CN109788227A

CN109788227A - A kind of image processing method and its device, equipment and storage medium

Info

Publication number: CN109788227A
Application number: CN201711116895.5A
Authority: CN
Inventors: 赵娜
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2017-11-13
Filing date: 2017-11-13
Publication date: 2019-05-21
Anticipated expiration: 2037-11-13
Also published as: CN109788227B

Abstract

The embodiment of the present invention provides a kind of image processing method and its device, equipment and storage medium, wherein the described method includes: obtaining fisrt feature information, the fisrt feature information is the characteristic information of target object in reference picture；Image to be adjusted and second feature information are obtained, the second feature information is the characteristic information of target object in the image to be adjusted；Determine the image to be adjusted relative to the first offset information with the reference picture according to the fisrt feature information and the second feature information；The image to be adjusted is adjusted according to first offset information, the image after being adjusted.

Description

A kind of image processing method and its device, equipment and storage medium

Technical field

The present invention relates to Internet technical field more particularly to a kind of image processing methods and its device, equipment and storage Medium.

Background technique

With the development of communication technology and intelligent terminal it is universal, the intelligent terminal for having camera almost becomes people Necessity, people can use that intelligent terminal shoots picture anywhere or anytime, video or terminal carry out video calling, when User using intelligent terminal in good friend's video call process, the picture taken can whip intelligent terminal mobile generation or More or few deflection in order to which good picture display effect needs to carry out video pictures angular adjustment, for example keeps image water Head up display is shown.The correction of video pictures can be carried out by holding third party's ancillary hardware (hand-held holder etc.) at present, but also do not solved Certainly whole issue, there are still following defects: (1) third party's ancillary hardware is at high cost；(2) third party's ancillary hardware is only able to maintain hand Machine is in some special angle；(3) when terminal does not deflect, image can not be corrected.It therefore in the prior art, cannot be complete Support the correction to video.

Summary of the invention

In view of this, an embodiment of the present invention is intended to provide a kind of image processing method and its device, equipment and storage medium, It solves the problems, such as that image flame detection can not be carried out in prior art under the premise of not using third party's ancillary hardware, is regarding Frequency call, video record and expected content of finding a view is not able to maintain yet by third party's ancillary hardware during taking pictures, from And good experience is brought to user.

The technical solution of the embodiment of the present invention is achieved in that

In a first aspect, the embodiment of the present invention provides a kind of image processing method, which comprises

Fisrt feature information is obtained, the fisrt feature information is the characteristic information of target object in reference picture；

Image to be adjusted and second feature information are obtained, the second feature information is target pair in the image to be adjusted The characteristic information of elephant；

According to the fisrt feature information and the second feature information determine the image to be adjusted relative to it is described First offset information of reference picture；

The image to be adjusted is adjusted according to first offset information, the image after being adjusted.

Second aspect, the embodiment of the present invention provide a kind of image processing apparatus, and described device includes: that described device includes: First obtains module, the second acquisition module, the first determining module and adjustment module, in which:

Described first obtains module, and for obtaining fisrt feature information, the fisrt feature information is mesh in reference picture Mark the characteristic information of object；

Described second obtains module, and for obtaining image to be adjusted and second feature information, the second feature information is The characteristic information of target object in the image to be adjusted；

First determining module, for according to the fisrt feature information and the second feature information determine it is described to Image is adjusted relative to the first offset information with the reference picture；

The adjustment module is adjusted for being adjusted according to first offset information to the image to be adjusted Image after whole.

The third aspect, the embodiment of the present invention provide a kind of image processing equipment, and the equipment includes at least: the equipment is extremely It less include: memory, communication bus and processor, in which:

The memory, for storing image processing program；

The communication bus, for realizing the connection communication between processor and memory；

The processor, for executing the image processing program stored in memory, to perform the steps of

Fourth aspect, the embodiment of the present invention provide a kind of computer readable storage medium, the computer-readable storage medium Image processing program is stored in matter, described image processing routine realizes figure provided in an embodiment of the present invention when being executed by processor As the step of processing method.

The embodiment of the present invention provides a kind of image processing method and its device, equipment and storage medium, wherein obtains first Fisrt feature information, the fisrt feature information are the characteristic information of target object in reference picture；Then figure to be adjusted is obtained Picture and second feature information, the second feature information are the characteristic information of target object in the image to be adjusted, and according to The fisrt feature information and the second feature information determine the image to be adjusted relative to the reference picture One offset information；Finally the image to be adjusted is adjusted according to first offset information, the image after being adjusted. It is not able to maintain by third party's ancillary hardware in expected find a view in video calling, video record and during take pictures so Hold, to bring good experience to user.

Detailed description of the invention

Figure 1A is the schematic diagram of the network architecture of the embodiment of the present invention；

Figure 1B is the schematic diagram of the network architecture of the embodiment of the present invention；

Fig. 2 is the implementation process schematic diagram of image processing method of the embodiment of the present invention；

Fig. 3 is the implementation process schematic diagram of image processing method of the embodiment of the present invention；

Fig. 4 is the implementation process schematic diagram of image processing method of the embodiment of the present invention；

Fig. 5 A is the terminal interface schematic diagram that marker of the embodiment of the present invention shifts；

Fig. 5 B is the schematic diagram of the direction of rotation of image of the embodiment of the present invention；

Fig. 5 C is the schematic diagram of the updated image of the embodiment of the present invention；

Fig. 5 D is the schematic diagram of the implementation environment of image processing method provided in an embodiment of the present invention；

Fig. 6 is the composed structure schematic diagram of image processing apparatus of the embodiment of the present invention；

Fig. 7 is the composed structure schematic diagram of image procossing of the embodiment of the present invention.

Specific embodiment

Following will be combined with the drawings in the embodiments of the present invention, is described in detail to the specific technical solution of invention.

Figure 1A is the schematic diagram of the network architecture of the embodiment of the present invention, and as shown in Figure 1A, which includes sender visitor Family end 11, recipient's client 12 and server 13, wherein all pacify in described sender client 11 and recipient's client Equipped with a instant messaging application, server 13 is that corresponding server is applied in the instant messaging.

