CN108573268A - Image-recognizing method and device, image processing method and device and storage medium - Google Patents

Abstract

The embodiment provides a kind of image-recognizing method and device, image processing method and device and storage mediums.Image-recognizing method includes：Obtain the first image and the second image；Calculate the similarity score between the first image and the second image；And similarity score is compared with first threshold and second threshold, to obtain recognition result, wherein, if similarity score is greater than or equal to first threshold, then recognition result is that the object that the first image includes and the object that the second image includes are identical, if similarity score is less than or equal to second threshold, then recognition result is that the object that the first image includes and the object that the second image includes are different, if similarity score is more than second threshold and is less than first threshold, whether then recognition result is the first image object for including and the second image includes object is identical to belong to fuzzy interval.It is whether identical that the object that two images include is weighed using dual threshold, it is possible to prevente effectively from misrecognition caused by single threshold mode.

Description

Image-recognizing method and device, image processing method and device and storage medium

Technical field

The present invention relates to computer realm, relate more specifically to a kind of image-recognizing method and device, image processing method With device and storage medium.

Background technology

Existing image identification system (such as face identification system) judges that object that two images include is (such as to be identified Known face in face and bottom library) whether it is identical when, the scoring taken is as follows：" recognition threshold " is pre-set, for example, Recognition threshold is set as 75 points, when the similarity score between two images is greater than or equal to recognition threshold, it is believed that two The object that image includes is identical, when the similarity score between two images is less than recognition threshold, it is believed that two images include Object it is different.

The defect of above-mentioned scoring is：Though when similarity score be less than recognition threshold with recognition threshold it is very close When (such as recognition threshold is 75 points, and similarity score is 74.9 points), the object that two images include may be actually same An object, but different objects are identified as, to cause identification inaccurate.

Invention content

The present invention is proposed in view of the above problem.The present invention provides a kind of image-recognizing methods and device, image Treating method and apparatus and storage medium.

According to an aspect of the present invention, a kind of image-recognizing method is provided.This method includes：Obtain the first image and second Image；Calculate the similarity score between described first image and second image；And by the similarity score and One threshold value compares with second threshold, to obtain recognition result, wherein if the similarity score is greater than or equal to described the One threshold value, then the recognition result object that be the described first image object that includes include with second image is identical, such as Similarity score described in fruit is less than or equal to the second threshold, then the recognition result is the object that described first image includes It is different with the object that second image includes, if the similarity score is more than the second threshold and less than described the One threshold value, then the recognition result is whether the object that described first image includes and the object that second image includes are identical Belong to fuzzy interval.

Illustratively, described image recognition methods further includes：Establish identification record list；And by the recognition result, The image information recording of the image information of described first image and second image is in the identification record list.

Illustratively, described first image is images to be recognized, and second image is stored in the known figure in the library of bottom One of as, the image information of described first image includes the title of described first image, storage location, acquisition time and acquisition position One or more in setting, the image information of second image includes the title and/or storage location of second image.

Illustratively, described image recognition methods further includes：Export the recognition result.

Illustratively, described image recognition methods further includes：It is pair that described first image includes in the recognition result As whether the object for including with second image identical belong to fuzzy interval in the case of, reception be used to indicate first figure As comprising object and the second image whether identical artificial markup information of object that includes；And according to the artificial mark It is identical that the recognition result is updated to the object that object that described first image includes and second image include by note information Or it is different.

According to a further aspect of the invention, a kind of image processing method is provided.This method includes：It obtains and is adopted in specific time period Multiple images to be recognized of collection；For each in the multiple images to be recognized, using above-mentioned image-recognizing method, by this Images to be recognized as described first image, using each known image being stored in the library of bottom as second image at Reason, to obtain the recognition result of the images to be recognized and each known image；And for the multiple images to be recognized In each, the processing of the images to be recognized is determined according to the recognition result of the images to be recognized and each known image As a result, wherein if the images to be recognized and the object phase for including in the presence of expression in the recognition result of each known image Same recognition result, then the handling result of the images to be recognized is that the object that the images to be recognized includes belongs to known object, such as Fruit images to be recognized recognition result different from the object that the recognition result of each known image includes there is only expression, Then the handling result of the images to be recognized is that the object that the images to be recognized includes belongs to unknown object, if the images to be recognized It is indicated there is no the identical recognition result of object for indicating to include and existing with the recognition result of each known image Belong to the recognition result of fuzzy interval, then the handling result of the images to be recognized is the images to be recognized object that includes and described Indicate to belong to whether object that the corresponding known image of recognition result of fuzzy interval includes is identical belongs to fuzzy interval.

Illustratively, described image processing method further includes：To it is in the multiple images to be recognized, all it includes Object and described indicate to belong to whether object that the corresponding known image of recognition result of fuzzy interval includes is identical belongs to fuzzy The images to be recognized in section carries out duplicate removal, to obtain unduplicated image collection；And it is all in output described image set The image information of images to be recognized.

Illustratively, described image processing method further includes：For each images to be recognized in described image set, connect Receipts are used to indicate the artificial the markup information whether object that the images to be recognized includes belongs to known object；And for the figure Each images to be recognized in image set conjunction, this is updated to according to the artificial markup information by the handling result of the images to be recognized The object that images to be recognized includes belongs to known object or unknown object.

Illustratively, the bottom library is stranger bottom library, and described image processing method further includes：It waits knowing for the multiple Each in other image belongs to known right if the handling result of the images to be recognized is the object that the images to be recognized includes As then by the image information recording of the handling result of the images to be recognized and the images to be recognized in stranger's identification record list In；For each in the multiple images to be recognized, if the handling result of the images to be recognized is the images to be recognized Including object belong to unknown object, then the images to be recognized is added in the library of the stranger bottom.

According to a further aspect of the invention, a kind of pattern recognition device is provided.The device includes：Image collection module is used In the first image of acquisition and the second image；Similarity calculation module, for calculate described first image and second image it Between similarity score；And comparison module, for the similarity score to be compared with first threshold and second threshold, with Obtain recognition result, wherein if the similarity score is greater than or equal to the first threshold, the recognition result is institute It is identical to state the object that object that the first image includes and second image include, if the similarity score is less than or equal to The second threshold, then the object that the recognition result is the described first image object that includes and second image includes is not Together, if the similarity score is more than the second threshold and is less than the first threshold, the recognition result is institute It states the object that the first image includes and whether the object that second image includes is identical belongs to fuzzy interval.

Illustratively, described image identification device further includes：Module is established in list, for establishing identification record list；With And logging modle, it is used for the image information of the recognition result, the image information and second image of described first image It is recorded in the identification record list.

Illustratively, described first image is images to be recognized, and second image is stored in the known figure in the library of bottom One of as, the image information of described first image includes the title of described first image, storage location, acquisition time and acquisition position One or more in setting, the image information of second image includes the title and/or storage location of second image.

Illustratively, described image identification device further includes：Output module, for exporting the recognition result.

Illustratively, described image identification device further includes：Receiving module, for being described first in the recognition result Whether object that the object and second image that image includes include identical belong to fuzzy interval in the case of, receive for referring to Show the object that described first image includes and the whether identical artificial markup information of object that second image includes；And more New module, for the recognition result to be updated to the object and institute that described first image includes according to the artificial markup information It is identical or different to state the object that the second image includes.

According to a further aspect of the invention, a kind of image processing apparatus is provided.The device includes：Image collection module is used In the multiple images to be recognized acquired in acquisition specific time period；Processing module, for in the multiple images to be recognized Each, it is every in the library of bottom by being stored in using the images to be recognized as described first image using above-mentioned pattern recognition device A known image is handled as second image, to obtain the identification of the images to be recognized and each known image As a result；And result determining module, for for each in the multiple images to be recognized, according to the images to be recognized with The recognition result of each known image determines the handling result of the images to be recognized, wherein if the images to be recognized with There is the object identical recognition result for indicating to include in the recognition result of each known image, then the images to be recognized Handling result is that the object that the images to be recognized includes belongs to known object, if the images to be recognized and each known figure The recognition result of picture is there is only the different recognition result of the object for including is indicated, then the handling result of the images to be recognized waits for for this The object that identification image includes belongs to unknown object, if in the recognition result of the images to be recognized and each known image There is no the identical recognition result of object for indicating to include and in the presence of the recognition result for indicating to belong to fuzzy interval, then this waits knowing The handling result of other image is the object that the images to be recognized includes and the recognition result correspondence for indicating to belong to fuzzy interval The known image object that includes whether identical belong to fuzzy interval.

Illustratively, described image processing unit further includes：Deduplication module, for in the multiple images to be recognized , it is all it includes object and described indicate that belonging to the object that the corresponding known image of recognition result of fuzzy interval includes is The no identical images to be recognized for belonging to fuzzy interval carries out duplicate removal, to obtain unduplicated image collection；And output module, it uses In the image information for exporting all images to be recognized in described image set.

Illustratively, described image processing unit further includes：Receiving module, for for each of described image set Images to be recognized, reception are used to indicate the artificial the markup information whether object that the images to be recognized includes belongs to known object； And update module, it is used for for each images to be recognized in described image set, it should according to the artificial markup information The handling result of images to be recognized is updated to the object that the images to be recognized includes and belongs to known object or unknown object.

