3
Referring to Fig. 3, the embodiment of the present application provides a kind of method of retrieving image, and this method can be applied to embodiment 1The network architecture of offer, and the executing subject of this method can be the terminal in the network architecture.This method comprises:
Step 301: each retrieval object for including in terminal iidentification picture to be retrieved.
In picture to be retrieved after mark, each object of retrieving is located in its corresponding target frame, and when there are someThe attribute of one target frame is set as top-level boxes when target frame is located in another target frame, and the attribute of another target frame is setFor bottom frame.
Picture to be retrieved can be user and input to terminal.Picture to be detected can be either passing through for user's shootingWhat the picture pick-up device in the network architecture shown in embodiment 1 was shot etc..Alternatively, picture to be retrieved can be last retrievalTo picture set in the maximum at least picture of the first similarity.
The present embodiment can be applied to policing system, and police is in the picture for taking the vehicle including suspect or suspectWhen, it may be necessary to the picture including the suspect or vehicle is retrieved from picture retrieval system, it at this time can be defeated by the pictureEnter into terminal, terminal receives the picture and as picture to be retrieved.
This step can be realized by following two steps, be respectively as follows:
3011: terminal identifies each retrieval object that picture to be detected includes by modeling to picture to be detected.
It include modeling in terminal, picture to be detected can be input to the modeling by terminal, which knowsIt Chu not picture to be detected each retrieval object for including.
Retrieval object in picture to be detected is the subject image in picture to be retrieved, such as can be facial image, peopleBody image or vehicle image etc..
For example, with reference to picture to be retrieved shown in Fig. 4, terminal can be by modeling from figure to be retrieved shown in Fig. 4The retrieval object identified in piece includes human body image, facial image, vehicle image and animal painting.
3012: terminal adds target frame on each retrieval object in picture to be retrieved, the mesh added on retrieval objectIt marks frame and is used for identification retrieval object.
The target frame of retrieval object can be the minimum frame for including the retrieval object, which can be rectangle frame or other are moreThe frame etc. of side shape.For example, target frame 1 is added on this retrieval object of facial image in picture to be retrieved shown in Fig. 4,Target frame 2 is added on this retrieval object of human body image, adds target frame 3 on this retrieval object of vehicle image, andTarget frame 4 is added on this retrieval object of animal painting, obtains picture to be retrieved shown in fig. 5.
Optionally, when some target frame is located within another target frame, the attribute of a target frame is arrangedFor top-level boxes, bottom frame is set by the attribute of another target frame.Terminal can be avoided to omit in subsequent processing in this way shouldOne target frame.
Optionally, when partly overlapping between certain two target frame, the attribute of one of target frame is set as top-level boxes,The attribute of another target frame is set as bottom frame.Top-level boxes can also be set by the target frame where focus, by anotherThe attribute of target frame is set as bottom frame.Such as it can be using the target frame that user chooses as the target frame where focus.It can be withAttribute setting is carried out according to the size of remainder of the target frame in addition to lap, such as by the small target frame of areaAttribute is set as top-level boxes, and the attribute of the big target frame of area is set as bottom frame.The application is defined not to this.
Step 302: terminal obtains the weight of each retrieval object in picture to be retrieved.
The weight for retrieving object is bigger, indicates that user uses the hope of retrieval object retrieving image in picture retrieval systemPrestige is bigger, conversely, the weight of retrieval object is smaller, indicates that user uses retrieval object retrieving image in picture retrieval systemHope it is smaller.
This step can be realized by following two steps, be respectively as follows:
3021: terminal shows picture to be retrieved, and each retrieval object in picture to be retrieved is shown in picture to be retrievedCorresponding weight input interface.
After terminal shows picture to be retrieved, the input-output equipment that terminal can be used in user is corresponding in each retrieval objectThe weight of each retrieval object is inputted in weight input interface respectively.
Wherein, for each retrieval object, if user uses retrieval object retrieving image in picture retrieval systemHope it is bigger, then the weight that user can input in the weight input interface of the retrieval object is bigger；If user usesRetrieval object hope of retrieving image in picture retrieval system is smaller, then user can input in the weight of the retrieval objectThe weight inputted in interface is smaller.
For example, with reference to Fig. 6, the corresponding weight input interface of each retrieval object, the weight are shown in picture to be retrievedInput interface can be input frame etc..Then user can input respectively in the corresponding weight input interface of each retrieval objectThe weight of each retrieval object.Such as in Fig. 6, the weight that facial image 1 is inputted in the weight input interface of facial image 1 is100, the weight that human body image 2 is inputted in the weight input interface of human body image 2 is 80, is inputted in the weight of vehicle image 3The weight that vehicle image 2 is inputted in interface is 90, and the weight that animal painting 4 is inputted in the weight input frame of animal painting 4 is0。
3022: terminal reads the corresponding power of each retrieval object respectively from the corresponding weight input interface of each retrieval objectValue.