It is connected firstly, sender's client 11 establishes video calling by server 13 with recipient's client 12, then If sender's client 11 to server initiate image processing requests, server 13 by sender's client 11 get into The reference picture of row image flame detection and the characteristic information of reference picture, next, if server 13 monitors sender visitor When the image that family end is sent needs to correct, the image is corrected according to reference picture, the image after being corrected, and will rectify Image after just is sent to sender's client 11 and recipient's client 12.

Image processing method provided in an embodiment of the present invention can be applied in implementation environment as shown in Figure 1B, Figure 1B For another implementation environment schematic diagram of image processing method provided in an embodiment of the present invention, as shown in Figure 1B, the implementation environment Including client 13 and server 14, wherein being equipped in client 13 can be taken pictures or the application of recorded video, clothes Business device 14 and client 13 establish communication connection.When client 13 is taken pictures or when video record starts image flame detection function Can, a selected sub-picture is first had to as reference picture and selects the target object in reference picture as marker, and The weight parameter of terminal when acquiring reference picture and the image data of marker are sent to server 14, server 14 is according to this The posture or marker of the image data of weight parameter and marker monitoring terminal whether shift when terminal posture or When marker shifts, server 14 according to the image data of weight parameter and/or marker, to the image to shift into Row correction, and the image after correction is sent to client 13.

In other embodiments, can also be client take pictures or recorded video before, itself get reference picture with And weight parameter when acquisition reference picture, and according to weight parameter and reference picture to subsequent acquisition to image rectify Just, and the image after correction is exported.

The schematic diagram in conjunction with shown in Figure 1A and Figure 1B, below to image processing method and image processing apparatus, equipment it is each Embodiment is illustrated.

The embodiment of the present invention provides a kind of image processing method, and Fig. 2 is the realization of image processing method of the embodiment of the present invention Flow diagram, as shown in Fig. 2, the described method comprises the following steps:

Step S201 obtains fisrt feature information.

In the present embodiment, the step S201 can be by server realization.The fisrt feature information is reference The characteristic information of target object in image.Reference picture is acquired image when terminal is in the first posture.First appearance State can be the posture of two sides and horizontal plane of terminal, that is, usually described vertical screen posture；First posture It can be the posture that two sides of terminal are parallel to the horizontal plane, that is, usually described transverse screen posture.

Target object can be what user selected, can also be that server is selected according to the image data in reference picture 's.For example, user does not have selection target image, server can determine reference picture according to the image data in reference picture In include a face, then face will be determined as target object.Alternatively, server can will be located in reference picture The image of central area is set as target object.

Characteristic information, which refers to, can most reflect features, basic, essential information.Obtain fisrt feature information Process namely the process of feature extraction.Feature extraction refers to determining the point of each image using computer extraction image information Whether a characteristics of image is belonged to.Feature extraction the result is that the point on image is divided into different subsets, these subsets are often Belong to isolated point, continuous curve or continuous region.

If target object is a face, can wrap in the fisrt feature information got by feature extraction Include: eyes, eyebrow, nose or the corresponding characteristic point of mouth and characteristic point position in the picture；If target object is One building, then the fisrt feature information got by feature extraction can be each top in the profile of building Point and vertex institute position in the picture.

In this example, it is assumed that target object is face, included at least in fisrt feature information: left eye corresponding first The position of characteristic point and fisrt feature point, the corresponding second feature point of right eye and second feature point position.With reference to figure As in, fisrt feature point is parallel or vertical with image border with the line of second feature point.

Step S202 obtains image to be adjusted and second feature information.

Here, the step S202 can be by server realization.The second feature information is the figure to be adjusted The characteristic information of target object as in.Also, the characteristic point in second feature information is and the characteristic point in fisrt feature information It is corresponding.

Step S203 determines that the image to be adjusted is opposite according to the fisrt feature information and the second feature information In the first offset information with the reference picture.

Here, the step S203 can be by server realization.It may include first in first offset information Deviation angle and the first offset direction, such as the first offset information can be and deviate 30 degree to the left, deviate 50 degree etc. to the right.

Step S204 is adjusted the image to be adjusted according to first offset information, the figure after being adjusted Picture.

Here, the step S204 can be by server realization.According to first offset information to described wait adjust Whole image is adjusted can be realized by least two modes: first is that according to the first deviation angle and the first offset direction pair The image to be adjusted is rotated accordingly；Second is that determining cutting portion according to the first deviation angle and the first offset direction Point, image to be adjusted is cut.

In other embodiments, after the step S204, the method also includes: the image adjusted is sent out Give second terminal and first terminal.Wherein, the first terminal is initiator's client in video calling, the second terminal For recipient's client.In some cases, the image adjusted only can also be sent to second terminal, without It is sent to first terminal.

In other embodiments, step S201 obtains fisrt feature information and may further include:

Step S2011 obtains the first image data of the target object in the reference picture；Wherein, described with reference to figure As being that the terminal is in collected when the first posture；

Step S2012 determines the type of the target object according to the first image data；

Step S2013 determines that the characteristic information for extracting the target object is used according to the type of the target object Preset algorithm；

Step S2014, according to the first image data and the preset algorithm, extract the target object first is special Reference breath.

In the embodiment where the step S2011 to step S2014, the reference can be obtained by two ways First image data of the target object in image:

First way: server obtains the first image data by terminal, it is possible to further pass through following step It is rapid to realize:

Step 31, the image processing requests that first terminal is sent are received；

Step 32, based on described image processing request, Xiang Suoshu first terminal sends the first request message；

Here, first request message is in for the first image data of request target object and the terminal The first weight parameter when the first posture.

Step 33, the first response message that the terminal is sent is received, wherein the mesh is carried in the response message The first image data and the terminal for marking object are in the first weight parameter when the first posture；

Step 34, first response message is parsed, the first image data of target object in the reference picture is obtained.

The second way: server obtains the first image data by the image itself stored, it is possible to further It is realized by following steps:

Step 41, according to image to be adjusted, reference picture is obtained from itself pre-stored image data；

Step 42, according to the second feature information and the reference picture, the target pair in the reference picture is obtained The first image data of elephant.