Illustratively, the bottom library is stranger bottom library, and described image processing unit further includes：Logging modle, for pair Each in the multiple images to be recognized includes if the handling result of the images to be recognized is the images to be recognized Object belongs to known object, then by the image information recording of the handling result of the images to be recognized and the images to be recognized strange In people's identification record list；Add module is used for for each in the multiple images to be recognized, if the figure to be identified The handling result of picture is that the object that the images to be recognized includes belongs to unknown object, then the images to be recognized is added to the footpath between fields In the library of stranger bottom.

According to a further aspect of the invention, a kind of pattern recognition device, including processor and memory are provided, wherein institute State and be stored with computer program instructions in memory, when the computer program instructions are run by the processor for execute with Lower step：Obtain the first image and the second image；Calculate the similarity score between described first image and second image； And the similarity score compares with first threshold and second threshold, to obtain recognition result, wherein if the phase It is greater than or equal to the first threshold like degree scoring, then the recognition result is the described first image object that includes and described the The object that two images include is identical, if the similarity score is less than or equal to the second threshold, the recognition result The object that the object for including for described first image and second image include is different, if the similarity score is more than institute It states second threshold and is less than the first threshold, then the recognition result is the described first image object that includes and described the Whether the object that two images include is identical to belong to fuzzy interval.

Illustratively, it is additionally operable to execute following steps when the computer program instructions are run by the processor：It establishes Identification record list；And the recognition result, the image information of described first image and the image of second image are believed Breath is recorded in the identification record list.

Illustratively, described first image is images to be recognized, and second image is stored in the known figure in the library of bottom One of as, the image information of described first image includes the title of described first image, storage location, acquisition time and acquisition position One or more in setting, the image information of second image includes the title and/or storage location of second image.

Illustratively, for executing following steps when the computer program instructions are run by the processor：Output institute State recognition result.

Illustratively, for executing following steps when the computer program instructions are run by the processor：Described Whether recognition result is the described first image object that includes and second image includes object is identical to belong to fuzzy interval In the case of, reception is used to indicate the object that described first image includes and whether the object that second image includes be identical Artificial markup information；And the recognition result is updated to by pair that described first image includes according to the artificial markup information As the object for including with second image is identical or different.

According to a further aspect of the invention, a kind of image processing apparatus, including processor and memory are provided, wherein institute State and be stored with computer program instructions in memory, when the computer program instructions are run by the processor for execute with Lower step：Obtain the multiple images to be recognized acquired in specific time period；For each in the multiple images to be recognized, profit With above-mentioned image-recognizing method the known figure in each of bottom library will be stored in using the images to be recognized as described first image As being handled as second image, to obtain the recognition result of the images to be recognized and each known image；With And for each in the multiple images to be recognized, according to the identification knot of the images to be recognized and each known image Fruit determines the handling result of the images to be recognized, wherein if the identification knot of the images to be recognized and each known image There is the identical recognition result of object for indicating to include in fruit, then the handling result of the images to be recognized is the images to be recognized packet The object contained belongs to known object, if there is only indicate to wrap for the recognition result of the images to be recognized and each known image The different recognition result of the object that contains, then the handling result of the images to be recognized is that the object that the images to be recognized includes belongs to not Know object, there is no the object for indicating to include is identical if the images to be recognized is with the recognition result of each known image Recognition result and in the presence of indicating to belong to the recognition result of fuzzy interval, then the handling result of the images to be recognized is that this waits knowing Whether the object that the corresponding known image of recognition result that the object and the expression that other image includes belong to fuzzy interval includes It is identical to belong to fuzzy interval.

Illustratively, it is additionally operable to execute following steps when the computer program instructions are run by the processor：To institute State it is in multiple images to be recognized, all it includes object and it is described indicate belong to fuzzy interval recognition result it is corresponding Know that the whether identical images to be recognized for belonging to fuzzy interval of object that image includes carries out duplicate removal, to obtain unduplicated image set It closes；And the image information of all images to be recognized in output described image set.

Illustratively, it is additionally operable to execute following steps when the computer program instructions are run by the processor：For Each images to be recognized in described image set, reception are used to indicate whether the object that the images to be recognized includes belongs to known The artificial markup information of object；And for each images to be recognized in described image set, believed according to the artificial mark The handling result of the images to be recognized is updated to the object that the images to be recognized includes and belongs to known object or unknown object by breath.

Illustratively, the bottom library is stranger bottom library, when the computer program instructions are run by the processor also For executing following steps：For each in the multiple images to be recognized, if the handling result of the images to be recognized The object for including for the images to be recognized belongs to known object, then by the handling result of the images to be recognized and the images to be recognized Image information recording in stranger's identification record list；For each in the multiple images to be recognized, if should The handling result of images to be recognized is that the object that the images to be recognized includes belongs to unknown object, then adds the images to be recognized Into stranger bottom library.

According to a further aspect of the invention, a kind of storage medium is provided, stores program instruction on said storage, institute Program instruction is stated to be used to execute following steps at runtime：Obtain the first image and the second image；Calculate described first image and Similarity score between second image；And the similarity score compares with first threshold and second threshold, To obtain recognition result, wherein if the similarity score is greater than or equal to the first threshold, the recognition result is The object that described first image includes is identical with the object that second image includes, if the similarity score is less than or waits In the second threshold, then the recognition result is the object that described first image includes and the object that second image includes Difference, if the similarity score is more than the second threshold and is less than the first threshold, the recognition result is Whether object that the object and second image that described first image includes include is identical to belong to fuzzy interval.

Illustratively, described program instruction is used to execute following steps at runtime：Establish identification record list；And it will The image information recording of the recognition result, the image information of described first image and second image is in the identification record In list.

Illustratively, described first image is images to be recognized, and second image is stored in the known figure in the library of bottom One of as, the image information of described first image includes the title of described first image, storage location, acquisition time and acquisition position One or more in setting, the image information of second image includes the title and/or storage location of second image.

Illustratively, described program instruction is used to execute following steps at runtime：Export the recognition result.

Illustratively, described program instruction is used to execute following steps at runtime：It is described the in the recognition result Whether object that the object and second image that one image includes include identical belong to fuzzy interval in the case of, reception is used for Indicate the object that described first image includes and the whether identical artificial markup information of object that second image includes；And The recognition result is updated to the object and second image that described first image includes according to the artificial markup information Including object it is identical or different.

According to a further aspect of the invention, a kind of storage medium is provided, stores program instruction on said storage, institute Program instruction is stated to be used to execute following steps at runtime：Obtain the multiple images to be recognized acquired in specific time period；For institute Each in multiple images to be recognized is stated, using above-mentioned image-recognizing method, using the images to be recognized as first figure Picture is handled each known image being stored in the library of bottom as second image, with obtain the images to be recognized with The recognition result of each known image；It is to be identified according to this and for each in the multiple images to be recognized Image and the recognition result of each known image determine the handling result of the images to be recognized, wherein if this is to be identified Image with there is the identical recognition result of object for indicating to include in the recognition result of each known image, then this is to be identified The handling result of image is that the images to be recognized object that includes belongs to known object, if the images to be recognized with it is described each The recognition result of known image is there is only the different recognition result of object for indicating to include, then the handling result of the images to be recognized The object for including for the images to be recognized belongs to unknown object, if the identification of the images to be recognized and each known image As a result there is no the identical recognition result of object for indicating to include and in the presence of the recognition result for indicating to belong to fuzzy interval in, then The handling result of the images to be recognized is the object that the images to be recognized includes and the identification knot for indicating to belong to fuzzy interval Whether object that the corresponding known image of fruit includes is identical to belong to fuzzy interval.

Illustratively, described program instruction is additionally operable to execute following steps at runtime：To the multiple images to be recognized In, it is all it includes object and the expression belong to the object that the corresponding known image of recognition result of fuzzy interval includes Whether the identical images to be recognized for belonging to fuzzy interval carries out duplicate removal, to obtain unduplicated image collection；And described in output The image information of all images to be recognized in image collection.

Illustratively, described program instruction is additionally operable to execute following steps at runtime：For in described image set Each images to be recognized receives and is used to indicate the artificial mark the letter whether object that the images to be recognized includes belongs to known object Breath；And the handling result of the images to be recognized is updated to by pair that the images to be recognized includes according to the artificial markup information As belonging to known object or unknown object.

Illustratively, the bottom library is stranger bottom library, and described program instruction is additionally operable to execute following steps at runtime： For each in the multiple images to be recognized, include if the handling result of the images to be recognized is the images to be recognized Object belong to known object, then by the image information recording of the handling result of the images to be recognized and the images to be recognized on footpath between fields In stranger's identification record list；For each in the multiple images to be recognized, if the processing knot of the images to be recognized Fruit is that the object that the images to be recognized includes belongs to unknown object, then the images to be recognized is added to stranger bottom library In.

Image-recognizing method and device according to the ... of the embodiment of the present invention, image processing method and device and storage medium, by It is whether identical in weighing the object that two images include using dual threshold, therefore it is possible to prevente effectively from caused by single threshold mode Misrecognition avoids being mistakenly identified as two images of the script comprising same object to include different objects.

Description of the drawings

The embodiment of the present invention is described in more detail in conjunction with the accompanying drawings, the above and other purposes of the present invention, Feature and advantage will be apparent.Attached drawing is used for providing further understanding the embodiment of the present invention, and constitutes explanation A part for book is not construed as limiting the invention for explaining the present invention together with the embodiment of the present invention.In the accompanying drawings, Identical reference label typically represents same parts or step.