For example, the weight 100 of facial image 1 is read in the weight input interface of facial image 1, in human body in Fig. 6The weight 80 that human body image 2 is read in the weight input interface of image 2, reads vehicle in the weight input interface of vehicle image 3The weight 90 of image 2, and read in the weight input frame of animal painting 4 weight 0 of animal painting 4.
Optionally, terminal can be with display condition input interface.User can in the condition entry interface input pictureFeature.
Terminal can show a condition entry interface, can also show that the corresponding condition entry of each object type connectsMouthful.
If terminal shows a condition entry interface, user can be inputted in the condition entry interface to every kind of objectThe characteristics of image of type.Such as user can input in the condition entry interface to the characteristics of image of facial image, to human bodyThe characteristics of image of image and/or characteristics of image to vehicle image.
If terminal shows the corresponding condition entry interface of each object type, user can be in each object type pairThe characteristics of image of each object type is inputted in the condition entry interface answered respectively.Such as user can be corresponding in facial imageThe characteristics of image of condition entry interface input facial image inputs human body image in the corresponding condition entry interface of human body imageCharacteristics of image and/or in the condition entry interface of vehicle image input vehicle image characteristics of image.
Wherein, whether the characteristics of image of facial image may include whether to smile, wear glasses and/or gender etc..Human figureThe characteristics of image of picture may include whether cycle, whether knapsack and/or gender etc..The characteristics of image of vehicle image may include vehicleBoard information, vehicle color and/or vehicle etc..
Step 303: terminal extracts the corresponding mesh of retrieval object that weight is greater than preset weights threshold value from picture to be retrievedThe image in frame is marked as target retrieval object, and forms object set.
Preset weights threshold value can be 10 or 20 equal numerical value.
Terminal can protect the image of extraction from the image extracted in the corresponding target frame of retrieval object in picture to be retrievedNew picture is saved as, which is a target retrieval object, and the target retrieval object that object set includes is that user needsThe retrieval object to be retrieved.
For example, it is assumed that preset weights threshold value 10, then terminal selection weight is more than 10 facial image 1, human body image 2 and vehicleImage 3 is used as target retrieval object, saves as a new picture from the facial image 1 in target frame 1 is extracted in picture to be retrieved1, target retrieval object 1 is obtained, the human body image 2 in target frame 2 is extracted and saves as new picture 2, obtain target retrieval object2, and extract the vehicle image 3 in target frame 3 and save as new picture 3, target retrieval object 3 is obtained, by target retrieval pairAs 1,2 and 3 composition object sets.
Step 304: terminal determines that each target is examined according to the object type of each target retrieval object in object setThe corresponding picture retrieval system of rope object, according to each target retrieval object respectively in the corresponding picture of each target retrieval objectThe corresponding at least picture of each target retrieval object is retrieved in searching system, and forms picture set.
For each target retrieval object in object set, terminal first determines the object type of the target retrieval object,It determines the corresponding picture retrieval system of the object type, sends retrieval request message, the retrieval request to the picture retrieval systemIt includes the target retrieval object that message, which carries,.The picture retrieval system receives the retrieval request message, extracts the retrieval requestThen the target retrieval object that message carries retrieves corresponding at least one figure of the target retrieval object from its picture libraryPiece.The picture retrieval system can also obtain between every picture and the target retrieval object in an at least pictureTwo similarities.
Wherein, which can be similar more than presetting to the second similarity between the target retrieval objectSpend the picture of threshold value.
The picture retrieval system can also obtain the pictorial information of every picture in an at least picture.The figure of picturePiece information may include camera site and/or shooting time of the picture etc.；If in the picture including facial image, the pictureInformation can also include whether the facial image smile, whether wear glasses and/or the information such as gender；If in the picture including peopleBody image, the pictorial information can also include human body image whether cycle, whether the information such as knapsack and/or gender；If shouldIt include vehicle image in picture, which can also include the information such as license board information, vehicle color and/or vehicle.
The picture retrieval system can will retrieve every picture in an at least picture, every picture and target and examineThe second similarity and pictorial information between rope object are sent to terminal.
Terminal can receive between every picture, every picture and target retrieval object that each picture retrieval system is sentSecond similarity and pictorial information and composition picture set.
For example, terminal can send retrieval request message to picture retrieval system for the target retrieval object 1 in Fig. 6,It includes target retrieval object 1 that the retrieval request message, which carries,.Picture retrieval system is according to target retrieval object 1, from its picture libraryIn retrieve the corresponding M1 picture of facial image 1 in target retrieval object 1, and obtain every picture in the M1 pictureThe pictorial information of the second similarity and every picture between target retrieval object 1 sends the M1 picture, every to terminalThe second similarity and pictorial information between picture and target retrieval object 1.
