CN110047102A

CN110047102A - Methods, devices and systems for output information

Info

Publication number: CN110047102A
Application number: CN201910314590.8A
Authority: CN
Inventors: 王旭
Original assignee: Beijing ByteDance Network Technology Co Ltd
Current assignee: Beijing ByteDance Network Technology Co Ltd
Priority date: 2019-04-18
Filing date: 2019-04-18
Publication date: 2019-07-23

Abstract

Embodiment of the disclosure discloses the methods, devices and systems for output information.One specific embodiment of this method includes: acquisition target image；Target image is input to target object location model trained in advance, obtain location information set, wherein, location information in location information set is used to indicate the position of the target object object in target image, the position for the target object object that target object location model is used to determine that image includes；Based on the location information in location information set, the key point information of the target object object in target image is determined；Export obtained key point information.The embodiment enriches the method for determination of key point, helps to improve the accuracy of the subject in detection image, facilitates under the premise of ensuring accuracy in detection, improves the recognition speed of key point.

Description

Methods, devices and systems for output information

Technical field

Embodiment of the disclosure is related to field of computer technology, and in particular to for output information method, apparatus and be System.

Background technique

Image recognition refers to and is handled image, analyzed and understood using computer, to identify various different modes The technology of target and object.In general industrial application, image acquiring device shooting picture or video are generallyd use, then Further identifying processing is done again.

Image object detection is a committed step in image recognition processes, and image object detection can be determined from image The interested target in position.Image object detection generally requires to judge the specific category of each target, and provides the side of each target Boundary's frame.Image object detection is in recognition of face, medical image, intelligent video monitoring, robot navigation, the image inspection based on content The fields such as rope, image basedrendering, picture editting and augmented reality are all widely used.

In the prior art, if the subject for including in image it is smaller (such as the subject pixel that includes compared with It is few), then it is generally difficult to detect the subject in image.

Summary of the invention

The present disclosure proposes the methods, devices and systems for output information.

In a first aspect, embodiment of the disclosure provides a kind of method for output information, this method comprises: obtaining mesh Logo image；Target image is input to target object location model trained in advance, obtains location information set, wherein position Location information in information aggregate is used to indicate the position of the target object object in target image, and target object location model is used In the position for determining target object object that image includes；Based on the location information in location information set, target image is determined In target object object key point information；Export obtained key point information.

In some embodiments, based on the location information in location information set, the target object in target image is determined The key point information of object, comprising: put the image-region at the position of the location information instruction in location information set Big processing, obtains the corresponding enlarged drawing region of location information；Obtained enlarged drawing region is sequentially input to preparatory instruction Experienced key point determines model, obtains being located at the key point information of the target object object at the position of location information instruction.

In some embodiments, the number of parameters Yu key point of target object location model determine the number of parameters of model Ratio is greater than 10.

In some embodiments, the number of parameters of target object location model is greater than 1,000,000.

In some embodiments, the object of target object object instruction is vehicle, and the key point of key point information instruction refers to Show the position where license plate.

In some embodiments, this method further include: meet in response to the recognition result identified to key point Preset condition sends the letter for forbidding vehicle belonging to the corresponding license plate of the key point to be travelled to target control equipment Number, wherein preset condition includes that recognition result indicating number plate number belongs to default license plate number set.

In some embodiments, target image is the video frame in predetermined video；And this method further include: will Video frame in video after target image is input to target object location model, obtains location information set again；It is right Location information in the location information set obtained again is based on the location information, determines and is located at location information instruction The key point information and output of target object object at position.

Second aspect, embodiment of the disclosure provides another method for output information, this method comprises: first Equipment obtains target video；First equipment selecting video frame from target video is such as issued as target video frame, and execution It send step: selected target video frame being input to target object location model trained in advance, obtains location information, In, location information is used to indicate the position of the target object object in video frame, and target object location model is for determining video The position for the target object object that frame includes；In response to obtaining location information, obtained location information is sent to second and is set It is standby；Second equipment is in response to receiving location information, based on received location information, determines the target object in video frame The key point information and output of object.

In some embodiments, this method further include: the first equipment is in response in the target video frame of choosing this The location information of target object object is sent, and the video frame of unselected mistake is chosen from target video as target video Frame continues to execute sending step.