Fig. 1 shows for realizing image-recognizing method according to the ... of the embodiment of the present invention and device and image processing method and dress The schematic block diagram for the exemplary electronic device set；

Fig. 2 shows the schematic flow charts of image-recognizing method according to an embodiment of the invention；

Fig. 3 shows the schematic flow chart of image processing method according to an embodiment of the invention；

Fig. 4 shows the schematic block diagram of pattern recognition device according to an embodiment of the invention；

Fig. 5 shows the schematic block diagram of image processing apparatus according to an embodiment of the invention；And

Fig. 6 shows the schematic block diagram of image processing system according to an embodiment of the invention.

Specific implementation mode

In order to enable the object, technical solutions and advantages of the present invention become apparent, root is described in detail below with reference to accompanying drawings According to example embodiments of the present invention.Obviously, described embodiment is only a part of the embodiment of the present invention, rather than this hair Bright whole embodiments, it should be appreciated that the present invention is not limited by example embodiment described herein.Based on described in the present invention The embodiment of the present invention, those skilled in the art's obtained all other embodiment in the case where not making the creative labor It should all fall under the scope of the present invention.

In order to solve problem as described above, the embodiment of the present invention is provided at a kind of image-recognizing method and device, image Method and apparatus and storage medium are managed, it is whether identical that the object that two images include is weighed using dual threshold scoring. Image-recognizing method and device provided in an embodiment of the present invention, image processing method and device and storage medium are suitable for various people Face identifies and similar scene.

First, it describes referring to Fig.1 for realizing image-recognizing method according to the ... of the embodiment of the present invention and device and image The exemplary electronic device 100 for the treatment of method and apparatus.

As shown in Figure 1, electronic equipment 100 include one or more processors 102, it is one or more storage device 104, defeated Enter device 106, output device 108 and image collecting device 110, these components pass through bus system 112 and/or other forms Bindiny mechanism's (not shown) interconnection.It should be noted that the component and structure of electronic equipment 100 shown in FIG. 1 are only exemplary, and Unrestricted, as needed, the electronic equipment can also have other assemblies and structure.

The processor 102 can be central processing unit (CPU) or have data-handling capacity and/or instruction execution The processing unit of the other forms of ability, and it is desired to execute to control other components in the electronic equipment 100 Function.

The storage device 104 may include one or more computer program products, and the computer program product can To include various forms of computer readable storage mediums, such as volatile memory and/or nonvolatile memory.It is described easy The property lost memory is such as may include random access memory (RAM) and/or cache memory (cache).It is described non- Volatile memory is such as may include read-only memory (ROM), hard disk, flash memory.In the computer readable storage medium On can store one or more computer program instructions, processor 102 can run described program instruction, to realize hereafter institute The client functionality (realized by processor) in the embodiment of the present invention stated and/or other desired functions.In the meter Can also store various application programs and various data in calculation machine readable storage medium storing program for executing, for example, the application program use and/or The various data etc. generated.

The input unit 106 can be the device that user is used for inputting instruction, and may include keyboard, mouse, wheat One or more of gram wind and touch screen etc..

The output device 108 can export various information (such as image and/or sound) to external (such as user), and And may include one or more of display, loud speaker etc..

Described image harvester 110 can acquire image (including video frame), and acquired image is stored in So that other components use in the storage device 104.Image collecting device 110 can be regular camera.It should be appreciated that figure As harvester 110 is only example, electronic equipment 100 can not include image collecting device 110.In such a case, it is possible to It is sent to electronic equipment 100 using other image acquisition device images, and by the image of acquisition.

Illustratively, for realizing image-recognizing method according to the ... of the embodiment of the present invention and device and image processing method and The exemplary electronic device of device can be realized in the equipment of personal computer or remote server etc..

In the following, image-recognizing method according to the ... of the embodiment of the present invention will be described with reference to figure 2.Fig. 2 shows according to the present invention one The schematic flow chart of the image-recognizing method 200 of a embodiment.As shown in Fig. 2, image-recognizing method 200 includes following step Suddenly.

In step S210, the first image and the second image are obtained.

It should be understood that clear in order to describe, distinguish that image recognition processes are related to using " first " and " second " herein two A image, " first " and " second " is not offered as sequence, also without any other particular meaning.First image and the second image can To be any suitable image for needing to compare identification, such as the facial image comprising face.In first image and the second image Either or both can be original image that the image acquisition devices such as camera arrive, can also be to original image into The image that row pretreatment obtains later.

Any one of first image and the second image or both can be sent by external equipment (such as cloud server) To electronic equipment 100 to be handled by the processor 102 of electronic equipment 100, the image that can also include by electronic equipment 100 Harvester 110 (such as camera) acquires and is transmitted to processor 102 and handled, and can also include by electronic equipment 100 Storage device 104 store and be transmitted to processor 102 and handled.

It will be understood by those skilled in the art that image recognition may include one-to-one image comparison or one-to-many image Comparison.For above-mentioned two situations, image-recognizing method as described herein is applicable in.For example, it is assumed that there are two facial images, User it is expected to know whether the face in the two facial images belongs to same people, image as described herein may be used at this time and know Other method handles the two facial images.In another example, it is assumed that district monitoring system collects a facial image, and user it is expected to obtain Know whether collected face belongs to neighbours living, then image-recognizing method as described herein may be used by collected face Image and the facial image of neighbours living are compared one by one, are confirmed whether that there are known neighbours livings and collected face Match.

In step S220, the similarity score between the first image and the second image is calculated.

It is exemplary and not restrictive, any suitable existing or phase in the cards in the future may be used in step S220 It is realized like degree algorithm, for example, the similarity calculating method based on distance, the similarity calculating method etc. based on hash method.

In step S230, similarity score is compared with first threshold and second threshold, to obtain recognition result, In, if similarity score is greater than or equal to first threshold, recognition result is the object and the second image that the first image includes Including object it is identical, if similarity score is less than or equal to second threshold, recognition result is the first image pair for including As the object for including with the second image is different, if similarity score is more than second threshold and is less than first threshold, identify As a result it is the first image object for including and whether object that the second image includes is identical belongs to fuzzy interval.

First threshold and second threshold can be any suitable values, can be set as needed, the present invention is not to this It is limited.For example, first threshold can be 75 points, second threshold can be 65 points.In actual treatment, when between two images Similarity score be greater than or equal to first threshold (such as >=75 points) when, it may be determined that the object that two images include is identical. When the similarity score between two images is less than or equal to second threshold (such as≤65 points), it may be determined that two image packets The object contained is different.When (such as ＞ 65 divides simultaneously the similarity score between two images between second threshold and first threshold And ＜ 75 divides) when, it may be determined that whether object that two images include is identical to belong to fuzzy interval, that is to say, that two images Including this whether identical event of object be fuzzy, temporarily there is no final conclusion.Mould is belonged to for recognition result expression The image for pasting section, may be used other modes to further determine that whether the object that two images include is identical, illustratively And not restrictive, artificial notation methods may be used to determine whether the object that two images include is identical.

Image-recognizing method according to the ... of the embodiment of the present invention, due to weighing the object that two images include using dual threshold It is whether identical, therefore it is possible to prevente effectively from misrecognition caused by single threshold mode, that is, avoid script including the two of same object A image is mistakenly identified as including different objects.

Illustratively, image-recognizing method according to the ... of the embodiment of the present invention can be in setting with memory and processor It is realized in standby, device or system.

Image-recognizing method according to the ... of the embodiment of the present invention can be deployed at independent equipment, for example, can be deployed in At personal computer or server.Alternatively, image-recognizing method according to the ... of the embodiment of the present invention is deployed in which can also be distributed At server end (or high in the clouds) and client.For example, can acquire or store in the first image and the second image in client Either or both, client sends the image for collecting or being stored to server end (or high in the clouds), by server end (or High in the clouds) carry out image recognition.

According to embodiments of the present invention, image processing method 200 can also include：Establish identification record list；And it will know The image information recording of other result, the image information of the first image and the second image is in identification record list.

Identification record list is established, it, can be by two images after carrying out image recognitions for two images every time Identification information is recorded, to facilitate user subsequently to check the identification information of image.User the identification information for checking image it Afterwards, it can receive, modify or supplement recognition result, recognition result can also be directed to and indicate belong to the image of fuzzy interval into pedestrian Work marks.Illustratively, identification information may include the image information of the recognition result and two images of two images.It is exemplary Ground, identification information can also include the similarity score between two images.

According to embodiments of the present invention, the first image can be images to be recognized, and the second image can be stored in the library of bottom One of known image, the image information of the first image includes the title of the first image, storage location, acquisition time and acquisition position One or more in setting, the image information of the second image includes the title and/or storage location of the second image.

A kind of with reference to image-recognizing method applying situation, i.e. the campus teaching building stream of people manages, to describe this implementation Example.Assuming that teaching building entrance is mounted with camera, camera can acquire the facial image of the personnel for the teaching building that comes in and goes out.For adopting For some facial image collected, itself and all student's facial images in the library of student bottom can be compared one by one. During the comparison process, the identification information of collected facial image and each student's facial image can be recorded.For example, In a line of identification record list, title (such as the name that camera distributes automatically of collected facial image can be recorded Claim or the title that manually marks), storage location, acquisition time or acquisition position (such as geographical location of camera) etc., may be used also To store current title, the storage location of student's facial image etc. for participating in comparison.Further, it is also possible in identification record list Record the recognition result of collected facial image and each student's facial image, for example, belong to same people, be not belonging to same people, Belong to fuzzy interval etc..

According to embodiments of the present invention, image-recognizing method 200 can also include：Export recognition result.