For the target retrieval object 2 in Fig. 6, terminal can send retrieval request message, the inspection to picture retrieval systemIt includes target retrieval object 2 that rope request message, which carries,.Picture retrieval system is examined from its picture library according to target retrieval object 2Rope goes out the corresponding M2 picture of human body image 2 in target retrieval object 2, and obtains every picture and mesh in the M2 pictureThe pictorial information of the second similarity and every picture between mark retrieval object 2, sends the M2 picture, every picture to terminalThe second similarity and pictorial information between target retrieval object 2.
For the target retrieval object 3 in Fig. 6, terminal can send retrieval request message, the inspection to picture retrieval systemIt includes target retrieval object 3 that rope request message, which carries,.Picture retrieval system is examined from its picture library according to target retrieval object 3Rope goes out the corresponding M3 picture of vehicle image 3 in target retrieval object 3, and obtains every picture and mesh in the M3 pictureThe pictorial information of the second similarity and every picture between mark retrieval object 3, sends the M3 picture, every picture to terminalThe second similarity and pictorial information between target retrieval object 3.
Terminal receives the second similarity between every picture and target retrieval object 1 in the M1 picture, the M1The second similarity and picture between the every picture and target retrieval object 2 in pictorial information, the M2 picture, the M2The second similarity and picture letter between every picture and target retrieval object 3 in information, the M3 picture and the M3Breath, and the contents such as the M1 picture, the M2 picture and the M3 picture are formed into picture set.
Optionally, in this step, according to the object type of each target retrieval object, each target retrieval is obtained respectivelyThe corresponding characteristics of image of object；According to each target retrieval object and the corresponding characteristics of image of each target retrieval object, respectivelyThe corresponding at least picture of each target retrieval object is retrieved in the corresponding picture retrieval system of each target retrieval object.
When realizing, for any one target retrieval object, terminal is by the image of the object type of the target retrieval objectFeature is corresponding from the object type then according to the target retrieval object as the corresponding characteristics of image of target retrieval objectPicture retrieval system in retrieve the pictorial information of the target retrieval object corresponding at least a picture and every picture, from thisPictorial information is selected to meet the picture of the corresponding characteristics of image of target retrieval object in an at least picture.Alternatively, for appointingOne target retrieval object, terminal is using the characteristics of image of the object type of the target retrieval object as the target retrieval object pairThe characteristics of image answered, then according to target retrieval object characteristics of image corresponding with the target retrieval object, from the object classThe corresponding at least picture of a picture and every picture letter of the target retrieval object is retrieved in the corresponding picture retrieval system of typeIt ceases, the pictorial information of every picture in an at least picture is all satisfied the corresponding characteristics of image of target retrieval object.
For example, terminal is after receiving the M1 picture, the characteristics of image of facial image is right as target retrieval object 1The characteristics of image answered selects pictorial information from the M1 picture according to the pictorial information of the characteristics of image and the M1 pictureMeet the M11 picture of the characteristics of image；After receiving the M2 picture, examined the characteristics of image of human body image as targetThe corresponding characteristics of image of rope object 2 is selected from the M2 picture according to the pictorial information of the characteristics of image and the M2 picturePictorial information meets the M12 picture of the characteristics of image；And after receiving the M3 picture, by the image of vehicle imageFeature is as the corresponding characteristics of image of target retrieval object 3, according to the pictorial information of the characteristics of image and the M3 picture, from thisPictorial information is selected to meet the M13 picture of the characteristics of image in M3 picture.Then, terminal by the M11 picture of selection,The M12 picture and the M13 picture form picture set.
Optionally, terminal, can be from this at least after receiving the corresponding at least picture of some target retrieval objectThe maximum picture of the second similarity is selected in one picture, according to the picture of selection, from the object type of the target retrieval objectThe corresponding at least picture of picture of the selection is retrieved in corresponding picture retrieval system, it can also be by an at least pictureIt is added to picture set；Alternatively, selection meets the picture of the characteristics of image of the object type from this at least a picture, then willThe picture of selection is added in picture set.
For example, terminal can select the second similarity maximum one after receiving the M1 picture from the M1 picturePicture retrieves the corresponding N1 picture of the picture according to the picture in face retrieval system, can be by the N1 figuresPiece is added to picture set；Alternatively, selection meets the N11 picture of the characteristics of image of facial image from the N1 picture, it willThe N11 picture is added in the picture set.
Terminal can select the maximum figure of the second similarity after receiving the M2 picture from the M2 picturePiece retrieves the corresponding N2 picture of the picture according to the picture in human body searching system, which can be addedIt is added to picture set；Alternatively, selection meets the N12 picture of the characteristics of image of human body image from the N2 picture, by the N12Picture is added in the picture set.