In some embodiments, based on received location information, the pass of the target object object in video frame is determined Key point information and output, comprising: amplify the image-region being located at the position that received location information indicates Processing, obtains the corresponding enlarged drawing region of location information；Obtained enlarged drawing region is input to pass trained in advance Key point determines model, obtains the key point information of the target object object at the position of location information instruction.

In some embodiments, this method further include: the second equipment is in response to the identification that is identified to key point As a result meet preset condition, send to target control equipment for forbidding vehicle belonging to the corresponding license plate of the key point to be gone The signal sailed, wherein preset condition includes that recognition result indicating number plate number belongs to default license plate number set.

The third aspect, embodiment of the disclosure provide a kind of device for output information, which includes: to obtain list Member is configured to obtain target image；First input unit is configured to for target image being input to object trained in advance Body location model obtains location information set, wherein the location information in location information set is used to indicate in target image The position of target object object, the position for the target object object that target object location model is used to determine that image includes；It determines Unit is configured to determine the key of the target object object in target image based on the location information in location information set Point information；First output unit is configured to export obtained key point information.

In some embodiments, determination unit includes: amplification module, is configured to the position letter in location information set It ceases the image-region at the position of instruction and amplifies processing, obtain the corresponding enlarged drawing region of location information；Input module, It is configured to sequentially input in obtained enlarged drawing region to key point trained in advance and determines model, obtain being located at position The key point information of target object object at the position of information instruction.

In some embodiments, device further include: transmission unit is configured in response to identify to key point To recognition result meet preset condition, to target control equipment send for forbidding vehicle belonging to the corresponding license plate of the key point The signal travelled, wherein preset condition includes that recognition result indicating number plate number belongs to default license plate number set.

In some embodiments, target image is the video frame in predetermined video；And the device further include: the Two input units are configured to the video frame in video after target image being input to target object location model, then It is secondary to obtain location information set；Second output unit is configured to for the position letter in the location information set that obtains again Breath, be based on the location information, determine be located at the location information instruction position at target object object key point information with And output.

Fourth aspect, embodiment of the disclosure provide a kind of system for output information, which sets including first Second equipment standby and with the communication connection of the first equipment, in which: the first equipment is configured to: obtaining target video；It is regarded from target Selecting video frame is as target video frame in frequency, and executes following sending step: selected target video frame is input to Trained target object location model in advance, obtains location information, wherein location information is used to indicate the object in video frame The position of body object, the position for the target object object that target object location model is used to determine that video frame includes；In response to To location information, obtained location information is sent to the second equipment；Second equipment is configured to: in response to receiving position Information determines the key point information and output of the target object object in video frame based on received location information.

In some embodiments, the first equipment is further configured to: in response to the mesh in the target video frame of choosing this The location information of mark subject is sent, and the video frame of unselected mistake is chosen from target video as target video Frame continues to execute sending step.

In some embodiments, the first equipment is further configured to: will be located at what received location information indicated Image-region at position amplifies processing, obtains the corresponding enlarged drawing region of location information；By obtained enlarged drawing Model is determined as region is input to key point trained in advance, obtains the target object object at the position of location information instruction Key point information.

In some embodiments, the second equipment is further configured to: in response to the identification knot identified to key point Fruit meets preset condition, sends to target control equipment for forbidding vehicle belonging to the corresponding license plate of the key point to be travelled Signal, wherein preset condition includes that recognition result indicating number plate number belongs to default license plate number set.

5th aspect, embodiment of the disclosure provide a kind of electronic equipment for output information, comprising: one or more A processor；Storage device is stored thereon with one or more programs, when said one or multiple programs are by said one or more A processor executes, so that the one or more processors are realized in such as above-mentioned first aspect for appointing in the method for output information The method of one embodiment.

6th aspect, embodiment of the disclosure provide a kind of computer-readable medium for output information, deposit thereon Computer program is contained, is realized when which is executed by processor in such as above-mentioned first aspect for appointing in the method for output information The method of one embodiment.

The methods, devices and systems for output information that embodiment of the disclosure provides, by obtaining target image, so Afterwards, target image is input to target object location model trained in advance, obtains location information set, wherein location information Location information in set is used to indicate the position of the target object object in target image, and target object location model is for true Determine the position for the target object object that image includes, later, based on the location information in location information set, determines target image In the key point information of target object object enrich the determination side of key point finally, exporting obtained key point information Formula helps to improve the accuracy of the subject in detection image, facilitates under the premise of ensuring accuracy in detection, improves The recognition speed of key point.