It should be noted that being to first carry out step if obtaining the first image data using the second way S202, then execute step S201's.That is, in other embodiments, server gets piece image first, and to this Image is judged determine whether the image can be as reference picture, if the image can be used as reference picture, will The image is stored, and when server gets another piece image, whether the interpretation image needs to adjust, if necessary to adjust It is whole, corresponding with image reference picture is obtained according to the image, according further to the of reference picture acquisition target object One image data.

In other embodiments, before the step S204, the method also includes:

Step 61, the 4th image for meeting the second preset condition is obtained according to the image to be adjusted；

Step 62, the third feature information of target object in the 4th image is obtained, and is believed according to the third feature Breath and the fisrt feature information, determine third offset information of the 4th image relative to the reference picture；

Step 63, if the third feature information and the fisrt feature information meet third preset condition, according to institute It states the first offset information to be adjusted the image to be adjusted, the image after being adjusted.

In the embodiment where step 61 to step 63, second preset condition be can be including the target object Image, that is to say, that in the 4th image include at least target object.And then extract the of target object in the 4th image Three characteristic informations.It is identical with the fisrt feature information that third preset condition can be third feature information, can also be basis Third feature information and fisrt feature information determine less than one preset threshold of the deviation angle in third offset information.

Before the step 63, the method also includes: judge the third feature information and fisrt feature letter Whether breath meets third preset condition, carries out the step of this step judge here actually in order to by storing in server Four image of big flow control carry out big data analysis to confirm the use habit of user, if third feature information and described first special Reference breath meet the third preset condition just illustrate user video calling or take pictures, video record when habit to image into Row correction, is at this time adjusted the image to be adjusted according to first offset information, the image after being adjusted；If Third feature information and fisrt feature information are unsatisfactory for third preset condition and indicate that user is not relevant for whether image occurs partially It moves, image is not corrected at this time.

In image processing method provided in an embodiment of the present invention, fisrt feature information, the fisrt feature are obtained first Information is the characteristic information of target object in reference picture；Then image to be adjusted and second feature information are obtained, described second Characteristic information is the characteristic information of target object in the image to be adjusted, and according to the fisrt feature information and described second Characteristic information determines the image to be adjusted relative to the first offset information with the reference picture；Finally according to described first Offset information is adjusted the image to be adjusted, the image after being adjusted.So video calling, video record and It is not able to maintain expected content of finding a view by third party's ancillary hardware during taking pictures, to bring good body to user It tests.

The embodiment of the present invention provides a kind of image processing method again, and the method can be applied to first terminal and second eventually End carries out the scene of video calling, can also be applied to first terminal and be taken pictures or the scene of recorded video.Fig. 3 is this hair The implementation process schematic diagram of bright embodiment image processing method, as shown in figure 3, the described method comprises the following steps:

Step S301, if first terminal receives the first operational order of starting image flame detection, output acquisition reference picture And the prompt information of selection target object.

Here, during realization, first terminal can provide the switch button of starting image flame detection function, or provide Start the option of image flame detection function.First operational order can be the switch that user clicks starting image flame detection function Button and the operational order triggered can also be the operational order that user selects the option of starting image flame detection function and triggers.

When receiving the first operational order of starting image flame detection, the first terminal can prompt user's acquisition with reference to figure Picture, and selection target object.Such as first terminal can export " please keep terminal in horizontal attitude, acquisition reference picture is simultaneously The prompt information of selection target object ".

In other embodiments, after exporting the prompt information, first terminal can detect the weight parameter of itself, such as top The parameter of spiral shell instrument, and judge the posture of first terminal.If detect be individually not horizontally or vertically direction gripping appearance State can export and user's confirmation is asked to use the prompt information of the Posture acquisition reference picture, be somebody's turn to do if receiving user and confirming to use The operational order of Posture acquisition reference picture acquired image will be determined as reference picture at this time.

Step S302 will be collected currently if the first terminal gets the second operational order of selection target object Image be determined as reference picture.

Here, in the present embodiment, after first terminal collects image, detect whether to receive user where image The touch control operation or clicking operation in region, if receiving user in the touch control operation or clicking operation of image region, The second operational order for thinking to receive selection target object, then current acquired image can be determined as with reference to figure Picture.

Step S303, the first terminal are based on second operational order, obtain the first image data of target object.

Here, the reference picture is that the terminal is in collected when the first posture.When the first terminal obtains When to the second operational order, the position that can be acted on first according to the operation for obtaining target object carries out scratching figure, obtains target First image data of object.It, can position to be acted on according to the operation for obtaining target object in addition, in other embodiments It is set to center, the image data in the region apart from the default pixel in the center is determined as to the first picture number of target object According to.

Step S304 determines the type of the target object according to the first image data.

Here, the type of target object can include but is not limited to: face, building, scenery.According to the first image Data determine that the type of the target object during realization, can determine target object according to the first image data Profile, and then determine according to the profile of target object the type of target object.

Step S305 determines that the characteristic information for extracting the target object is used according to the type of the target object Preset algorithm.

Here, it can be previously stored with the type of target object in the first terminal and the corresponding of the algorithm of use is closed System.Such as the type of target object be face when, using the first preset algorithm；When the type of target object is building, use Second preset algorithm.

Step S306, according to the first image data and the preset algorithm, extract the target object first is special Reference breath.

It here, include at least two characteristic points and each characteristic point in fisrt feature information at reference picture corresponding end Hold the position on screen.In this example, it is assumed that target object is face, fisrt feature point is pupil of left eye, second feature Point is pupil of right eye.

Step S307, the first terminal obtain image to be adjusted and second feature information.

Here, the second feature information is the characteristic information of target object in the image to be adjusted.Described second is special The corresponding characteristic point of characteristic point in fisrt feature information is included at least in reference breath.In the present embodiment, the second feature The position of third feature point, the position of fourth feature point and third feature point and fourth feature point is included at least in information, In, the third feature point is the corresponding characteristic point of the fisrt feature point, and fourth feature point is the second feature point phase Corresponding characteristic point.That is, third feature point is the pupil of left eye in image to be adjusted, fourth feature point is figure to be adjusted Pupil of right eye as in.