Terminal can select the maximum figure of the second similarity after receiving the M3 picture from the M3 picturePiece retrieves the corresponding N3 picture of the picture according to the picture in vehicle retrieval system, which can be addedIt is added to picture set；Alternatively, selection meets the N13 picture of the characteristics of image of vehicle image from the N3 picture, by the N13Picture is added in the picture set.
Step 305: terminal calculates the first similarity between every picture and object set in picture set.
This step can be with are as follows: for any picture in the picture set, the picture is referred to as mesh for ease of descriptionIt marks on a map piece, determines that Target Photo includes n target retrieval object in object set, according to each target retrieval in the nThe second similar value between the weight and Target Photo of object and each target retrieval object calculates Target Photo and object setBetween the first similarity.
For example, it is assumed that including n target retrieval object in Target Photo, the weight of the n target retrieval object is respectivelyX1, X2 ... Xn, the Target Photo the second phase between each target retrieval object in the n target retrieval object respectivelyIt is respectively Y1, Y2 ... Yn like degree；So the first similarity=X1*Y1+X2*Y2 between the Target Photo and object set+ ...+Xn*Yn, * are multiplying.
Step 306: terminal shows every picture in picture set according to the first similarity.
Optionally, terminal can be shown every picture in picture set by the sequence of the first similarity from big to small.
It in the embodiment of the present application, is more than the retrieval object of preset weights threshold value by selecting weight from picture to be retrievedAs target retrieval object, and object set is formed, each target retrieval retrieved required for automatically identifying in this way by terminalObject determines each target retrieval object pair then according to the object type of each target retrieval object in the object setThe picture retrieval system answered retrieves each target retrieval pair in the corresponding picture retrieval system of each target retrieval object respectivelyAs a corresponding at least picture, and picture set is formed, due to automatically identifying target retrieval object and composition object set,Determine the corresponding picture retrieval system of each target retrieval object, such terminal can be automatically using each mesh in object setMark retrieval object retrieving image in picture retrieval system so as to avoid the problem that manual retrieval goes out to omit, and can mentionHigh recall precision.
Subsequently calculate the first similarity between the every picture and the object set in the picture set；By the first phaseEvery picture in the picture set is shown like the sequence of degree from big to small, it in this way can be by the retrieval of multiple picture retrieval systemsAs a result integrated ordered displaying, so that the result of display is more comprehensively.
Embodiment 4
Referring to Fig. 7, the embodiment of the present application provides a kind of method of retrieving image, and this method can be applied to embodiment 1The network architecture of offer, and the executing subject of this method can be the terminal in the network architecture.This method comprises:
Step 401: terminal obtains object set, which includes at least one characteristics of image.
Terminal can be with display condition input interface.User can in the condition entry interface input picture feature.EventuallyEnd can show a condition entry interface, can also show the corresponding condition entry interface of each object type.
If terminal shows a condition entry interface, user can be inputted in the condition entry interface to every kind of objectThe characteristics of image of type.Such as user can input in the condition entry interface to the characteristics of image of facial image, to human bodyThe characteristics of image of image and/or characteristics of image to vehicle image.Terminal obtains every kind of object class from the condition entry interfaceThe characteristics of image of type.
If terminal shows the corresponding condition entry interface of each object type, user can be in each object type pairThe characteristics of image of each object type is inputted in the condition entry interface answered respectively.Such as user can be corresponding in facial imageThe characteristics of image of condition entry interface input facial image inputs human body image in the corresponding condition entry interface of human body imageCharacteristics of image and/or in the condition entry interface of vehicle image input vehicle image characteristics of image.Terminal is from each rightCharacteristics of image as obtaining every kind of object type in the corresponding condition entry interface of type.
Wherein, whether the characteristics of image of facial image may include whether to smile, wear glasses and/or gender etc..Human figureThe characteristics of image of picture may include whether cycle, whether knapsack and/or gender etc..The characteristics of image of vehicle image may include vehicleBoard information, vehicle color and/or vehicle etc..
Step 402: terminal determines the corresponding object type of each characteristics of image in the object set, determines each imageThe corresponding picture retrieval system of feature.
For belonging at least one characteristics of image of same target type in object set, terminal is corresponding by the object typePicture retrieval system be determined as at least one characteristics of image corresponding figure sign searching system.
For example, at least one characteristics of image for belonging to facial image, the corresponding picture of at least one characteristics of imageSearching system is face searching system；For belonging at least one characteristics of image of human body image, at least one characteristics of imageCorresponding picture retrieval system is people's physical examination cable system；For belonging at least one characteristics of image of vehicle image, this at least oneThe corresponding picture retrieval system of a characteristics of image is vehicle retrieval system.