The way of output may include showing or being exported to it by wired or wireless way on the display apparatus

His equipment.It is alternatively possible to directly above-mentioned identification record list be exported, so that user checks.

Illustratively, it is the object and the second image that the first image includes in the recognition result of the first image and the second image Including object it is whether identical belong to fuzzy interval in the case of, can select to export the recognition result, so that user checks. User can then select artificial notation methods to determine whether the object that the first image and the second image include is identical.

According to embodiments of the present invention, image-recognizing method 200 can also include：It is that the first image includes in recognition result Whether object that object and the second image include identical belong to fuzzy interval in the case of, reception is used to indicate the first image and includes Object and the second image whether identical artificial markup information of object for including；And identification is tied according to artificial markup information It is identical or different that fruit is updated to the object that the object that the first image includes and the second image include.

Fuzzy interval (similarity score ＞ 65 as described above divides and ＜ 75 divides) is belonged to for recognition result expression The first image and the second image, can judge whether the object that two images include identical by user.User can carry out Artificial mark, such as inputted by above-mentioned input unit 106 and to be used to indicate object that the first image includes and the second image includes The whether identical artificial markup information of object.It, can be according to the letter when image identification system receives artificial markup information The recognition result of instruction update the first image and the second image of breath.

Situation about can not be determined come auxiliary machinery by artificial notation methods can make whole recognition results obtain in this way Processing thereby may be ensured that the accurate of recognition result and define.

According to a further aspect of the invention, a kind of image processing method is provided.Fig. 3 shows according to an embodiment of the invention The schematic flow chart of image processing method 300.As shown in figure 3, image processing method 300 includes the following steps.

In step S310, the multiple images to be recognized acquired in specific time period are obtained.

In step S320, for each in multiple images to be recognized, using above-mentioned image-recognizing method 200, by this Images to be recognized is handled each known image being stored in the library of bottom as the second image, as the first image to obtain Obtain the recognition result of the images to be recognized and each known image.

In step S330, for each in multiple images to be recognized, according to the images to be recognized and each known figure The recognition result of picture determines the handling result of the images to be recognized, wherein if the images to be recognized and each known image There is the identical recognition result of object for indicating to include in recognition result, then the handling result of the images to be recognized is that this is to be identified The object that image includes belongs to known object, if there is only expressions for the recognition result of the images to be recognized and each known image Including the different recognition result of object, then the handling result of the images to be recognized is that the object that the images to be recognized includes belongs to Unknown object, if the images to be recognized is with the recognition result of each known image, there is no the object for indicating to include is identical Recognition result and the recognition result for belonging to fuzzy interval in the presence of expression, then the handling result of the images to be recognized is that this is to be identified The object that the corresponding known image of recognition result that the object and expression that image includes belong to fuzzy interval includes whether same genus In fuzzy interval.

The example of above-mentioned campus teaching building stream of people management is continued to use below to describe the realization method of image processing method 300.

It as described above, can be by all student's facial images in the collected facial image of camera and student bottom library It is compared one by one.Since whithin a period of time, camera can will be acquired with many facial images of continuous acquisition, therefore To each facial image compared with all student's facial images in the library of student bottom, in each facial image of determination Whether face belongs to student.

Assuming that camera is in certain specific time period (such as all a whole mornings 8：30-11：30 instructional blocks of time) in collect 1000 A facial image, number 1-1000.For No. 1 facial image, if by all in the facial image and student bottom library After stranger's face image compares, matched student's facial image (similarity score is greater than or equal to first threshold) is found, Then illustrate that the collected personnel of camera are known student.However, if No. 1 facial image and all students in the library of student bottom Facial image is different (similarity score all less than or be equal to second threshold), then illustrates that the collected personnel of camera are footpath between fields Stranger.Furthermore, it is also possible to which there are following situations：It is not found in the library of student bottom and the matched student's face figure of No. 1 facial image It is as (similarity score is greater than or equal to first threshold), but in the presence of the object for being confirmed as including together with No. 1 facial image The no identical student's facial image (similarity score is more than second threshold and is less than first threshold) for belonging to fuzzy interval, then may be used Belong to mould to determine whether face that face that the collected facial image of camera includes and student's facial image include is identical Paste section, i.e., can not temporarily determine it includes object be known student or stranger.In the above manner, 1 can be determined The object that number facial image includes belongs to student, stranger or fuzzy interval.Class may be used in remaining 2-1000 facial images It handles, repeats no more like mode.

According to embodiments of the present invention, image processing method 300 can also include：To in multiple images to be recognized, all It includes object and expression belong to whether object that the corresponding known image of recognition result of fuzzy interval includes identical to belong to The images to be recognized of fuzzy interval carries out duplicate removal, to obtain unduplicated image collection；And it is all in output image collection The image information of images to be recognized.

Assuming that in above-mentioned 1-1000 facial images, it has been determined that the face that 850 facial images include belongs to known Raw, the face that 50 facial images include belongs to stranger, and the face that 100 facial images include belongs to fuzzy interval, then may be used Facial image this 100 faces for including are belonged to fuzzy interval carries out duplicate removal.The mode of duplicate removal again may be by phase It realizes that multiple facial images that are larger to similarity, may belonging to same personnel screen like the mode of degree comparison, retains One of facial image.By operating above, 100 facial images can be filtered into such as 60 facial images.It crosses The object that filter includes belongs to the extra images to be recognized in the images to be recognized of fuzzy interval, can remove repetition and useless figure As information, for the output of subsequent image or artificial mark and etc. the processing for being conveniently conducive to save machine or manpower is provided when Between.

The way of output of the image information of images to be recognized may include display on the display apparatus or by wired or Wireless mode is exported to other equipment.

According to embodiments of the present invention, image processing method 300 can also include：It is to be identified for each of image collection Image, reception are used to indicate the artificial the markup information whether object that the images to be recognized includes belongs to known object；And it is right The handling result of the images to be recognized is updated to this by each images to be recognized in image collection according to artificial markup information The object that images to be recognized includes belongs to known object or unknown object.

Continue to use above-mentioned example, it is assumed that by after duplicate removal, remaining 60 facial images belong to fuzzy interval, and user can be with needle This 60 facial images are manually marked.For example, user can by above-mentioned input unit 106 input be used to indicate it is each The object that facial image includes is the artificial markup information of known student or stranger.When image processing system receives manually , can be according to the handling result of the instruction update facial image of the information when markup information, i.e., the object for including by facial image It is classified as known student or stranger.

According to embodiments of the present invention, bottom library is stranger bottom library, and image processing method 300 can also include：For multiple Each in images to be recognized belongs to if the handling result of the images to be recognized is the object that the images to be recognized includes Know object, then by the image information recording of the handling result of the images to be recognized and the images to be recognized in stranger's identification record In list；For each in multiple images to be recognized, if the handling result of the images to be recognized is the images to be recognized Including object belong to unknown object, then the images to be recognized is added in the library of stranger bottom.

In the example of above-mentioned campus teaching building stream of people management, student bottom library is described.In another example, it can also set Set stranger bottom library.Stranger bottom library can store the facial image of the previously stranger of disengaging teaching building.In one example, It can be by camera currently collected images to be recognized (being known as " facial image to be identified " in this example) and student bottom Library be confirmed as after being compared it includes object be not belonging to the facial image to be identified of known student in the library of stranger bottom Stranger's facial image compared one by one.It in another example, can be current collected all to be identified by camera Facial image is compared one by one with stranger's facial image in the library of stranger bottom.

When stranger's facial image in each facial image to be identified and stranger bottom library compares, if on footpath between fields Matched stranger's facial image (similarity score is greater than or equal to first threshold) is found in the library of stranger bottom, then illustrates camera Collected personnel are the stranger previously occurred.In such a case, it is possible to which the identification of the facial image to be identified is believed Breath is recorded in stranger's identification record list.Stranger's identification record list can pre-establish, whenever identifying The stranger previously occurred, then can add a record in stranger's identification record list.Optionally, images to be recognized Identification information may include the handling result of the images to be recognized and the image information of the images to be recognized.Optionally, it waits knowing The identification information of other image can also include the image letter for the corresponding known image of the identical recognition result of object for indicating to include Breath.

When stranger's facial image in each facial image to be identified and stranger bottom library compares, if waiting knowing Others' face image is different from all stranger's facial images (similarity score all less than or equal to second threshold), then illustrates The collected personnel of camera are stranger that is new, not occurring previously.In such a case, it is possible to by the people to be identified Face image is added to as the facial image of emerging stranger in the library of stranger bottom.In this way, when the stranger goes out next time Now, it can be found from the library of stranger bottom and the matched stranger's facial image of the stranger.

According to a further aspect of the invention, a kind of pattern recognition device is provided.Fig. 4 is shown according to an embodiment of the present invention Pattern recognition device 400 schematic block diagram.

As shown in figure 4, image processing apparatus 400 according to the ... of the embodiment of the present invention includes image collection module 410, similarity Computing module 420 and comparison module 430.The modules can execute the image recognition side described above in conjunction with Fig. 2 respectively Each step/function of method.Only the major function of each component of the pattern recognition device 400 is described below, and is omitted The detail content having been described above.

Image collection module 410 is for obtaining the first image and the second image.Image collection module 410 can be by Fig. 1 institutes The program instruction that is stored in 102 Running storage device 104 of processor in the electronic equipment shown is realized.

Similarity calculation module 420 is used to calculate the similarity score between the first image and the second image.Similarity meter Calculate the program instruction stored in 102 Running storage device 104 of processor in the electronic equipment that module 420 can be as shown in Figure 1 To realize.