Step S308 determines the fisrt feature according to the position of the position of the fisrt feature point and second feature point First relative positional relationship of point and the second feature point.

Here, the principle of straight line where can determining two points according to known two points, according to the position of fisrt feature point The expression formula of straight line where can determining fisrt feature point and second feature point with the position of second feature point, the expression formula is also It is the first relative positional relationship of fisrt feature point and second feature point.

Step S309 determines the third feature according to the position of the position of third feature point and fourth feature point Second relative positional relationship of point and the fourth feature point.

Here, similar to the first relative positional relationship, it can be according to the position of third feature point and the position of fourth feature point Set the second relative positional relationship of determining third feature point and fourth feature point.

Step S310 is determined described wait adjust according to first relative positional relationship and second relative positional relationship First offset information of the whole image relative to the reference picture.

Here, first offset information includes at least first deviation angle of the image to be adjusted relative to reference picture.

Step S310, can be by according to first relative positional relationship and described second opposite during realization Straight line where straight line and third feature point where positional relationship calculates fisrt feature point and second feature point and fourth feature point Angle, to determine the first offset information according to the angle.

Step S311 is adjusted the image to be adjusted according to first offset information, the figure after being adjusted Picture.

Here, the step S311 can at least be realized by three kinds of modes: first way, by the image to be adjusted It is rotated, the image after being adjusted.The second way cuts the image to be adjusted, the figure after being adjusted Picture.The third mode, be adjusted by big data analysis after image.

In other embodiments, if it is in the scene that first terminal and second terminal carry out video calling, described After step S311, the method also includes: the image adjusted is sent to second terminal by the first terminal.

In other embodiments, the method also includes:

Step 51, first terminal is obtained from the first weight parameter being in when the first posture.

Here, first weight parameter can by first terminal when the first posture gyro sensor parameter come It obtains.

Step 52, if the first terminal is in the second posture when acquiring image to be adjusted, the first terminal is obtained Itself is in the second weight parameter when the second posture.

Step 53, the first terminal determines first terminal according to first weight parameter and second weight parameter The second offset information when the first posture is in relative to the terminal when in the second posture.

Step 54, the first terminal judges whether first offset information and second offset information meet first Preset condition.

Here, it is identical as second offset information to can be the first offset information for first preset condition, and described One preset condition can also be the second deviation angle in the first deviation angle and the second offset information in the first offset information Difference absolute value in values.

If first offset information and second offset information meet the first preset condition, illustrate figure to be adjusted As the offset occurred relative to reference picture is almost to enter step 55 at this time due to caused by the posture of terminal；Such as First offset information described in fruit and second offset information are unsatisfactory for the first preset condition, then illustrate image to be adjusted relative to The offset that reference picture occurs not exclusively is to enter step 56 at this time due to caused by the posture of terminal.

Step 55, the first terminal determines first terminal according to second weight parameter and first weight parameter The adjustment direction and adjustment amplitude of posture.

Step 56, the first terminal is true according to second deviation angle, the first deviation angle and the first weight parameter Determine third weight parameter.

Step 57, the first terminal determines terminal posture according to the third weight parameter and second weight parameter Adjustment direction and adjustment amplitude.

Step 58, the first terminal generates the prompt information of simultaneously output adjustment first terminal posture, wherein the prompt It include the adjustment direction and the adjustment amplitude in information.

It, can be according to the adjustment direction in prompt information after user receives the prompt information of adjustment first terminal posture It is adjusted with posture of the adjustment amplitude to first terminal, to guarantee that the acquired image of first terminal is that there is no inclined The image of shifting.

In image processing method provided in an embodiment of the present invention, if first terminal receives starting image flame detection first First operational order, the prompt information of output acquisition reference picture and selection target object；If the first terminal gets choosing Current acquired image is determined as reference picture by the second operational order for selecting target object；The first terminal is based on institute The second operational order is stated, the first image data of target object is obtained；And according to the first image data, the target is determined Then the type of object according to the type of the target object, further determines that the characteristic information for extracting the target object is adopted Preset algorithm；And then according to the first image data and the preset algorithm, extract the target object first is special Reference breath；Then first terminal obtains image to be adjusted and second feature information；And according to the position of the fisrt feature point and The position of second feature point determines the first relative positional relationship of the fisrt feature point and second feature point；Further according to The position of third feature point and the position of fourth feature point determine the of the third feature point and fourth feature point Two relative positional relationships；Then according to first relative positional relationship and second relative positional relationship, so that it is determined that institute State first offset information of the image to be adjusted relative to the reference picture；Finally according to first offset information to it is described to Adjustment image is adjusted, the image after being adjusted, and so, it is possible to shoot photo or recorded video in first terminal, or It is determined when person carries out video calling with second terminal by reference to the positional relationship of the characteristic point in image and image to be adjusted Offset information between image and reference picture to be adjusted, to be adjusted to image to be adjusted, and then user obtain One terminal acquired image is that there is no offsets relative to reference picture, improves user experience.

The embodiment of the present invention provides a kind of image processing method again, and the method can be using whole with first terminal and second End carries out the scene of video calling.It the described method comprises the following steps:

Step 71, second terminal receives the first image that first terminal is sent.

Step 72, the second terminal judges whether itself is stored with reference picture.

Here, if the second terminal itself is stored with reference picture, 75 are entered step；If itself is not stored with Reference picture enters step 73.

Step 73, judge whether can be using the first image as reference picture for the second terminal.

Here, if it is possible to using the first image as reference picture, enter step 74；If cannot be by described first Image enters step 75 as reference picture.

Judge whether that the first image an energy can be preset as reference picture during realization Enough as the preset condition of reference picture, by judging whether the first image meets the preset condition determine whether can be by institute The first image is stated as reference picture.

For example, in the first image include face, then in available first image two eyes of face line, should The line that preset condition can be two eyes is parallel with image border.If the line and image of two eyes in the first image Sides aligned parallel, at this time can be using the first image as reference picture；If the line of two eyes and image side in the first image Edge is not parallel, then just cannot be using the first image as reference picture.