Detailed description of the invention

By reading a detailed description of non-restrictive embodiments in the light of the attached drawings below, the disclosure is other Feature, objects and advantages will become more apparent upon:

Fig. 1 is that one embodiment of the disclosure can be applied to exemplary system architecture figure therein；

Fig. 2 is the flow chart according to one embodiment of first of the disclosure method for output information；

Fig. 3 is the schematic diagram according to an application scenarios of first of the disclosure method for output information；

Fig. 4 is the flow chart according to one embodiment of second of the disclosure method for output information；

Fig. 5 is the structural schematic diagram according to one embodiment of the device for output information of the disclosure；

Fig. 6 is the interactive process schematic diagram according to the system for output information of the disclosure；

Fig. 7 is another interactive process schematic diagram according to the system for output information of the disclosure；

Fig. 8 is adapted for the structural schematic diagram for the computer system for realizing the electronic equipment of embodiment of the disclosure.

Specific embodiment

The disclosure is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is retouched The specific embodiment stated is used only for explaining related invention, rather than the restriction to the invention.It also should be noted that in order to Convenient for description, part relevant to related invention is illustrated only in attached drawing.

It should be noted that in the absence of conflict, the feature in embodiment and embodiment in the disclosure can phase Mutually combination.The disclosure is described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.

Fig. 1 is shown can the method for output information using embodiment of the disclosure or the dress for output information The exemplary system architecture 100 for the embodiment set.

As shown in Figure 1, system architecture 100 may include terminal device 101,102,103, network 104 and server 105. Network 104 between terminal device 101,102,103 and server 105 to provide the medium of communication link.Network 104 can be with Including various connection types, such as wired, wireless communication link or fiber optic cables etc..

User can be used terminal device 101,102,103 and be interacted by network 104 with server 105, to receive or send out Send data (such as image) etc..Various client applications, such as video playing can be installed on terminal device 101,102,103 Software, the application of Domestic News class, image processing class application, web browser applications, the application of shopping class, searching class are applied, immediately Means of communication, mailbox client, social platform software etc..

Terminal device 101,102,103 can be hardware, be also possible to software.As an example, when terminal device 101, 102,103 when being hardware, can be with display screen and/or the various electronic equipments of photographic device, including but not limited to intelligent hand Machine, tablet computer, E-book reader, MP3 player (Moving Picture Experts Group Audio Layer III, dynamic image expert's compression standard audio level 3), MP4 (Moving Picture Experts Group Audio Layer IV, dynamic image expert's compression standard audio level 4) player, pocket computer on knee and desktop computer etc. Deng.When terminal device 101,102,103 is software, may be mounted in above-mentioned cited electronic equipment.It may be implemented At multiple softwares or software module (such as providing the software of Distributed Services or software module), also may be implemented into single Software or software module.It is not specifically limited herein.

Server 105 can be to provide the server of various services, such as to the figure that terminal device 101,102,103 is sent Background server as carrying out target object object detection.Background server can be to the target object object for including in the image Positioned, key point identification etc. processing.And processing result (such as positioning result, key point recognition result) is stored in this Ground, alternatively, feeding back to terminal device.As an example, server 105 can be cloud server, it is also possible to physical server

It should be noted that server can be hardware, it is also possible to software.When server is hardware, may be implemented At the distributed server cluster that multiple servers form, individual server also may be implemented into.It, can when server is software To be implemented as multiple softwares or software module (such as providing the software of Distributed Services or software module), also may be implemented At single software or software module.It is not specifically limited herein.

It should also be noted that, the method provided by embodiment of the disclosure for output information can be held by server Row, can also be executed, can also be fitted to each other execution by server and terminal device by terminal device.Correspondingly, for exporting The various pieces (such as each unit, subelement, module, submodule) that the device of information includes can all be set to server In, it can also all be set in terminal device, can also be respectively arranged in server and terminal device.

It should be understood that the number of terminal device, network and server in Fig. 1 is only schematical.According to realization need It wants, can have any number of terminal device, network and server.For example, when being used for the operation of output information method thereon Electronic equipment when not needing to carry out data transmission with other electronic equipments, which can only include being used for output information The electronic equipment (such as server or terminal device) of method operation thereon.

With continued reference to Fig. 2, the one embodiment for the method that output information is used for according to first of the disclosure is shown Process 200.This is used for the method for output information, comprising the following steps:

Step 201, target image is obtained.