Comparison module 430 is for comparing similarity score with first threshold and second threshold, to obtain recognition result, Wherein, if similarity score is greater than or equal to first threshold, recognition result is the object and the second figure that the first image includes As comprising object it is identical, if similarity score be less than or equal to second threshold, recognition result be the first image include The object that object and the second image include is different, if similarity score is more than second threshold and is less than first threshold, knows Whether other result is the first image object for including and the second image includes object is identical to belong to fuzzy interval.Comparison module 430 can be as shown in Figure 1 electronic equipment in 102 Running storage device 104 of processor in the program instruction that stores realize.

According to embodiments of the present invention, pattern recognition device 400 further includes：Module (not shown) is established in list, for establishing Identification record list；And logging modle (not shown), it is used for image information and the second image of recognition result, the first image Image information recording in identification record list.

According to embodiments of the present invention, the first image is images to be recognized, and the second image is stored in the known figure in the library of bottom One of as, the image information of the first image includes one in title, storage location, acquisition time and the acquisition position of the first image Item is multinomial, and the image information of the second image includes the title and/or storage location of the second image.

According to embodiments of the present invention, pattern recognition device 400 further includes：Output module (not shown), for exporting identification As a result.

According to embodiments of the present invention, pattern recognition device 400 further includes：Receiving module (not shown), for being tied in identification Whether fruit is the first image object for including and the second image includes object identical belong to fuzzy interval in the case of, receive and use The whether identical artificial markup information of object that the object and the second image for including in the first image of instruction include；And update mould Block (not shown), for recognition result to be updated to the object and the second image packet that the first image includes according to artificial markup information The object contained is identical or different.

According to a further aspect of the invention, a kind of image processing apparatus is provided.Fig. 5 is shown according to an embodiment of the present invention Image processing apparatus 500 schematic block diagram.

As shown in figure 5, image processing apparatus 500 according to the ... of the embodiment of the present invention includes image collection module 510, processing mould Block 520 and result determining module 530.The modules can execute the image processing method described above in conjunction with Fig. 3 respectively Each step/function.Only the major function of each component of the image processing apparatus 500 is described below, and omit with On the detail content that has been noted above.

Image collection module 510 is used to obtain the multiple images to be recognized acquired in specific time period.Image collection module 510 The program instruction that is stored in 102 Running storage device 104 of processor in electronic equipment that can be as shown in Figure 1 is realized.

Processing module 520 is used to, for each in multiple images to be recognized, utilize such as above-mentioned pattern recognition device 400, using the images to be recognized as the first image, using each known image being stored in the library of bottom as the second image at Reason, to obtain the recognition result of the images to be recognized and each known image.Processing module 520 can be as shown in Figure 1 electronics The program instruction that is stored in 102 Running storage device 104 of processor in equipment is realized.

As a result determining module 530 is used for for each in multiple images to be recognized, according to the images to be recognized and often The recognition result of a known image determines the handling result of the images to be recognized, wherein if the images to be recognized with it is each Know the identical recognition result of object for existing in the recognition result of image and indicating to include, then the handling result of the images to be recognized is The object that the images to be recognized includes belongs to known object, if the recognition result of the images to be recognized and each known image is only The recognition result different in the presence of the object for including is indicated, then the handling result of the images to be recognized is that the images to be recognized includes Object belongs to unknown object, if the images to be recognized is with the recognition result of each known image, there is no pairs for indicating to include Belong to the recognition result of fuzzy interval as identical recognition result and in the presence of expression, then the handling result of the images to be recognized is The object that the corresponding known image of recognition result that the object and expression that the images to be recognized includes belong to fuzzy interval includes is It is no identical to belong to fuzzy interval.As a result the operation of processor 102 in the electronic equipment that determining module 530 can be as shown in Figure 1 is deposited The program instruction that is stored in storage device 104 is realized.

According to embodiments of the present invention, image processing apparatus 500 further includes：Deduplication module (not shown), for waiting for multiple It is in identification image, all it includes object and expression belong to the corresponding known image of recognition result of fuzzy interval and include The whether identical images to be recognized for belonging to fuzzy interval of object carries out duplicate removal, to obtain unduplicated image collection；And output Module (not shown), the image information for exporting all images to be recognized in image collection.

According to embodiments of the present invention, image processing apparatus 500 further includes：Receiving module (not shown), for for image Each images to be recognized in set, reception are used to indicate the people whether object that the images to be recognized includes belongs to known object Work markup information；And update module (not shown), for for each images to be recognized in image collection, being marked according to artificial The handling result of the images to be recognized is updated to the object that the images to be recognized includes and belongs to known object or unknown by note information Object.

According to embodiments of the present invention, bottom library is stranger bottom library, and image processing apparatus 500 further includes：Logging modle is used for For each in multiple images to be recognized, if the handling result of the images to be recognized is pair that the images to be recognized includes As belonging to known object, then by the image information recording of the handling result of the images to be recognized and the images to be recognized in stranger In identification record list；Add module is used for for each in multiple images to be recognized, if the place of the images to be recognized Reason result is that the object that the images to be recognized includes belongs to unknown object, then the images to be recognized is added to stranger bottom library In.

Those of ordinary skill in the art may realize that lists described in conjunction with the examples disclosed in the embodiments of the present disclosure Member and algorithm steps can be realized with the combination of electronic hardware or computer software and electronic hardware.These functions are actually It is implemented in hardware or software, depends on the specific application and design constraint of technical solution.Professional technician Each specific application can be used different methods to achieve the described function, but this realization is it is not considered that exceed The scope of the present invention.

Fig. 6 shows the schematic block diagram of image processing system 600 according to an embodiment of the invention.Image recognition system System 600 includes image collecting device 610, storage device 620 and processor 630.

Image collecting device 610 is for acquiring image.Image collecting device 610 is optional, and image processing system 600 can Not include image collecting device 610.In such a case, it is possible to using the image needed for other image acquisition devices, And the image of acquisition is sent to image processing system 600.In addition, image can also be stored in storage device 620, Huo Zheyou Other External memory equipments are sent to image processing system 600.

The storage device 620 stores for realizing image-recognizing method according to the ... of the embodiment of the present invention or image processing method The program code (i.e. computer program instructions) of corresponding steps in method.

The processor 630 is for running the program code stored in the storage device 620, to execute according to the present invention The image-recognizing method of embodiment or the corresponding steps of image processing method, and for realizing figure according to the ... of the embodiment of the present invention As image collection module 410, similarity calculation module 420 and the comparison module 430 or image processing apparatus in identification device 400 Image collection module 510, processing module 520 in 500 and result determining module 530.

The processor 630 is described below for running the program code stored in the storage device 620, to execute root According to the embodiment of the corresponding steps of the image-recognizing method of the embodiment of the present invention.

In one embodiment, make described image processing system 600 when said program code is run by the processor 630 Execute following steps：Obtain the first image and the second image；It calculates similar between described first image and second image Degree scoring；And the similarity score compares with first threshold and second threshold, to obtain recognition result, wherein such as Similarity score described in fruit is greater than or equal to the first threshold, then the recognition result is the object that described first image includes It is identical with the object that second image includes, it is described if the similarity score is less than or equal to the second threshold Recognition result is the object difference that the object that described first image includes includes with second image, if the similarity is commented Divide more than the second threshold and be less than the first threshold, then the recognition result is the object that described first image includes Whether the object that includes with second image is identical to belong to fuzzy interval.

In one embodiment, also make described image processing system when said program code is run by the processor 630 600 execute：Establish identification record list；And by the recognition result, the image information and described second of described first image The image information recording of image is in the identification record list.

In one embodiment, described first image is images to be recognized, and second image is stored in the library of bottom One of known image, the image information of described first image include the title of described first image, storage location, acquisition time and One or more in acquisition position, the image information of second image includes title and/or the storage of second image Position.

In one embodiment, also make described image processing system when said program code is run by the processor 630 600 execute：Export the recognition result.

In one embodiment, also make described image processing system when said program code is run by the processor 630 600 execute：The object that the recognition result is the described first image object that includes and second image includes whether phase In the case of belonging to fuzzy interval, reception is used to indicate the object that described first image includes and second image includes The whether identical artificial markup information of object；And the recognition result is updated to described according to the artificial markup information The object that the object and second image that one image includes include is identical or different.

The processor 630 is described below for running the program code stored in the storage device 620, to execute root According to the embodiment of the corresponding steps of the image processing method of the embodiment of the present invention.

In one embodiment, make described image processing system 600 when said program code is run by the processor 630 Execute following steps：Obtain the multiple images to be recognized acquired in specific time period；For every in the multiple images to be recognized One, each of bottom library will be stored in using the images to be recognized as described first image using above-mentioned image-recognizing method Known image is handled as second image, to obtain the identification knot of the images to be recognized and each known image Fruit；And for each in the multiple images to be recognized, according to the images to be recognized and each known image Recognition result determines the handling result of the images to be recognized, wherein if the images to be recognized and each known image There is the identical recognition result of object for indicating to include in recognition result, then the handling result of the images to be recognized is that this is to be identified The object that image includes belongs to known object, if the recognition result of the images to be recognized and each known image there is only The different recognition result of object that expression includes, then the handling result of the images to be recognized is the object that the images to be recognized includes Belong to unknown object, there is no pairs for indicating to include if the images to be recognized is with the recognition result of each known image Belong to the recognition result of fuzzy interval as identical recognition result and in the presence of expression, then the handling result of the images to be recognized is The object and the expression that the images to be recognized includes belong to pair that the corresponding known image of recognition result of fuzzy interval includes As if no identical belong to fuzzy interval.