Step 403: at least one characteristics of image for belonging to same target type in object set, terminal according to this extremelyA few characteristics of image, retrieves at least one characteristics of image in the corresponding picture retrieval system of at least one characteristics of imageA corresponding at least picture, and the picture retrieved composition pictures are closed.
For example, at least one characteristics of image for belonging to facial image in object set, according at least one figureAs feature, retrieval meets an at least picture at least one characteristics of image in face retrieval system；For object setIn at least one characteristics of image for belonging to human body image examined in human body searching system according at least one characteristics of imageRope meets an at least picture at least one characteristics of image；And for belonging to vehicle image extremely in object setA few characteristics of image, according at least one characteristics of image, retrieval meets at least one image in vehicle retrieval systemAn at least picture for feature.It is closed finally, the picture of retrieval can be formed pictures.
For at least one characteristics of image for belonging to same target type in object set, terminal is at least one figureAs the corresponding picture retrieval system transmission retrieval request message of feature, it includes at least one image which, which carries,Feature.The picture retrieval system receives the retrieval request message, extracts at least one figure of retrieval request message carryingAs feature, the corresponding at least picture of at least one characteristics of image is then retrieved from its picture library.The picture retrievalSystem can also obtain the second similarity between every picture and at least one characteristics of image in an at least picture.Second similarity of picture is bigger, and the degree that is consistent between the characteristics of image that picture includes and at least one characteristics of image is higher.
Wherein, which can be the picture that the second similarity is more than default similarity threshold.
The picture retrieval system can also obtain the pictorial information of every picture in an at least picture.The figure of picturePiece information may include camera site and/or shooting time of the picture etc.；If in the picture including facial image, the pictureInformation can also include whether the facial image smile, whether wear glasses and/or the information such as gender；If in the picture including peopleBody image, the pictorial information can also include human body image whether cycle, whether the information such as knapsack and/or gender；If shouldIt include vehicle image in picture, which can also include the information such as license board information, vehicle color and/or vehicle.
The picture retrieval system can will retrieve the second phase of every picture in an at least picture, every pictureTerminal is sent to like degree and pictorial information.
Terminal can receive every picture of each picture retrieval system transmission, the second similarity of every picture and picture letterIt ceases and forms picture set.
Step 404: terminal calculates the first similarity between every picture and object set in the picture set.
This step can be with are as follows: for any picture in the picture set, the picture is referred to as mesh for ease of descriptionIt marks on a map piece, determines that the pictorial information of Target Photo includes m characteristics of image in object set, according to numerical value m and Target PhotoThe second similar value, calculate the first similarity between Target Photo and object set.For example, can be with logarithm m and target figureThe first similarity between Target Photo and object set is calculated in the mode of second similar value of piece being weighted.
Step 405: terminal shows every picture in picture set according to the first similarity.
Optionally, terminal can be shown every picture in picture set by the sequence of the first similarity from big to small.
Optionally, after having run a picture retrieval, first can also be selected from the picture set this time obtainedThe maximum at least picture of similarity carries out picture retrieval next time, specific picture retrieval process as picture to be retrievedIdentical as step 301-306, details are not described herein.New condition can also be gradually increased during continuing retrieval, such asFor each retrieval object setting weight, the characteristics of image for inputting every kind of object type etc. in picture to be retrieved, by incrementallyRetrieval, improve the accuracy rate of search result.
In the embodiment of the present application, for belonging at least one characteristics of image of same target type in the object set,The corresponding picture retrieval system of the object type is determined as the corresponding picture retrieval system of at least one characteristics of image, at thisThe corresponding at least picture of at least one characteristics of image is retrieved in the corresponding picture retrieval system of at least one characteristics of image,And picture set is formed, due to the corresponding picture retrieval system of characteristics of image, such terminal can be automatically using in object setCharacteristics of image in picture retrieval system retrieving image and can be mentioned so as to avoid the problem that manual retrieval goes out to omitHigh recall precision.
Following is the application Installation practice, can be used for executing the application embodiment of the method.It is real for the application deviceUndisclosed details in example is applied, the application embodiment of the method is please referred to.
Embodiment 5
Referring to Fig. 8, the embodiment of the present application provides a kind of device 500 of retrieving image, and described device 500 includes:
First obtains module 501, and for obtaining object set, the object set includes at least one target retrieval pairAs target retrieval object is the subject image either characteristics of image in picture to be retrieved；
Determining module 502 determines institute for the object type according to each target retrieval object in the object setState the corresponding picture retrieval system of each target retrieval object；
Retrieval module 503 is used for according to each target retrieval object, respectively in each target retrieval object pairThe corresponding at least picture of each target retrieval object is retrieved in the picture retrieval system answered.
Optionally, the target retrieval object is the subject image in the picture to be retrieved, and described device 500 is also wrappedIt includes:
Second obtains module, for the object type according to each target retrieval object, obtains respectively described eachThe corresponding characteristics of image of target retrieval object；
The retrieval module, for corresponding according to each target retrieval object and each target retrieval objectCharacteristics of image retrieves each target retrieval pair in the corresponding picture retrieval system of each target retrieval object respectivelyAs a corresponding at least picture.