In the present embodiment, for the executing subject of the method for output information, (such as server shown in FIG. 1 or terminal are set It is standby) target image can be obtained from other electronic equipments or locally by wired connection mode or radio connection.

Wherein, above-mentioned target image can be arbitrary image, for example, target image can be object therein to be determined The image of the position of body object.Target image can be the video frame in video, be also possible to an independent image.Above-mentioned mesh Mark subject can be the object of the physical entity presented on the image shot to any physical entity.Example Such as, vehicle is shot to have obtained vehicle image, then, the target object object which includes can be the vehicle Vehicle in image.As an example, target object object can include but is not limited to following any one: Vehicle Object, face pair As, human object, trees object, house object etc..

Step 202, target image is input to target object location model trained in advance, obtains location information set.

In the present embodiment, target image can be input to target object positioning mould trained in advance by above-mentioned executing subject Type obtains location information set.Wherein, the location information in location information set is used to indicate the target object in target image Position of the object in the target image, the position for the target object object that target object location model is used to determine that image includes It sets.

Herein, the location information in above-mentioned location information set can indicate each pixel of composition target object object Position where point can also indicate that the position of the rectangle frame in target image including target object object.

It is appreciated that practice in, above-mentioned executing subject can by target image include target object object each square The smallest rectangle frame (such as rectangle frame where minimum circumscribed rectangle) in shape frame is determined as target object object in the target Position in image.It is possible thereby to improve positional accuracy, and improve the accuracy that subsequent step executes.

As an example, above-mentioned target object location model can be the convolutional Neural obtained using the training of deep learning algorithm Network is also possible to bivariate table or database that associated storage has image and location information.

Step 203, based on the location information in location information set, the pass of the target object object in target image is determined Key point information.

In the present embodiment, above-mentioned executing subject can determine target figure based on the location information in location information set The key point information of target object object as in.

Wherein, the key point information of above-mentioned target object object can serve to indicate that the position of the key point of target object object It sets.In practice, different key points can be set to be directed to different subjects according to specific requirements.

As an example, the key point information of target object object can be used for when target object object is Vehicle Object The position of four angle points of indicating number plate.When target object object is face object, the key point information of target object object It can serve to indicate that following any one or more position: eyes, nose, mouth.

It is appreciated that location information and key point information may be incorporated for the target object object in instruction target image Position.When the two is used to the position of the target object object in instruction target image, relative to location information, key point letter Breath can serve to indicate that the position of more fine-grained target object object.For example, when target object object is license plate object, it should The location information of license plate object can be the biggish rectangle frame including the license plate object and (be also possible to include the license plate object Four angular coordinates of biggish rectangle frame), and the key point information of the license plate object can be the license plate pair in target image The minimum circumscribed rectangle of elephant (sit by four angle points for being also possible to the biggish rectangle frame in target image including the license plate object Mark).

As an example, above-mentioned executing subject can be using such as lower section for each location information in location information set Formula is based on the location information, determines the key point information of the target object object in target image:

Firstly, determining the image-region of location information instruction from target image.

Then, it is determined that whether each pixel in the image-region determined is key point, to will be key point Each pixel location information in the target image, be determined as the key point letter of the target object object in target image Breath.

In some optional implementations of the present embodiment, above-mentioned executing subject can also execute in the following way The step 203:

Firstly, the image-region at the position of the location information instruction in location information set is amplified processing, obtain To the corresponding enlarged drawing region of location information.

Herein, above-mentioned executing subject can use the image interpolation algorithm based on region, the image interpolation based on edge The methods of algorithm amplifies processing to above-mentioned image-region, so that it is corresponding as location information to obtain enlarged drawing region Enlarged drawing region.It is appreciated that above-mentioned executing subject can obtain an enlarged drawing region for each location information.

Then, obtained enlarged drawing region is sequentially input to key point trained in advance and determines model, obtained in place The key point information of target object object at the position of location information instruction.

Wherein, above-mentioned key point determines that model is determined for the key point of target object object in enlarged drawing region Information.

As an example, above-mentioned key point determines that model can be the convolutional Neural net obtained based on machine learning algorithm training Network model is also possible to bivariate table or database that associated storage has enlarged drawing region and key point information.

It is appreciated that when key point determines that model interaction is stored with enlarged drawing region and key point information, it is above-mentioned to hold Row main body can as follows, to determine the key point information of target object object:

Firstly, determining that key point determines in each enlarged drawing region stored in model, with the enlarged drawing inputted As the highest enlarged drawing region of the similarity in region.