Certainly, sometimes parallel such condition is excessively harsh, then can set preset condition to the company of two eyes Angle between line and image border is less than preset value, which can be 5 degree, 3 degree etc..If two in the first image Angle between the line and image border of eyes is less than preset value, and indicating that can be using the first image as reference picture；Such as Angle in the first image of fruit between the line and image border of two eyes is not less than preset value, and indicating that cannot be by first Image is as reference picture.

Step 74, the first image is determined as reference picture by the second terminal, and saves the reference picture.

Here, if the first image can be used as reference picture, illustrate that the first image does not need to correct, go at this time Step 71 continues to next first image of first terminal transmission.

Step 75, the second terminal obtains reference picture.

Here, step 75 can be realized by two ways: first way: if being stored with reference picture in itself, Reference picture is so obtained directly from the storage medium of itself.The second way: if not being stored with ginseng in itself Image is examined, and the first image cannot function as reference picture again, needs to obtain reference picture by first terminal at this time.

When obtaining reference picture by the second way, the second terminal needs to send an acquisition to first terminal The request message of reference picture after first terminal receives the request message, prompts user to acquire reference picture, and will acquisition To reference picture be carried in response message and be sent to second terminal.

Step 76, the second terminal obtains fisrt feature information.

Here, the fisrt feature information is the characteristic information of target object in the reference picture.

Before the step 76, the method also includes: the second terminal obtains the target in the reference picture Object.During realization, obtaining the target object in the reference picture, to can be second terminal automatic according to reference picture It obtains, such as the image for being located at preset middle section in reference picture is determined as target object by second terminal；It can be with It is to be determined by obtaining user based on the operational order that touch control operation of the reference picture in the second terminal triggers 's.

In the fisrt feature information include at least fisrt feature point, second feature point and fisrt feature point position and The position of second feature point.

Step 77, the first image is determined as image to be adjusted by the second terminal, obtains the second of the first image Characteristic information.

Here, third feature point, fourth feature point and third feature point are included at least in the second feature information The position of position and fourth feature point, and third feature point is corresponding with fisrt feature point, fourth feature point and second feature Point is corresponding.

Step 78, the second terminal determines the image phase to be adjusted according to second feature information and fisrt feature information For the first offset information with the reference picture.

Step 79, the second terminal is adjusted the image to be adjusted according to first offset information, obtains Image adjusted.

Here, the step 79 can at least be realized by two ways:

First way rotates the image to be adjusted, the image after being adjusted.It is possible to further logical Cross following steps realization:

Step 79a1 determines the first adjustment angle and the first adjustment direction according to first offset information；

It here, may include the first deviation angle and the first offset direction, the first adjustment angle etc. in the first offset information In the first deviation angle, the first adjustment direction is the opposite direction of the first offset direction.Such as first offset information be clockwise partially 20 degree are moved, then the first adjustment angle is 20 degree, the first adjustment direction is counterclockwise.

The image to be adjusted is rotated the first adjustment angle according to the first adjustment direction, obtained by step 79a2 To image adjusted.

The second way cuts the image to be adjusted, the image after being adjusted.It is possible to further logical Cross following steps realization:

Step 79b1 is determined according to first deflection angle and the image to be adjusted and is cut target image；

Step 79b2 removes the cutting target image, the image after being adjusted from the image to be adjusted.

In other embodiments, after can also being adjusted by big data analysis after getting image to be adjusted Image.It can not need to obtain reference picture at this time, but what is stored in acquisition server largely includes in image to be adjusted Target object the second image, and the fourth feature information of target object in the second image is obtained, then to fourth feature information Carry out clustering, determine one can as the fisrt feature information of reference standard, thus according to fisrt feature information with The second feature information of target object determines the first offset information in image to be adjusted, treats tune further according to the first offset information Whole image is adjusted, the image after being adjusted.

Step 80, the second terminal exports the image adjusted.

It should be noted that can refer in the present embodiment with the explanation of same steps in other embodiments or concept other Description in embodiment.

It is provided in a kind of image processing method in the embodiment of the present invention, second terminal receives the first figure that first terminal is sent Picture simultaneously judges whether itself is stored with reference picture；If itself is not stored with reference picture, judge that the first image is It is no to can be used as reference picture, if so, the first image can be used as reference picture, next first image is continued to, If it is not, then obtaining reference picture.Then second terminal obtains the fisrt feature information of target object in reference picture, by first Image obtains the second feature information of the first image further according to second feature information and fisrt feature information as image to be adjusted The image to be adjusted is determined relative to the first offset information with the reference picture, then according to first offset information The image to be adjusted is adjusted, the image after being adjusted, the second last terminal exports the image adjusted.

The embodiment of the present invention provides a kind of image processing method, and Fig. 4 is the realization of image processing method of the embodiment of the present invention Flow diagram, as shown in figure 4, the described method comprises the following steps:

Step S401, the machine A and good friend B carry out video conversation.

Step S402, the machine A client send to server and send correction request.

Here, during video conversation, the correction corrected to current video can be initiated by button or option Request.

Step S403, server receive correction request.

Here, after server receives correction request, the setting of correction data is begun preparing.

Step S404, the machine A user end to server report sensor parameters.

Here, server issues the request message of mobile client acquisition standard value, and the machine A client carries out standard value and adopts Collection:

Reference direction: the machine A Client-Prompt local user A selection standard direction extracts and records each of terminal this moment The detail parameters of class gravity sensor (gyroscope etc.), and by the parameter upload server of gravity sensor.

Step S405, the machine A user end to server report marker parameter.

Here, the machine A client chooses object (such as recognition of face) and does marker, and marker parameter is uploaded and is taken Business device.

Step S406, server carry out data analysis to the sensor parameters and marker parameter received.

Here, server carries out data analysis: the server data of upload passes through artificial intelligence (Artificial Intelligence, AI) identification, big data analysis, marker parameter and sensor parameters are parsed, more new algorithm, and Storing and resolving is as a result, standard value setting is completed.