Optionally, described first module 501 is obtained, comprising:
Mark unit is retrieved in the picture to be retrieved for identifying each retrieval object for including in picture to be retrievedObject is located in its corresponding target frame, and one target frame when being located in another target frame there are a target frameAttribute be set as top-level boxes, the attribute of another target frame is set as bottom frame；
Acquiring unit, for obtaining the weight of each retrieval object in the picture to be retrieved；
Extraction unit, the retrieval object that preset weights threshold value is greater than for extracting weight in the picture to be retrieved correspond toTarget frame in image as target retrieval object.
Optionally, described device 500 further include:
Computing module, for calculating the first similarity between every picture and the object set in picture set,The picture set includes the corresponding at least picture of each target retrieval object；
Display module, for showing every picture according to the first similarity of every picture.
Optionally, the computing module includes:
Determination unit, for determining n target retrieval object in the object set that Target Photo includes, the meshPiece of marking on a map is any picture in the picture set, and n is the integer greater than 0；
Computing unit, for according to the weight of each target retrieval object in the n target retrieval object and describedThe second similarity between Target Photo and each target retrieval object, calculates the Target Photo and the object setBetween the first similarity.
Optionally, described device 500 further include:
Selecting module, for selecting the maximum at least picture of the first similarity from the picture set that the last time obtainsAs picture to be retrieved.
It in the embodiment of the present application, is more than the retrieval object of preset weights threshold value by selecting weight from picture to be retrievedAs target retrieval object, and object set is formed, each target retrieval retrieved required for automatically identifying in this way by terminalObject, then according to each target retrieval object in the object set, respectively in pair of each target retrieval objectAs retrieving the corresponding at least picture of each target retrieval object in the corresponding picture retrieval system of type, and form figurePiece set；Calculate the first similarity between the every picture and the object set in the picture set；It is similar by firstThe sequence of degree from big to small shows every picture in the picture set.Due to automatically identifying target retrieval object and compositionObject set, such terminal can be retrieved in picture retrieval system using each target retrieval object in object set automaticallySo as to avoid the problem that manual retrieval goes out to omit, and recall precision can be improved in picture.
About the device in above-described embodiment, wherein modules execute the concrete mode of operation in related this methodEmbodiment in be described in detail, no detailed explanation will be given here.
Embodiment 6
Fig. 9 shows the structural block diagram of the terminal 600 of an illustrative embodiment of the invention offer.The terminal 600 can be withIt is portable mobile termianl, such as: smart phone, tablet computer, MP3 player (Moving Picture ExpertsGroup Audio Layer III, dynamic image expert's compression standard audio level 3), MP4 (Moving PictureExperts Group Audio Layer IV, dynamic image expert's compression standard audio level 4) player, laptopOr desktop computer.Terminal 600 is also possible to referred to as other names such as user equipment, portable terminal, laptop terminal, terminal consoleClaim.
In general, terminal 600 includes: processor 601 and memory 602.
Processor 601 may include one or more processing cores, such as 4 core processors, 8 core processors etc..PlaceReason device 601 can use DSP (Digital Signal Processing, Digital Signal Processing), FPGA (Field-Programmable Gate Array, field programmable gate array), PLA (Programmable Logic Array, may be programmedLogic array) at least one of example, in hardware realize.Processor 601 also may include primary processor and coprocessor, masterProcessor is the processor for being handled data in the awake state, also referred to as CPU (Central ProcessingUnit, central processing unit)；Coprocessor is the low power processor for being handled data in the standby state.?In some embodiments, processor 601 can be integrated with GPU (Graphics Processing Unit, image processor),GPU is used to be responsible for the rendering and drafting of content to be shown needed for display screen.In some embodiments, processor 601 can also be wrappedAI (Artificial Intelligence, artificial intelligence) processor is included, the AI processor is for handling related machine learningCalculating operation.
Memory 602 may include one or more computer readable storage mediums, which canTo be non-transient.Memory 602 may also include high-speed random access memory and nonvolatile memory, such as oneOr multiple disk storage equipments, flash memory device.In some embodiments, the non-transient computer in memory 602 canStorage medium is read for storing at least one instruction, at least one instruction performed by processor 601 for realizing this ShenPlease in embodiment of the method provide retrieving image method.
In some embodiments, terminal 600 is also optional includes: peripheral device interface 603 and at least one peripheral equipment.It can be connected by bus or signal wire between processor 601, memory 602 and peripheral device interface 603.Each peripheral equipmentIt can be connected by bus, signal wire or circuit board with peripheral device interface 603.Specifically, peripheral equipment includes: radio circuit604, at least one of touch display screen 605, camera 606, voicefrequency circuit 607, positioning component 608 and power supply 609.