Then, it is determined that key point determines the key point information in model with the enlarged drawing region associated storage determined It is located at the key point information of the target object object at the position of location information instruction as the resulting.

It should be noted that amplifying the enlarged drawing region obtained after processing relative to original image region, composition figure As the visibility of the pixel in region will become higher, the pass of target object object is determined based on enlarged drawing region as a result, Key point information can be obtained more accurately relative to the key point information for being determined target object object based on original image region The key point information of target object object.

In some optional implementations of the present embodiment, the number of parameters and key point of target object location model are true The ratio of the number of parameters of cover half type is greater than 10.

It is appreciated that training obtains the process of model, the as parameter value of the parameter of adjustment model so that model meet it is pre- If the process of condition (such as the functional value of predetermined loss function is less than preset threshold).As an example, above-mentioned model Parameter can include but is not limited at least one of following: weight (weight), biasing (bias), input layer neuron number, Neuron number of output layer etc..

In practice, the number of parameters of model is more, and it is bigger usually to characterize the model.For an input data, large-sized model Required calculation amount is often bigger.Herein, when the number of parameters of target object location model and key point determine model When the ratio of number of parameters is greater than 10, target object location model determines that model is bigger compared to key point, target object positioning Model is bigger to the calculation amount of input data, and it is often longer to calculate the time.

In some optional implementations of the present embodiment, the number of parameters of target object location model is greater than 100 Ten thousand.

It should be understood that the number of parameters of model is more, it is bigger usually to characterize the model.For an input data, big mould Calculation amount required for type is often bigger, and the learning ability of model will be stronger.It was verified that working as the ginseng of target object location model When number quantity is greater than 1,000,000, the target object object of very little in target image can be identified.For example, when target object is fixed When the number of parameters of bit model is greater than 1,000,000, the target object pair that shooting distance is 5 meters in target image can be identified As.Wherein, when shooting distance is photographic subjects image, between physical entity indicated by image taking point and target object object Distance.And under normal conditions, the model of existing number of parameters less (such as number of parameters is 20,000) can not then identify The target object object that shooting distance is 5 meters in target image.

As an example, the number of parameters of target object location model can be 5,000,000,10,000,000 etc..

It is appreciated that the number of parameters of target object location model is more, can identify smaller in target image Target object object, also, the accuracy of the target object object identified is also higher.Thus, it is possible to reduce missing inspection, erroneous detection The case where generation, and then obtain more accurate location information and key point information.

Step 204, obtained key point information is exported.

In the present embodiment, above-mentioned executing subject can export the obtained key point information of step 203.

In some optional implementations of the present embodiment, the object of target object object instruction is vehicle, key point Position where the key point indicating number plate of information instruction.

It is appreciated that when the object of target object object instruction is vehicle, also, the key point of key point information instruction refers to When showing the position where license plate, the scheme of embodiment of the disclosure can be applied to Car license recognition, help to improve Car license recognition Accuracy.

In some optional implementations of the present embodiment, meet in the recognition result identified to key point In the case where preset condition, above-mentioned executing subject can also be sent to target control equipment for forbidding the corresponding vehicle of the key point The signal that vehicle belonging to board is travelled.

Wherein, preset condition includes that recognition result indicating number plate number belongs to default license plate number set.Target control is set The standby equipment that can be the vehicle for controlling the instruction of target object object, for example, vehicle obstructing device, vehicle start prevent dress It sets.Above-mentioned default license plate number set can be the set of the license plate number in escape state vehicle violating the regulations.Knowing as a result, In the case that other result meets preset condition, above-mentioned executing subject can also be sent to target control equipment for forbidding above-mentioned mesh The signal that the vehicle for the Vehicle Object instruction that logo image includes is travelled, to prevent in escape state vehicle row violating the regulations It sails, so that related management personnel are managed in escape state vehicle violating the regulations and its driver.

In some optional implementations of the present embodiment, target image is the video frame in predetermined video. Following steps can also be performed in above-mentioned executing subject:

Video frame in video after target image is input to above-mentioned target object location model by step 1, then It is secondary to obtain location information set.

Wherein, the video frame after target image can be it is adjacent with target video frame and be located at target video frame it Video frame afterwards is also possible to target video frame period preset quantity video frame (such as 10 frames, 15 frames) and positioned at target Video frame after video frame.