In one embodiment, also make described image processing system when said program code is run by the processor 630 600 execute：To it is in the multiple images to be recognized, all it includes object and described indicate to belong to the identification of fuzzy interval As a result the whether identical images to be recognized for belonging to fuzzy interval of object that corresponding known image includes carries out duplicate removal, to obtain not The image collection repeated；And the image information of all images to be recognized in output described image set.

In one embodiment, also make described image processing system when said program code is run by the processor 630 600 execute：For each images to be recognized in described image set, reception is used to indicate the object that the images to be recognized includes Whether the artificial markup information of known object is belonged to；And for each images to be recognized in described image set, according to institute It states artificial markup information the handling result of the images to be recognized is updated to the object that the images to be recognized includes and belong to known right As or unknown object.

In one embodiment, the bottom library is stranger bottom library, when said program code is run by the processor 630 Also execute described image processing system 600：For each in the multiple images to be recognized, if the images to be recognized Handling result be that the images to be recognized object that includes belongs to known object, then by the handling result of the images to be recognized and should The image information recording of images to be recognized is in stranger's identification record list；For each in the multiple images to be recognized It is a, if the handling result of the images to be recognized, which is the object that the images to be recognized includes, belongs to unknown object, this is waited knowing Other image is added in the library of the stranger bottom.

In addition, according to embodiments of the present invention, additionally providing a kind of storage medium, storing program on said storage Instruction, image-recognizing method when described program instruction is run by computer or processor for executing the embodiment of the present invention or The corresponding steps of image processing method, and filled for realizing pattern recognition device according to the ... of the embodiment of the present invention or image procossing Corresponding module in setting.The storage medium for example may include the storage unit, a of the storage card of smart phone, tablet computer Hard disk, read-only memory (ROM), Erasable Programmable Read Only Memory EPROM (EPROM), the portable compact disc of people's computer are read-only The arbitrary combination of memory (CD-ROM), USB storage or above-mentioned storage medium.

In one embodiment, the computer program instructions can to calculate when by computer or processor operation Machine or processor realize each function module of pattern recognition device or image processing apparatus according to the ... of the embodiment of the present invention, and And/or person can execute image-recognizing method or image processing method according to the ... of the embodiment of the present invention.

The computer program instructions are described below and make computer or processor when by computer or processor operation Execute the embodiment of the corresponding steps of image-recognizing method according to the ... of the embodiment of the present invention.

In one embodiment, the computer program instructions execute the computer when being run by computer following Step：Obtain the first image and the second image；Calculate the similarity score between described first image and second image；With And the similarity score compares with first threshold and second threshold, to obtain recognition result, wherein if described similar Degree scoring is greater than or equal to the first threshold, then the recognition result is the object and described second that described first image includes The object that image includes is identical, if the similarity score is less than or equal to the second threshold, the recognition result is The object that the object that described first image includes and second image include is different, if the similarity score is more than described Second threshold and it is less than the first threshold, then the recognition result is the described first image object and described second that includes Whether the object that image includes is identical to belong to fuzzy interval.

In one embodiment, the computer program instructions when being run by computer execute also the computer： Establish identification record list；And by the recognition result, the figure of the image information and second image of described first image As information is recorded in the identification record list.

In one embodiment, described first image is images to be recognized, and second image is stored in the library of bottom One of known image, the image information of described first image include the title of described first image, storage location, acquisition time and One or more in acquisition position, the image information of second image includes title and/or the storage of second image Position.

In one embodiment, the computer program instructions when being run by computer execute also the computer： Export the recognition result.

In one embodiment, the computer program instructions when being run by computer execute also the computer： Belong to mould whether the object that the recognition result is the described first image object that includes and second image includes is identical In the case of pasting section, whether reception is used to indicate the object that described first image includes and the object that second image includes Identical artificial markup information；And the recognition result is updated to by described first image packet according to the artificial markup information The object that the object and second image contained includes is identical or different.

The computer program instructions are described below and make computer or processor when by computer or processor operation Execute the embodiment of the corresponding steps of image processing method according to the ... of the embodiment of the present invention.

In one embodiment, the computer program instructions execute the computer when being run by computer following Step：Obtain the multiple images to be recognized acquired in specific time period；For each in the multiple images to be recognized, utilize Above-mentioned image-recognizing method, using the images to be recognized as described first image, each known image that will be stored in the library of bottom It is handled as second image, to obtain the recognition result of the images to be recognized and each known image；And For each in the multiple images to be recognized, according to the recognition result of the images to be recognized and each known image Determine the handling result of the images to be recognized, wherein if the recognition result of the images to be recognized and each known image The middle identical recognition result of object for including in the presence of expression, then the handling result of the images to be recognized is that the images to be recognized includes Object belong to known object, if there is only indicate to include for the recognition result of the images to be recognized and each known image The different recognition result of object, then the handling result of the images to be recognized is that the object that the images to be recognized includes belongs to unknown Object, if the images to be recognized is with the recognition result of each known image, there is no the object for indicating to include is identical Recognition result and the recognition result for belonging to fuzzy interval in the presence of expression, then the handling result of the images to be recognized is that this is to be identified Object that image includes and it is described indicate object that the corresponding known image of recognition result for belonging to fuzzy interval includes whether phase Belong to fuzzy interval.

In one embodiment, the computer program instructions when being run by computer execute also the computer： To it is in the multiple images to be recognized, all it includes object and it is described indicate belong to fuzzy interval recognition result correspondence The whether identical images to be recognized for belonging to fuzzy interval of the known image object that includes carry out duplicate removal, to obtain unduplicated figure Image set closes；And the image information of all images to be recognized in output described image set.

In one embodiment, the computer program instructions when being run by computer execute also the computer： For each images to be recognized in described image set, reception is used to indicate whether the object that the images to be recognized includes belongs to The artificial markup information of known object；And for each images to be recognized in described image set, according to the artificial mark The handling result of the images to be recognized is updated to the object that the images to be recognized includes and belongs to known object or unknown by note information Object.

In one embodiment, the bottom library is stranger bottom library, and the computer program instructions are run by computer When also execute the computer：For each in the multiple images to be recognized, if the processing of the images to be recognized As a result be that the object that the images to be recognized includes belongs to known object, then it is the handling result of the images to be recognized is to be identified with this The image information recording of image is in stranger's identification record list；For each in the multiple images to be recognized, such as The handling result of the fruit images to be recognized is that the object that the images to be recognized includes belongs to unknown object, then by the images to be recognized It is added in the library of the stranger bottom.

Each module in image processing system according to the ... of the embodiment of the present invention can pass through reality according to the ... of the embodiment of the present invention The processor for applying the electronic equipment of image recognition or image procossing runs the computer program instructions stored in memory in fact Calculating that is existing, or being stored in the computer readable storage medium of computer program product according to the ... of the embodiment of the present invention Machine instruction is realized when being run by computer.

Image-recognizing method and device according to the ... of the embodiment of the present invention, image processing method and device and storage medium, by It is whether identical in weighing the object that two images include using dual threshold, therefore it is possible to prevente effectively from caused by single threshold mode Misrecognition avoids being mistakenly identified as two images of the script comprising same object to include different objects.

Although describing example embodiment by reference to attached drawing here, it should be understood that the above example embodiment is merely exemplary , and be not intended to limit the scope of the invention to this.Those of ordinary skill in the art can carry out various changes wherein And modification, it is made without departing from the scope of the present invention and spiritual.All such changes and modifications are intended to be included in appended claims Within required the scope of the present invention.

Those of ordinary skill in the art may realize that lists described in conjunction with the examples disclosed in the embodiments of the present disclosure Member and algorithm steps can be realized with the combination of electronic hardware or computer software and electronic hardware.These functions are actually It is implemented in hardware or software, depends on the specific application and design constraint of technical solution.Professional technician Each specific application can be used different methods to achieve the described function, but this realization is it is not considered that exceed The scope of the present invention.

In several embodiments provided herein, it should be understood that disclosed device and method can pass through it Its mode is realized.For example, apparatus embodiments described above are merely indicative, for example, the division of the unit, only Only a kind of division of logic function, formula that in actual implementation, there may be another division manner, such as multiple units or component can be tied Another equipment is closed or is desirably integrated into, or some features can be ignored or not executed.

In the instructions provided here, numerous specific details are set forth.It is to be appreciated, however, that the implementation of the present invention Example can be put into practice without these specific details.In some instances, well known method, structure is not been shown in detail And technology, so as not to obscure the understanding of this description.

Similarly, it should be understood that in order to simplify the present invention and help to understand one or more of each inventive aspect, To in the description of exemplary embodiment of the present invention, each feature of the invention be grouped together into sometimes single embodiment, figure, Or in descriptions thereof.However, the method for the present invention should be construed to reflect following intention：It is i.e. claimed The present invention claims the more features of feature than being expressly recited in each claim.More precisely, such as corresponding power As sharp claim reflects, inventive point is that the spy of all features less than some disclosed single embodiment can be used It levies to solve corresponding technical problem.Therefore, it then follows thus claims of specific implementation mode are expressly incorporated in this specific Embodiment, wherein each claim itself is as a separate embodiment of the present invention.