Peripheral device interface 603 can be used for I/O (Input/Output, input/output) is relevant outside at least onePeripheral equipment is connected to processor 601 and memory 602.In some embodiments, processor 601, memory 602 and peripheral equipmentInterface 603 is integrated on same chip or circuit board；In some other embodiments, processor 601, memory 602 and outerAny one or two in peripheral equipment interface 603 can realize on individual chip or circuit board, the present embodiment to this notIt is limited.
Radio circuit 604 is for receiving and emitting RF (Radio Frequency, radio frequency) signal, also referred to as electromagnetic signal.It penetratesFrequency circuit 604 is communicated by electromagnetic signal with communication network and other communication equipments.Radio circuit 604 turns electric signalIt is changed to electromagnetic signal to be sent, alternatively, the electromagnetic signal received is converted to electric signal.Optionally, radio circuit 604 wrapsIt includes: antenna system, RF transceiver, one or more amplifiers, tuner, oscillator, digital signal processor, codec chipGroup, user identity module card etc..Radio circuit 604 can be carried out by least one wireless communication protocol with other terminalsCommunication.The wireless communication protocol includes but is not limited to: WWW, Metropolitan Area Network (MAN), Intranet, each third generation mobile communication network (2G, 3G,4G and 5G), WLAN and/or WiFi (Wireless Fidelity, Wireless Fidelity) network.In some embodiments, it penetratesFrequency circuit 604 can also include NFC (Near Field Communication, wireless near field communication) related circuit, thisApplication is not limited this.
Display screen 605 is for showing UI (User Interface, user interface).The UI may include figure, text, figureMark, video and its their any combination.When display screen 605 is touch display screen, display screen 605 also there is acquisition to showThe ability of the touch signal on the surface or surface of screen 605.The touch signal can be used as control signal and be input to processor601 are handled.At this point, display screen 605 can be also used for providing virtual push button and/or dummy keyboard, also referred to as soft button and/orSoft keyboard.In some embodiments, display screen 605 can be one, and the front panel of terminal 600 is arranged；In other embodimentsIn, display screen 605 can be at least two, be separately positioned on the different surfaces of terminal 600 or in foldover design；In still other realityIt applies in example, display screen 605 can be flexible display screen, be arranged on the curved surface of terminal 600 or on fold plane.Even, it showsDisplay screen 605 can also be arranged to non-rectangle irregular figure, namely abnormity screen.Display screen 605 can use LCD (LiquidCrystal Display, liquid crystal display), OLED (Organic Light-Emitting Diode, Organic Light Emitting Diode)Etc. materials preparation.
CCD camera assembly 606 is for acquiring image or video.Optionally, CCD camera assembly 606 include front camera andRear camera.In general, the front panel of terminal is arranged in front camera, the back side of terminal is arranged in rear camera.OneIn a little embodiments, rear camera at least two is main camera, depth of field camera, wide-angle camera, focal length camera shooting respectivelyAny one in head, to realize that main camera and the fusion of depth of field camera realize background blurring function, main camera and wide-angleCamera fusion realizes that pan-shot and VR (Virtual Reality, virtual reality) shooting function or other fusions are clappedCamera shooting function.In some embodiments, CCD camera assembly 606 can also include flash lamp.Flash lamp can be monochromatic warm flash lamp,It is also possible to double-colored temperature flash lamp.Double-colored temperature flash lamp refers to the combination of warm light flash lamp and cold light flash lamp, can be used for notWith the light compensation under colour temperature.
Voicefrequency circuit 607 may include microphone and loudspeaker.Microphone is used to acquire the sound wave of user and environment, and willSound wave, which is converted to electric signal and is input to processor 601, to be handled, or is input to radio circuit 604 to realize voice communication.For stereo acquisition or the purpose of noise reduction, microphone can be separately positioned on the different parts of terminal 600 to be multiple.MikeWind can also be array microphone or omnidirectional's acquisition type microphone.Loudspeaker is then used to that processor 601 or radio circuit will to be come from604 electric signal is converted to sound wave.Loudspeaker can be traditional wafer speaker, be also possible to piezoelectric ceramic loudspeaker.WhenWhen loudspeaker is piezoelectric ceramic loudspeaker, the audible sound wave of the mankind can be not only converted electrical signals to, it can also be by telecommunicationsNumber the sound wave that the mankind do not hear is converted to carry out the purposes such as ranging.In some embodiments, voicefrequency circuit 607 can also includeEarphone jack.
Positioning component 608 is used for the current geographic position of positioning terminal 600, to realize navigation or LBS (LocationBased Service, location based service).Positioning component 608 can be the GPS (Global based on the U.S.Positioning System, global positioning system), China dipper system or Russia Galileo system positioning groupPart.