Step 2 is based on the location information for the location information in the location information set that obtains again, and determination is located at The key point information and output of target object object at the position of location information instruction.

Herein, the executive mode of above-mentioned steps one and step 2 respectively with above-mentioned steps 202 and the basic class of step 203 Seemingly, details are not described herein.

It is appreciated that this optional implementation can determine each video frame in video or conform to a predetermined condition The key point information of target object object in video frame improves it is possible thereby to realize the tracking of target object object in video The accuracy of subject in tracking video, facilitates under the premise of ensuring accuracy in detection, improves the knowledge of key point Other speed.

It is the one of the application scenarios for the method that output information is used for according to first of the present embodiment with continued reference to Fig. 3, Fig. 3 A schematic diagram.In the application scenarios of Fig. 3, server 301 has got target image 3011 from terminal device 302.Then, it takes Target image 3011 is input to target object location model 3012 trained in advance by business device 301, obtains location information set (packet Include location information 30131 and location information 30132).Wherein, the location information in location information set is used to indicate target image In target object object (such as Vehicle Object) position, target object location model is for determining the object that image includes The position of body object.In diagram, location information is by including that the rectangle frame of target object object characterizes.Later, server 301 is based on Location information (i.e. location information 30131 and location information 30132) in location information set, determines in target image 3011 The key point information 3014 of target object object.Finally, server 301 exports obtained key point letter to terminal device 302 Breath 3014.

Then target image is input to pre- by the method provided by the above embodiment of the disclosure by obtaining target image First trained target object location model, obtains location information set, wherein the location information in location information set is for referring to Show the position of the target object object in target image, target object location model is for determining the target object pair that image includes The position of elephant based on the location information in location information set, determines the key of the target object object in target image later Point information enriches the method for determination of key point finally, exporting obtained key point information, by determining in image first Target object object location information, be then based on the key point for the target object object that location information determines in target image The accuracy of the subject in detection image can be improved in the mode of information, facilitates in the premise for ensuring accuracy in detection Under, improve the recognition speed of key point.

With continued reference to Fig. 4, the one embodiment for the method that output information is used for according to second of the disclosure is shown Process 400.This is used for the method for output information, comprising the following steps:

Step 401, the first equipment obtains target video.

In the present embodiment, the first equipment can be set by wired connection mode or radio connection from other electronics Standby or local acquisition target video.

Wherein, above-mentioned first equipment can be any electronic equipment.For example, the first equipment can be terminal device or clothes Business device.Above-mentioned target video can be any video, for example, target video can be to the position to target object object therein The video track set.Above-mentioned target object object can be institute in the video shot to any physical entity The object of the physical entity presented.For example, vehicle is shot to have obtained automobile video frequency, then, which includes Target object object can be the vehicle in the automobile video frequency.As an example, target object object can include but is not limited to It is any one of following: Vehicle Object, face object, human object, trees object, house object etc..

Step 402, the first equipment selecting video frame from target video sends and walks as target video frame, and execution Suddenly.

In the present embodiment, the first equipment and can be held selecting video frame from target video as target video frame Row sending step.Wherein, sending step includes step 4021 and step 4022.

Herein, above-mentioned first equipment can randomly select video frame as target video frame from target video, can also Using the first frame in selection target video as target video frame.

Step 4021, selected target video frame is input to target object location model trained in advance, is obtained in place Confidence breath.

In the present embodiment, selected target video frame can be input to target object trained in advance by the first equipment Location model obtains location information.Wherein, location information is used to indicate the position of the target object object in video frame, target The position for the target object object that object location model is used to determine that video frame includes.

Herein, above-mentioned location information can indicate the position where each pixel of composition target object object, It can indicate the position of the rectangle frame in target image including target object object.

It is appreciated that practice in, above-mentioned first equipment can by target image include target object object each square Position (such as rectangle frame where minimum circumscribed rectangle) where the smallest rectangle frame in shape frame is determined as target object pair As the position in the target image.It is possible thereby to improve positional accuracy, and improve the accuracy that subsequent step executes.

Step 4022, in response to obtaining location information, obtained location information is sent to the second equipment.

In the present embodiment, in response to obtaining location information, obtained location information can be sent to by the first equipment Second equipment.Wherein, the second equipment can be any electronic equipment.For example, the second equipment can be terminal device or service Device.

It is appreciated that first, second in above-mentioned first equipment, the second equipment, it is used only as distinguishing equipment, composition pair The particular determination of equipment.First equipment can be two physically relatively independent equipment, two be also possible in an equipment A unit.