Step S407, server monitor the marker of sensor parameters and area of finding a view；

Here, after the completion of the setting of client standard value, server monitors the marker of sensor parameters and area of finding a view.

Step S408, server judge whether sensor parameters change.

Whether step S409, server judgement symbol object parameter change.

Here, the machine A client and good friend's B client continue video conversation, the rotation of the machine A client or area of finding a view Interior object rotates, and server can listen to change.

Step S410, server carries out AI analysis and is supplied to the image that user suggests adjustment, or acquires to the machine A client To image rotated, cut to obtain after image.

Here, right 1. if listening to sensor parameters change: the parameter of server force revert primary standard gyroscope value Area's data of finding a view are cut, and image issues the machine customer end A and good friend's customer end B after cutting；

2. if listening to marker parameter modification: marker is face or other objects, root by the result that AI is identified According to the angle that object deflects, rotation update is carried out to image；

In embodiments of the present invention, it is assumed that the reference direction of terminal is vertical screen direction, and marker is face, and two in face The line of a human eye is parallel with image border.As shown in Figure 5A, there is no the marks that variation, user are chosen for the posture of terminal Will object is face shown in 511.It, can be according to the parameter of the marker after change after server listens to marker parameter modification Direction of rotation and rotation angle are determined with the marker parameter of standard, as shown in Figure 5 B, need to carry out 521 institutes to image at this time That shows rotates clockwise.As shown in Figure 5 C, there is no inclined relative to the marker of standard for marker 531 in updated image Move namely updated image in marker 531 line of two human eyes with image border be parallel.

3. if listening to sensor parameters+marker parameter modification: server is supplied to user by big data analysis It is recommended that the image of adjustment, also provides customized three-dimensional (3Dimension, 3D) the gyroscope button of user, user is manually adjustable to be taken Scenic spot image rotation, and report server.

Step S411, after image update, after image issues the machine customer end A and good friend's customer end B.

In the present embodiment by taking two clients carry out video calling as an example, instant messaging apply by under message window into The customized camera of row is recorded short-sighted frequency, video record is carried out using camera applications, is taken pictures, using figure provided in this embodiment As processing method carries out image flame detection.

Fig. 5 D is the implementation environment schematic diagram of image procossing of the embodiment of the present invention, as described in Fig. 5 D, in the implementation environment It include: the machine A client 541, good friend B client 542, server 543.

Apply firstly, sender's client 541 is sent to server, during video conversation, can by button or Option initiates the application corrected to current video, and server receives request, begins preparing correction data setting；Then, it takes Business device 543 notifies sender's client 541 to carry out standard value acquisition, and 541 selection standard direction of sender's client is extracted and remembered Record terminal various kinds of sensors (gyroscope etc.) detail parameters this moment, upload server；Sender's client 541 chooses object (such as face, building) does marker, upload server；Server is by AI identification, big data analysis, to marker and biography Sensor parameter is parsed, and storing and resolving is as a result, standard value setting is completed；Hereafter server monitors sender's client 541 Sensor parameters and the marker in area of finding a view；If object rotates in the rotation of sender's client 541 or area of finding a view, take After business device listens to change, the Adjusted Option to the image after offset is determined according to standard value, and carry out to the image after offset Adjustment, obtains updated image, and then updated image is sent to transmission hair client 541 and received by server 543 Hold client 542.

The embodiment of the invention provides identified based on AI, the combination of big data analysis technology, mobile terminal video is rectified Positive processing realizes the scheme of audio-video, the expected content of finding a view of video record holding: first during video record, setting Sensor standard value: the deflection angle of record standard video obtains the gyroscope parameters of mobile device this moment；And in area of finding a view It is interior, landmark or face are chosen, is identified in conjunction with AI, data analysis is carried out to the mark, extracts characteristic value, and record spy Value indicative deflection；When mobile device, which carries out mobile or object, to be rotated, by sensor, characteristic calculating, to finding a view Picture material is deflected, is cut in area, and data are transmitted after carrying out pressure correction to image data.So pass through mobile phone gyroscope Equal sensors and the mode for setting marker are cut and are overturn to viewfinder range in conjunction with AI identification and big data analysis, real Now the pressure of video is corrected, realizes the occasion in audio-video dialogue, recorded video, expected angle is kept to video, is improved The image quality of video.

The embodiment of the present invention provides a kind of image processing apparatus, and Fig. 6 is the composition of image processing apparatus of the embodiment of the present invention Structural schematic diagram, as shown in fig. 6, described device 600 includes: that the first acquisition module 601, second acquisition module 602, first determines Module 603 and adjustment module 604, in which:

Described first obtains module 601, and for obtaining fisrt feature information, the fisrt feature information is in reference picture The characteristic information of target object；

Described second obtains module 602, for obtaining image to be adjusted and second feature information, the second feature information For the characteristic information of target object in the image to be adjusted；

First determining module 603, for determining institute according to the fisrt feature information and the second feature information Image to be adjusted is stated relative to the first offset information with the reference picture；

The adjustment module 604 is obtained for being adjusted according to first offset information to the image to be adjusted Image adjusted.

In other embodiments, the first acquisition module 601 further comprises:

First acquisition unit, for obtaining the first image data of the target object in the reference picture；Wherein, described Reference picture is that the terminal is in collected when the first posture；

First determination unit, for determining the type of the target object according to the first image data；

Second determination unit determines the feature letter for extracting the target object for the type according to the target object Preset algorithm used by ceasing；

Extraction unit, for according to the first image data and the preset algorithm, extracting the of the target object One characteristic information.

In other embodiments, described device further include:

First receiving module, for receiving the image processing requests of terminal transmission；

First sending module, for being requested based on described image processing, Xiang Suoshu terminal sends the first request message, In, when first request message is in the first posture for the first image data of request target object and the terminal The first weight parameter.

In other embodiments, the first acquisition unit, is also used to: receiving the first response that the terminal is sent and disappears Breath, wherein the first image data and the terminal that the target object is carried in the response message are in the first posture When the first weight parameter；First response message is parsed, the first picture number of target object in the reference picture is obtained According to.