It will be understood to those skilled in the art that other than mutually exclusive between feature, any combinations pair may be used All features disclosed in this specification (including adjoint claim, abstract and attached drawing) and so disclosed any method Or all processes or unit of equipment are combined.Unless expressly stated otherwise, this specification (including want by adjoint right Ask, make a summary and attached drawing) disclosed in each feature can be replaced by providing the alternative features of identical, equivalent or similar purpose.

In addition, it will be appreciated by those of skill in the art that although some embodiments described herein include other embodiments In included certain features rather than other feature, but the combination of the feature of different embodiments means in of the invention Within the scope of and form different embodiments.For example, in detail in the claims, embodiment claimed it is one of arbitrary It mode can use in any combination.

The all parts embodiment of the present invention can be with hardware realization, or to run on one or more processors Software module realize, or realized with combination thereof.It will be understood by those of skill in the art that can use in practice Microprocessor or digital signal processor (DSP) realize pattern recognition device or image procossing according to the ... of the embodiment of the present invention The some or all functions of some modules in device.The present invention is also implemented as executing method as described herein Some or all program of device (for example, computer program and computer program product).Such realization present invention Program can may be stored on the computer-readable medium, or can with one or more signal form.Such letter It number can download and to obtain from internet website, either provide on carrier signal or provide in any other forms.

It should be noted that the present invention will be described rather than limits the invention for above-described embodiment, and ability Field technique personnel can design alternative embodiment without departing from the scope of the appended claims.In the claims, Any reference mark between bracket should not be configured to limitations on claims.Word "comprising" does not exclude the presence of not Element or step listed in the claims.Word "a" or "an" before element does not exclude the presence of multiple such Element.The present invention can be by means of including the hardware of several different elements and being come by means of properly programmed computer real It is existing.In the unit claims listing several devices, several in these devices can be by the same hardware branch To embody.The use of word first, second, and third does not indicate that any sequence.These words can be explained and be run after fame Claim.

The above description is merely a specific embodiment or to the explanation of specific implementation mode, protection of the invention Range is not limited thereto, and any one skilled in the art in the technical scope disclosed by the present invention, can be easily Expect change or replacement, should be covered by the protection scope of the present invention.Protection scope of the present invention should be with claim Subject to protection domain.

Claims (36)

1. a kind of image-recognizing method, including：

Obtain the first image and the second image；

Calculate the similarity score between described first image and second image；And

The similarity score is compared with first threshold and second threshold, to obtain recognition result, wherein if the phase It is greater than or equal to the first threshold like degree scoring, then the recognition result is the described first image object that includes and described the The object that two images include is identical, if the similarity score is less than or equal to the second threshold, the recognition result The object that the object for including for described first image and second image include is different, if the similarity score is more than institute It states second threshold and is less than the first threshold, then the recognition result is the described first image object that includes and described the Whether the object that two images include is identical to belong to fuzzy interval.

2. image-recognizing method as described in claim 1, wherein described image recognition methods further includes：

Establish identification record list；And

By the image information recording of the recognition result, the image information of described first image and second image in the knowledge It Ji Lu not be in list.

3. image-recognizing method as claimed in claim 2, wherein described first image is images to be recognized, second figure Seem one of the known image being stored in the library of bottom, the image information of described first image include the title of described first image, One or more in storage location, acquisition time and acquisition position, the image information of second image includes described second The title and/or storage location of image.

4. image-recognizing method as described in claim 1, wherein described image recognition methods further includes：

Export the recognition result.

5. image-recognizing method as described in claim 1, wherein described image recognition methods further includes：

The object that the recognition result is the described first image object that includes and second image includes whether same genus In the case of fuzzy interval, reception is used to indicate the object that described first image includes and the object that second image includes Whether identical artificial markup information；And

The recognition result is updated to the object and described second that described first image includes according to the artificial markup information The object that image includes is identical or different.

6. a kind of image processing method, including：

Obtain the multiple images to be recognized acquired in specific time period；

For each in the multiple images to be recognized,

Using image-recognizing method as described in any one of claims 1 to 3, using the images to be recognized as first figure Picture is handled each known image being stored in the library of bottom as second image, with obtain the images to be recognized with The recognition result of each known image；And

The handling result of the images to be recognized is determined according to the recognition result of the images to be recognized and each known image, In, if the images to be recognized with there is the identical identification of object for indicating to include in the recognition result of each known image As a result, then the handling result of the images to be recognized is that the object that the images to be recognized includes belongs to known object, if this waits knowing Other image and the recognition result of each known image are there is only the different recognition result of the object for including is indicated, then this waits knowing The handling result of other image is that the images to be recognized object that includes belongs to unknown object, if the images to be recognized with it is described every Belong to fuzzy there is no the identical recognition result of object for indicating to include in the recognition result of a known image and in the presence of expression The recognition result in section, then the handling result of the images to be recognized is that the images to be recognized object that includes and the expression belong to Whether object that the corresponding known image of recognition result of fuzzy interval includes is identical to belong to fuzzy interval.

7. image processing method as claimed in claim 6, wherein described image processing method further includes：

To it is in the multiple images to be recognized, all it includes object and described indicate to belong to the recognition result of fuzzy interval The whether identical images to be recognized for belonging to fuzzy interval of the object that corresponding known image includes carries out duplicate removal, not repeated Image collection；And

Export the image information of all images to be recognized in described image set.

8. image processing method as claimed in claim 7, wherein described image processing method further includes：

For each images to be recognized in described image set,

Reception is used to indicate the artificial the markup information whether object that the images to be recognized includes belongs to known object；And

The handling result of the images to be recognized is updated to the object that the images to be recognized includes according to the artificial markup information Belong to known object or unknown object.

9. image processing method as claimed in claim 6, wherein the bottom library is stranger bottom library, described image processing side Method further includes：

For each in the multiple images to be recognized,

If the handling result of the images to be recognized, which is the object that the images to be recognized includes, belongs to known object, this is waited knowing The handling result of other image and the image information recording of the images to be recognized are in stranger's identification record list；

If the handling result of the images to be recognized, which is the object that the images to be recognized includes, belongs to unknown object, this is waited knowing Other image is added in the library of the stranger bottom.

10. a kind of pattern recognition device, including：

Image collection module, for obtaining the first image and the second image；

Similarity calculation module, for calculating the similarity score between described first image and second image；And

Comparison module, for the similarity score to be compared with first threshold and second threshold, to obtain recognition result, In, if the similarity score is greater than or equal to the first threshold, the recognition result includes for described first image Object and second image object that includes it is identical, if the similarity score is less than or equal to the second threshold, Then the recognition result is the object difference that the object that described first image includes includes with second image, if the phase More than the second threshold and it is less than the first threshold like degree scoring, then the recognition result includes for described first image Object and whether second image object that includes identical belongs to fuzzy interval.

11. pattern recognition device as claimed in claim 10, wherein described image identification device further includes：

Module is established in list, for establishing identification record list；And

Logging modle, for believing the image information of the recognition result, described first image and the image of second image Breath is recorded in the identification record list.

12. pattern recognition device as claimed in claim 11, wherein described first image is images to be recognized, described second Image is stored in one of the known image in the library of bottom, and the image information of described first image includes the name of described first image One or more in title, storage location, acquisition time and acquisition position, the image information of second image includes described The title and/or storage location of two images.

13. pattern recognition device as claimed in claim 10, wherein described image identification device further includes：

Output module, for exporting the recognition result.

14. pattern recognition device as claimed in claim 10, wherein described image identification device further includes：

Receiving module, in pair that the recognition result is the object that described first image includes and second image includes As if it is no it is identical belong to fuzzy interval in the case of, reception be used to indicate the object and second figure that described first image includes As comprising the whether identical artificial markup information of object；And

Update module, for the recognition result to be updated to pair that described first image includes according to the artificial markup information As the object for including with second image is identical or different.

15. a kind of image processing apparatus, including：

Image collection module, for obtaining the multiple images to be recognized acquired in specific time period；

Processing module, for for each in the multiple images to be recognized, utilizing any one of such as claim 10 to 12 The pattern recognition device will be stored in the known figure in each of bottom library using the images to be recognized as described first image As being handled as second image, to obtain the recognition result of the images to be recognized and each known image；With And

As a result determining module, for for each in the multiple images to be recognized, according to the images to be recognized with it is described The recognition result of each known image determines the handling result of the images to be recognized, wherein if the images to be recognized with it is described There is the identical recognition result of object for indicating to include, the then processing of the images to be recognized in the recognition result of each known image As a result it is that the object that the images to be recognized includes belongs to known object, if the images to be recognized and each known image Recognition result is there is only the different recognition result of the object for including is indicated, then the handling result of the images to be recognized is that this is to be identified The object that image includes belongs to unknown object, if the images to be recognized is not deposited with the recognition result of each known image In the identical recognition result of object for indicating to include and there is the recognition result for indicating to belong to fuzzy interval, then the figure to be identified The handling result of picture is that the object that the images to be recognized includes and the recognition result that the expression belongs to fuzzy interval are corresponding Know whether object that image includes is identical and belongs to fuzzy interval.

16. image processing apparatus as claimed in claim 15, wherein described image processing unit further includes：

Deduplication module, for it is in the multiple images to be recognized, all it includes object and it is described expression belong to fuzzy The whether identical images to be recognized for belonging to fuzzy interval of object that the corresponding known image of recognition result in section includes is gone Weight, to obtain unduplicated image collection；And

Output module, the image information for exporting all images to be recognized in described image set.