It should be noted that due to may include one or more (at least two) target objects in selected video frame Object, and each target object object can correspond to a location information, thus the available one or more positions of the first equipment Confidence breath.When the available multiple location informations of the first equipment, once the first equipment obtains a location information, then first set The standby location information that can be obtained this is sent to the second equipment.

It is appreciated that herein, " the first equipment obtains location information " can be used as that " the first equipment is by obtained position Information is sent to the second equipment " trigger condition.

Step 403, the second equipment, based on received location information, determines video in response to receiving location information The key point information and output of target object object in frame.

In the present embodiment, above-mentioned second equipment can determine target figure based on the location information in location information set The key point information of target object object as in.

Wherein, the key point information of above-mentioned target object object can serve to indicate that the key point of target object object.It is real It, can be according to specific requirements, to be directed to the position that different key points is arranged in different subjects in trampling.As an example, working as When target object object is Vehicle Object, the key point information of target object object can serve to indicate that four angle points of license plate Position.When target object object is face object, the key point information of target object object can serve to indicate that following any One or more position: eyes, nose, mouth.

It should be noted that the first equipment can be to the second equipment when the available multiple location informations of the first equipment Send multiple location informations.Once the second equipment obtains a location information, then the second equipment can be based on received position Confidence breath, determines the key point information and output of the target object object in video frame.

It is appreciated that herein, " the second equipment receives location information " can be used as that " the second equipment is based on received Location information, determine the key point information and output of the target object object in video frame " trigger condition.

As an example, above-mentioned second equipment can be using such as lower section for each location information in location information set Formula is based on the location information, determines the key point information of the target object object in target image:

In some optional implementations of the present embodiment, above-mentioned second equipment can also execute in the following way The step 403:

Firstly, the image-region being located at the position that received location information indicates is amplified processing, obtain The corresponding enlarged drawing region of location information

Herein, above-mentioned second equipment can use the image interpolation algorithm based on region, the image interpolation based on edge The methods of algorithm amplifies processing to above-mentioned image-region, so that it is corresponding as location information to obtain enlarged drawing region Enlarged drawing region.It is appreciated that above-mentioned second equipment can obtain an enlarged drawing region for each location information.

Then, obtained enlarged drawing region is input to key point trained in advance and determines model, obtain being located at position The key point information of target object object at the position of confidence breath instruction.

It is appreciated that when key point determines that model interaction is stored with enlarged drawing region and key point information, above-mentioned Two equipment can as follows, to determine the key point information of target object object:

It is appreciated that training obtains the process of model, the as parameter value of the parameter of adjustment model so that model meet it is pre- If the process of condition.As an example, the parameter of above-mentioned model can include but is not limited to it is at least one of following: weight (weight), (bias), the neuron number of input layer, the neuron number of output layer etc. are biased.

Step 404, obtained key point information is exported.

In the present embodiment, above-mentioned second equipment can export the obtained key point information of step 403.

In some optional implementations of the present embodiment, meet in the recognition result identified to key point In the case where preset condition, above-mentioned second equipment can also be sent to target control equipment for forbidding the corresponding vehicle of the key point The signal that vehicle belonging to board is travelled.

Wherein, preset condition includes that recognition result indicating number plate number belongs to default license plate number set.Target control is set The standby equipment that can be the vehicle for controlling the instruction of target object object, for example, vehicle obstructing device, vehicle start prevent dress It sets.Above-mentioned default license plate number set can be the set of the license plate number in escape state vehicle violating the regulations.Knowing as a result, In the case that other result meets preset condition, above-mentioned second equipment can also be sent to target control equipment for forbidding above-mentioned mesh The signal that the vehicle for the Vehicle Object instruction that logo image includes is travelled, to prevent in escape state vehicle row violating the regulations It sails, so that related management personnel are managed in escape state vehicle violating the regulations and its driver.

It is appreciated that above-mentioned second equipment can be identified by the license plate to vehicle, thus by judging the license plate Whether predetermined license plate set in violating the regulations escape state is belonged to, to determine whether attribute information instruction vehicle is in separated Chapter escape state.

In some optional implementations of the present embodiment, in response to the target in the target video frame of choosing this The location information of subject is sent, and the video frame that the first equipment can also choose unselected mistake from target video is made For target video frame, above-mentioned sending step is continued to execute.