Power supply 609 is used to be powered for the various components in terminal 600.Power supply 609 can be alternating current, direct current,Disposable battery or rechargeable battery.When power supply 609 includes rechargeable battery, which can be wired charging electricityPond or wireless charging battery.Wired charging battery is the battery to be charged by Wireline, and wireless charging battery is by wirelessThe battery of coil charges.The rechargeable battery can be also used for supporting fast charge technology.
In some embodiments, terminal 600 further includes having one or more sensors 610.The one or more sensors610 include but is not limited to: acceleration transducer 611, gyro sensor 612, pressure sensor 613, fingerprint sensor 614,Optical sensor 615 and proximity sensor 616.
The acceleration that acceleration transducer 611 can detecte in three reference axis of the coordinate system established with terminal 600 is bigIt is small.For example, acceleration transducer 611 can be used for detecting component of the acceleration of gravity in three reference axis.Processor 601 canWith the acceleration of gravity signal acquired according to acceleration transducer 611, touch display screen 605 is controlled with transverse views or longitudinal viewFigure carries out the display of user interface.Acceleration transducer 611 can be also used for the acquisition of game or the exercise data of user.
Gyro sensor 612 can detecte body direction and the rotational angle of terminal 600, and gyro sensor 612 canTo cooperate with acquisition user to act the 3D of terminal 600 with acceleration transducer 611.Processor 601 is according to gyro sensor 612Following function may be implemented in the data of acquisition: when action induction (for example changing UI according to the tilt operation of user), shootingImage stabilization, game control and inertial navigation.
The lower layer of side frame and/or touch display screen 605 in terminal 600 can be set in pressure sensor 613.Work as pressureWhen the side frame of terminal 600 is arranged in sensor 613, user can detecte to the gripping signal of terminal 600, by processor 601Right-hand man's identification or prompt operation are carried out according to the gripping signal that pressure sensor 613 acquires.When the setting of pressure sensor 613 existsWhen the lower layer of touch display screen 605, the pressure operation of touch display screen 605 is realized to UI circle according to user by processor 601Operability control on face is controlled.Operability control includes button control, scroll bar control, icon control, menuAt least one of control.
Fingerprint sensor 614 is used to acquire the fingerprint of user, collected according to fingerprint sensor 614 by processor 601The identity of fingerprint recognition user, alternatively, by fingerprint sensor 614 according to the identity of collected fingerprint recognition user.It is identifyingWhen the identity of user is trusted identity out, the user is authorized to execute relevant sensitive operation, the sensitive operation packet by processor 601Include solution lock screen, check encryption information, downloading software, payment and change setting etc..Terminal can be set in fingerprint sensor 614600 front, the back side or side.When being provided with physical button or manufacturer Logo in terminal 600, fingerprint sensor 614 can be withIt is integrated with physical button or manufacturer Logo.
Optical sensor 615 is for acquiring ambient light intensity.In one embodiment, processor 601 can be according to opticsThe ambient light intensity that sensor 615 acquires controls the display brightness of touch display screen 605.Specifically, when ambient light intensity is higherWhen, the display brightness of touch display screen 605 is turned up；When ambient light intensity is lower, the display for turning down touch display screen 605 is brightDegree.In another embodiment, the ambient light intensity that processor 601 can also be acquired according to optical sensor 615, dynamic adjustThe acquisition parameters of CCD camera assembly 606.
Proximity sensor 616, also referred to as range sensor are generally arranged at the front panel of terminal 600.Proximity sensor 616For acquiring the distance between the front of user Yu terminal 600.In one embodiment, when proximity sensor 616 detects useWhen family and the distance between the front of terminal 600 gradually become smaller, touch display screen 605 is controlled from bright screen state by processor 601It is switched to breath screen state；When proximity sensor 616 detects user and the distance between the front of terminal 600 becomes larger,Touch display screen 605 is controlled by processor 601 and is switched to bright screen state from breath screen state.
It will be understood by those skilled in the art that the restriction of the not structure paired terminal 600 of structure shown in Fig. 9, can wrapIt includes than illustrating more or fewer components, perhaps combine certain components or is arranged using different components.
Those skilled in the art will readily occur to its of the application after considering specification and practicing application disclosed hereinIts embodiment.This application is intended to cover any variations, uses, or adaptations of the application, these modifications, purposes orPerson's adaptive change follows the general principle of the application and including the undocumented common knowledge in the art of the applicationOr conventional techniques.The description and examples are only to be considered as illustrative, and the true scope and spirit of the application are by followingClaim is pointed out.
It should be understood that the application is not limited to the precise structure that has been described above and shown in the drawings, andAnd various modifications and changes may be made without departing from the scope thereof.Scope of the present application is only limited by the accompanying claims.