In other embodiments, the first acquisition unit, is also used to: according to image to be adjusted, depositing in advance from itself Reference picture is obtained in the image data of storage；According to the second feature information and the reference picture, obtain described with reference to figure First image data of the target object as in.

In other embodiments, including at least the position of fisrt feature point and second feature point in the fisrt feature information Position, the position of the third feature point and the position of fourth feature point are included at least in second feature information, accordingly Ground, first determining module 603, comprising:

Third determination unit, described in determining according to the position of the fisrt feature point and the position of second feature point First relative positional relationship of fisrt feature point and second feature point；

4th determination unit, described in determining according to the position of third feature point and the position of fourth feature point Second relative positional relationship of third feature point and fourth feature point, wherein the third feature point is described first special The corresponding characteristic point of sign point, fourth feature point are the corresponding characteristic point of the second feature point；

5th determination unit, for determining according to first relative positional relationship and second relative positional relationship First offset information of the image to be adjusted relative to the reference picture.

In other embodiments, described device further include:

Third obtains module, for obtaining the first weight parameter when the terminal is in the first posture；

4th obtains module, is in the second posture when if acquiring image to be adjusted for the terminal, obtains the end The second weight parameter when end is in the second posture；

Second determining module, for determining that terminal is in the according to first weight parameter and second weight parameter The second offset information when the first posture is in relative to the terminal when two postures.

In other embodiments, the adjustment module 604 includes:

6th determination unit, for determining the first adjustment angle and the first adjustment direction according to first offset information；

Rotary unit, for the image to be adjusted to be rotated the first adjustment angle according to the first adjustment direction Degree, the image after being adjusted.

In other embodiments, the first deviation angle, accordingly, the adjustment are included at least in first offset information Module 604 includes:

7th determination unit, for determining and cutting target figure according to first deflection angle and the image to be adjusted Picture；

Cut unit, the image for removing the cutting target image from the image to be adjusted, after being adjusted.

In other embodiments, described device further include:

5th obtains module, for obtaining the 4th image for meeting the second preset condition according to the image to be adjusted；

6th obtains module, for obtaining the third feature information of target object in the 4th image, and according to described Third feature information and the fisrt feature information determine third offset letter of the 4th image relative to the reference picture Breath；

Judgment module presets item for whether third feature information described in interpretation and the fisrt feature information to meet third Part, wherein if the third feature information and the fisrt feature information meet third preset condition, partially according to described first It moves information to be adjusted the image to be adjusted, the image after being adjusted.

In other embodiments, second offset information at least further includes the second deviation angle, described device further include:

Third determining module, if being unsatisfactory for described first for first offset information and second offset information Preset condition determines third weight parameter according to second deviation angle, the first deviation angle and the first weight parameter；

4th determining module, for determining terminal posture according to the third weight parameter and second weight parameter Adjustment direction；

Output module, for generating the prompt information of simultaneously output adjustment terminal posture, wherein include in the prompt information The adjustment direction.

The description of apparatus above embodiment, be with the description of above method embodiment it is similar, have same embodiment of the method Similar beneficial effect.For undisclosed technical detail in apparatus of the present invention embodiment, embodiment of the present invention method is please referred to Description and understand.

The embodiment of the present invention provides a kind of image processing equipment, and Fig. 7 is the composed structure of image procossing of the embodiment of the present invention Schematic diagram, as shown in fig. 7, the equipment 700 at least one processor 701, at least one communication bus 702, user interface 703, at least one external communication interface 704 and memory 705.Wherein, communication bus 702 is arranged for carrying out between these components Connection communication.Wherein, user interface 703 may include display screen, and external communication interface 704 may include that the wired of standard connects Mouth and wireless interface.The wherein processor 701, is configured that

In other embodiments, the acquisition fisrt feature information, comprising:

Obtain the first image data of the target object in the reference picture；Wherein, the reference picture is the end It is collected when end is in the first posture；

According to the first image data, the type of the target object is determined；

According to the type of the target object, determines and imputed in advance used by the characteristic information for extracting the target object Method；

According to the first image data and the preset algorithm, the fisrt feature information of the target object is extracted.

In other embodiments, the step of obtaining the first image data of the target object in the reference picture it Before, the processor 701 is additionally configured to:

Receive the image processing requests that terminal is sent；

It is requested based on described image processing, Xiang Suoshu terminal sends the first request message, wherein first request message The first weight parameter when the first posture is in for the first image data of request target object and the terminal.

In the embodiment of the present invention, if realizing above-mentioned image processing method in the form of software function module, and conduct Independent product when selling or using, also can store in a computer readable storage medium.Based on this understanding, Substantially the part that contributes to existing technology can be with the shape of software product in other words for the technical solution of the embodiment of the present invention Formula embodies, which is stored in a storage medium, including some instructions are used so that a calculating Machine equipment (can be personal computer, server or network equipment etc.) executes each embodiment the method for the present invention It is all or part of.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (ROM, Read Only Memory), the various media that can store program code such as magnetic or disk.In this way, the embodiment of the present invention is not limited to appoint What specific hardware and software combines.

The embodiment of the present invention provides a kind of computer storage medium, and being stored with computer in the computer storage medium can It executes instruction, the computer executable instructions are for executing image processing method provided in an embodiment of the present invention.

It should be understood that " one embodiment " or " embodiment " that specification is mentioned in the whole text mean it is related with embodiment A particular feature, structure, or characteristic is included at least one embodiment of the present invention.Therefore, occur everywhere in the whole instruction " in one embodiment " or " in one embodiment " not necessarily refer to identical embodiment.It should be understood that of the invention various In embodiment, magnitude of the sequence numbers of the above procedures are not meant that the order of the execution order, and the execution of each process sequence should be with Its function and internal logic determine that the implementation process of the embodiments of the invention shall not be constituted with any limitation.

It should be noted that, in this document, the terms "include", "comprise" or its any other variant are intended to non-row His property includes, in the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that wrapping Include in process, method, article or the device of the element that there is also other identical elements.