17. image processing apparatus as claimed in claim 16, wherein described image processing unit further includes：

Receiving module, for for each images to be recognized in described image set, reception to be used to indicate the images to be recognized Including object whether belong to the artificial markup information of known object；And

Update module is used for for each images to be recognized in described image set, should according to the artificial markup information The handling result of images to be recognized is updated to the object that the images to be recognized includes and belongs to known object or unknown object.

18. image processing apparatus as claimed in claim 15, wherein the bottom library is stranger bottom library, described image processing Device further includes：

Logging modle is used for for each in the multiple images to be recognized, if the handling result of the images to be recognized The object for including for the images to be recognized belongs to known object, then by the handling result of the images to be recognized and the images to be recognized Image information recording in stranger's identification record list；

Add module is used for for each in the multiple images to be recognized, if the handling result of the images to be recognized The object for including for the images to be recognized belongs to unknown object, then the images to be recognized is added in the library of the stranger bottom.

19. a kind of pattern recognition device, including processor and memory, wherein be stored with computer program in the memory Instruction, for executing following steps when the computer program instructions are run by the processor：

Obtain the first image and the second image；

Calculate the similarity score between described first image and second image；And

The similarity score is compared with first threshold and second threshold, to obtain recognition result, wherein if the phase It is greater than or equal to the first threshold like degree scoring, then the recognition result is the described first image object that includes and described the The object that two images include is identical, if the similarity score is less than or equal to the second threshold, the recognition result The object that the object for including for described first image and second image include is different, if the similarity score is more than institute It states second threshold and is less than the first threshold, then the recognition result is the described first image object that includes and described the Whether the object that two images include is identical to belong to fuzzy interval.

20. pattern recognition device as claimed in claim 19, wherein the computer program instructions are run by the processor When be additionally operable to execute following steps：

Establish identification record list；And

By the image information recording of the recognition result, the image information of described first image and second image in the knowledge It Ji Lu not be in list.

21. pattern recognition device as claimed in claim 20, wherein described first image is images to be recognized, described second Image is stored in one of the known image in the library of bottom, and the image information of described first image includes the name of described first image One or more in title, storage location, acquisition time and acquisition position, the image information of second image includes described The title and/or storage location of two images.

22. pattern recognition device as claimed in claim 19, wherein the computer program instructions are run by the processor When for executing following steps：

Export the recognition result.

23. pattern recognition device as claimed in claim 19, wherein the computer program instructions are run by the processor When for executing following steps：

The object that the recognition result is the described first image object that includes and second image includes whether same genus In the case of fuzzy interval, reception is used to indicate the object that described first image includes and the object that second image includes Whether identical artificial markup information；And

The recognition result is updated to the object and described second that described first image includes according to the artificial markup information The object that image includes is identical or different.

24. a kind of image processing apparatus, including processor and memory, wherein be stored with computer program in the memory Instruction, for executing following steps when the computer program instructions are run by the processor：

Obtain the multiple images to be recognized acquired in specific time period；

For each in the multiple images to be recognized,

Using image-recognizing method as described in any one of claims 1 to 3, using the images to be recognized as first figure Picture is handled each known image being stored in the library of bottom as second image, with obtain the images to be recognized with The recognition result of each known image；And

The handling result of the images to be recognized is determined according to the recognition result of the images to be recognized and each known image, In, if the images to be recognized with there is the identical identification of object for indicating to include in the recognition result of each known image As a result, then the handling result of the images to be recognized is that the object that the images to be recognized includes belongs to known object, if this waits knowing Other image and the recognition result of each known image are there is only the different recognition result of the object for including is indicated, then this waits knowing The handling result of other image is that the images to be recognized object that includes belongs to unknown object, if the images to be recognized with it is described every Belong to fuzzy there is no the identical recognition result of object for indicating to include in the recognition result of a known image and in the presence of expression The recognition result in section, then the handling result of the images to be recognized is that the images to be recognized object that includes and the expression belong to Whether object that the corresponding known image of recognition result of fuzzy interval includes is identical to belong to fuzzy interval.

25. image processing apparatus as claimed in claim 24, wherein the computer program instructions are run by the processor When be additionally operable to execute following steps：

To it is in the multiple images to be recognized, all it includes object and described indicate to belong to the recognition result of fuzzy interval The whether identical images to be recognized for belonging to fuzzy interval of the object that corresponding known image includes carries out duplicate removal, not repeated Image collection；And

Export the image information of all images to be recognized in described image set.

26. image processing apparatus as claimed in claim 25, wherein the computer program instructions are run by the processor When be additionally operable to execute following steps：

For each images to be recognized in described image set,

Reception is used to indicate the artificial the markup information whether object that the images to be recognized includes belongs to known object；And

The handling result of the images to be recognized is updated to the object that the images to be recognized includes according to the artificial markup information Belong to known object or unknown object.

27. image processing apparatus as claimed in claim 24, wherein the bottom library is stranger bottom library, the computer journey Sequence instruction is additionally operable to execute following steps when being run by the processor：

For each in the multiple images to be recognized,

If the handling result of the images to be recognized, which is the object that the images to be recognized includes, belongs to known object, this is waited knowing The handling result of other image and the image information recording of the images to be recognized are in stranger's identification record list；

If the handling result of the images to be recognized, which is the object that the images to be recognized includes, belongs to unknown object, this is waited knowing Other image is added in the library of the stranger bottom.

28. a kind of storage medium, stores program instruction on said storage, described program instruction is used to hold at runtime Row following steps：

Obtain the first image and the second image；

Calculate the similarity score between described first image and second image；And

The similarity score is compared with first threshold and second threshold, to obtain recognition result, wherein if the phase It is greater than or equal to the first threshold like degree scoring, then the recognition result is the described first image object that includes and described the The object that two images include is identical, if the similarity score is less than or equal to the second threshold, the recognition result The object that the object for including for described first image and second image include is different, if the similarity score is more than institute It states second threshold and is less than the first threshold, then the recognition result is the described first image object that includes and described the Whether the object that two images include is identical to belong to fuzzy interval.

29. storage medium as claimed in claim 28, wherein described program instruction is used to execute following steps at runtime：

Establish identification record list；And

By the image information recording of the recognition result, the image information of described first image and second image in the knowledge It Ji Lu not be in list.

30. storage medium as claimed in claim 29, wherein described first image is images to be recognized, second image The image information of one of the known image being stored in the library of bottom, described first image includes the title of described first image, deposits Storage space sets, is in acquisition time and acquisition position one or more, and the image information of second image includes second figure The title and/or storage location of picture.

31. storage medium as claimed in claim 28, wherein described program instruction is used to execute following steps at runtime：

Export the recognition result.

32. storage medium as claimed in claim 28, wherein described program instruction is used to execute following steps at runtime：

The object that the recognition result is the described first image object that includes and second image includes whether same genus In the case of fuzzy interval, reception is used to indicate the object that described first image includes and the object that second image includes Whether identical artificial markup information；And

The recognition result is updated to the object and described second that described first image includes according to the artificial markup information The object that image includes is identical or different.

33. a kind of storage medium, stores program instruction on said storage, described program instruction is used to hold at runtime Row following steps：

Obtain the multiple images to be recognized acquired in specific time period；

For each in the multiple images to be recognized,

Using image-recognizing method as described in any one of claims 1 to 3, using the images to be recognized as first figure Picture is handled each known image being stored in the library of bottom as second image, with obtain the images to be recognized with The recognition result of each known image；And

The handling result of the images to be recognized is determined according to the recognition result of the images to be recognized and each known image, In, if the images to be recognized with there is the identical identification of object for indicating to include in the recognition result of each known image As a result, then the handling result of the images to be recognized is that the object that the images to be recognized includes belongs to known object, if this waits knowing Other image and the recognition result of each known image are there is only the different recognition result of the object for including is indicated, then this waits knowing The handling result of other image is that the images to be recognized object that includes belongs to unknown object, if the images to be recognized with it is described every Belong to fuzzy there is no the identical recognition result of object for indicating to include in the recognition result of a known image and in the presence of expression The recognition result in section, then the handling result of the images to be recognized is that the images to be recognized object that includes and the expression belong to Whether object that the corresponding known image of recognition result of fuzzy interval includes is identical to belong to fuzzy interval.

34. storage medium as claimed in claim 32, wherein described program instruction is additionally operable to execute following step at runtime Suddenly：

To it is in the multiple images to be recognized, all it includes object and described indicate to belong to the recognition result of fuzzy interval The whether identical images to be recognized for belonging to fuzzy interval of the object that corresponding known image includes carries out duplicate removal, not repeated Image collection；And

Export the image information of all images to be recognized in described image set.

35. storage medium as claimed in claim 34, wherein described program instruction is additionally operable to execute following step at runtime Suddenly：

For each images to be recognized in described image set,

Reception is used to indicate the artificial the markup information whether object that the images to be recognized includes belongs to known object；And

The handling result of the images to be recognized is updated to the object that the images to be recognized includes according to the artificial markup information Belong to known object or unknown object.

36. storage medium as claimed in claim 34, wherein the bottom library is stranger bottom library, and described program instruction is being transported It is additionally operable to execute following steps when row：

For each in the multiple images to be recognized,

If the handling result of the images to be recognized, which is the object that the images to be recognized includes, belongs to known object, this is waited knowing The handling result of other image and the image information recording of the images to be recognized are in stranger's identification record list；

If the handling result of the images to be recognized, which is the object that the images to be recognized includes, belongs to unknown object, this is waited knowing Other image is added in the library of the stranger bottom.