In several embodiments provided herein, it should be understood that disclosed device and method can pass through it Its mode is realized.Apparatus embodiments described above are merely indicative, for example, the division of the unit, only A kind of logical function partition, there may be another division manner in actual implementation, such as: multiple units or components can combine, or It is desirably integrated into another system, or some features can be ignored or not executed.

Above-mentioned unit as illustrated by the separation member, which can be or may not be, to be physically separated, aobvious as unit The component shown can be or may not be physical unit；Both it can be located in one place, and may be distributed over multiple network lists In member；Some or all of units can be selected to achieve the purpose of the solution of this embodiment according to the actual needs.

In addition, each functional unit in various embodiments of the present invention can be fully integrated in one processing unit, it can also To be each unit individually as a unit, can also be integrated in one unit with two or more units.

Claims

1. a kind of image processing method, which is characterized in that the described method includes:

Image to be adjusted and second feature information are obtained, the second feature information is target object in the image to be adjusted Characteristic information；

According to the fisrt feature information and the second feature information determine the image to be adjusted relative to the reference First offset information of image；

2. method according to claim 1, which is characterized in that the acquisition fisrt feature information, comprising:

Obtain the first image data of the target object in the reference picture；Wherein, the reference picture is at the terminal It is collected when the first posture；

According to the type of the target object, preset algorithm used by the characteristic information for extracting the target object is determined；

3. method according to claim 2, which is characterized in that obtaining the of the target object in the reference picture Before the step of one image data, the method also includes:

Receive the image processing requests that terminal is sent；

It is requested based on described image processing, Xiang Suoshu terminal sends the first request message, wherein first request message is used for The first image data and the terminal of request target object are in the first weight parameter when the first posture.

4. method according to claim 3, which is characterized in that described to obtain target object in the reference picture First image data, comprising:

Receive the first response message that the terminal is sent, wherein the of the target object is carried in the response message One image data and the terminal are in the first weight parameter when the first posture；

First response message is parsed, the first image data of target object in the reference picture is obtained.

5. method according to claim 2, which is characterized in that described to obtain target object in the reference picture First image data, comprising:

According to image to be adjusted, reference picture is obtained from itself pre-stored image data；

According to the second feature information and the reference picture, the first image of the target object in the reference picture is obtained Data.

6. method according to claim 1, which is characterized in that include at least fisrt feature in the fisrt feature information The position of point and the position of second feature point include at least the position and described the of the third feature point in second feature information The position of four characteristic points, it is correspondingly, described determining described wait adjust according to the fisrt feature information and the second feature information Whole image is relative to the first offset information with the reference picture, comprising:

According to the position of the position of the fisrt feature point and second feature point, the fisrt feature point and second spy are determined Levy the first relative positional relationship of point；

According to the position of the position of third feature point and fourth feature point, the third feature point and the 4th spy are determined Levy point the second relative positional relationship, wherein the third feature point be the corresponding characteristic point of the fisrt feature point, the 4th Characteristic point is the corresponding characteristic point of the second feature point；

According to first relative positional relationship and second relative positional relationship, determine the image to be adjusted relative to institute State the first offset information of reference picture.

7. method according to claim 6, which is characterized in that the method also includes:

Obtain the first weight parameter when the terminal is in the first posture；

If the terminal is in the second posture when acquiring image to be adjusted, the second when terminal is in the second posture is obtained Weight parameter；

Relative to the end when determining that terminal is in the second posture according to first weight parameter and second weight parameter The second offset information when end is in the first posture.

8. any method in -5 according to claim 1, which is characterized in that it is described according to first offset information to institute It states image to be adjusted to be adjusted, the image after being adjusted includes:

The first adjustment angle and the first adjustment direction are determined according to first offset information；

The image to be adjusted is rotated into the first adjustment angle according to the first adjustment direction, the figure after being adjusted Picture.

9. any method in -5 according to claim 1, which is characterized in that include at least the in first offset information One deviation angle, it is accordingly, described that the image to be adjusted is adjusted according to first offset information, after being adjusted Image includes:

According to first deflection angle and the image to be adjusted, determines and cut target image；

The cutting target image, the image after being adjusted are removed from the image to be adjusted.

10. any method in -5 according to claim 1, which is characterized in that described according to first offset information The image to be adjusted is adjusted, the step of image after being adjusted before, the method also includes:

The 4th image for meeting the second preset condition is obtained according to the image to be adjusted；

The third feature information of target object in the 4th image is obtained, and according to the third feature information and described first Characteristic information determines third offset information of the 4th image relative to the reference picture；

If the third feature information and the fisrt feature information meet third preset condition, believed according to first offset Breath is adjusted the image to be adjusted, the image after being adjusted.

11. method according to claim 7, which is characterized in that second offset information at least further includes second inclined Angle is moved, the method also includes:

If first offset information and second offset information are unsatisfactory for first preset condition, according to described second Deviation angle, the first deviation angle and the first weight parameter determine third weight parameter；

The adjustment direction of terminal posture is determined according to the third weight parameter and second weight parameter；

Generate the prompt information of simultaneously output adjustment terminal posture, wherein include the adjustment direction in the prompt information.

12. a kind of image processing apparatus, which is characterized in that described device includes: the first acquisition module, second obtains module, the One determining module and adjustment module, in which:

Described first obtains module, and for obtaining fisrt feature information, the fisrt feature information is target pair in reference picture The characteristic information of elephant；

Described second obtains module, and for obtaining image to be adjusted and second feature information, the second feature information is described The characteristic information of target object in image to be adjusted；

First determining module, it is described to be adjusted for being determined according to the fisrt feature information and the second feature information Image is relative to the first offset information with the reference picture；

The adjustment module, for being adjusted according to first offset information to the image to be adjusted, after being adjusted Image.

13. a kind of image processing equipment, which is characterized in that the equipment includes at least: the equipment includes at least: memory, Communication bus and processor, in which:

The memory, for storing image processing program；

14. a kind of computer readable storage medium, which is characterized in that be stored at image on the computer readable storage medium Program is managed, is realized when described image processing routine is executed by processor at the image as described in any one of claims 1 to 11 The step of reason method.