It should be noted that the first equipment can be successively to second when the available multiple location informations of the first equipment Equipment sends multiple location informations.When the position of the target object object in the target video frame that the first equipment chooses this is believed Breath is when being sent, and the first equipment can choose the video frame of unselected mistake as target video frame from target video, after It is continuous to execute above-mentioned sending step

It is appreciated that herein, " first the second equipment of equipment receives location information " can be used as that " the first equipment is from mesh Mark video in choose unselected mistake video frame as target video frame, continue to execute above-mentioned sending step " trigger condition.

The method that second provided by the above embodiment of the disclosure is used for output information obtains target by the first equipment Video, then, the first equipment selecting video frame from target video and execute following sending step as target video frame: Selected target video frame is input to target object location model trained in advance, obtains location information, wherein position letter The position for the target object object being used to indicate in video frame is ceased, target object location model is for determining the mesh that video frame includes Mark the position of subject；In response to obtaining location information, obtained location information is sent to the second equipment and second Equipment is in response to receiving location information, based on received location information, determines the target object object in video frame Key point information and output enrich the method for determination of key point, by determining the target object object in image first Location information is then based on the mode that location information determines the key point information of target object object in target image, can be with The accuracy for improving the subject in detection image, helps to improve key point under the premise of ensuring accuracy in detection Recognition speed.

With further reference to Fig. 5, as the realization to method shown in above-mentioned each figure, present disclose provides one kind for exporting letter One embodiment of the device of breath, the Installation practice is corresponding with embodiment of the method shown in Fig. 2, except following documented special Sign is outer, which can also include feature identical or corresponding with embodiment of the method shown in Fig. 2, and generates and scheme Embodiment of the method shown in 2 is identical or corresponding effect.The device specifically can be applied in various electronic equipments.

As shown in figure 5, the device 500 for output information of the present embodiment includes: that the input of acquiring unit 501, first is single Member 502, determination unit 503 and the first output unit 504.Wherein, acquiring unit 501 are configured to obtain target image；First Input unit 502 is configured to for target image being input to target object location model trained in advance, obtains position information set It closes, wherein the location information in location information set is used to indicate the position of the target object object in target image, object The position for the target object object that body location model is used to determine that image includes；Determination unit 503 is configured to believe based on position Location information in breath set, determines the key point information of the target object object in target image；First output unit 504, It is configured to export obtained key point information.

In the present embodiment, for the acquiring unit 501 of the device of output information 500 can by wired connection mode or Person's radio connection obtains target image from other electronic equipments or locally.Wherein, above-mentioned target image can be arbitrarily Image, for example, target image can be the image of the position of target object object therein to be determined.Above-mentioned target object object It can be the object of the physical entity presented on the image shot to any physical entity.

In the present embodiment, target image can be input to object trained in advance by above-mentioned first input unit 502 Body location model obtains location information set.Wherein, the location information in location information set is used to indicate in target image Position of the target object object in the target image, target object location model is for determining the target object pair that image includes The position of elephant.

In the present embodiment, above-mentioned determination unit 503 can determine target based on the location information in location information set The key point information of target object object in image.Wherein, the key point information of above-mentioned target object object can be used for referring to Show the key point of target object object.

In the present embodiment, above-mentioned first output unit 504 can export the obtained key point of determination unit 503 letter Breath.

In some optional implementations of the present embodiment, determination unit 503 includes: amplification module (not shown) It is configured to the image-region at the position of the location information instruction in location information set amplifying processing, obtains position The corresponding enlarged drawing region of information；Input module (not shown) is configured to obtained enlarged drawing region successively It is input to key point trained in advance and determines model, obtain being located at the pass of the target object object at the position of location information instruction Key point information.

In some optional implementations of the present embodiment, device 500 further include: transmission unit (not shown) quilt The recognition result for being configured to identify key point meets preset condition, is used for the transmission of target control equipment The signal for forbidding vehicle belonging to the corresponding license plate of the key point to be travelled, wherein preset condition includes recognition result instruction License plate number belongs to default license plate number set.

In some optional implementations of the present embodiment, target image is the video frame in predetermined video； And device 500 further include: the second input unit (not shown) is configured to be located at after target image in video Video frame is input to target object location model, obtains location information set again；Second output unit (not shown) quilt It is configured to be based on the location information for the location information in the location information set that obtains again, determines that being located at the position believes Cease the key point information and output of the target object object at the position of instruction.

The device provided by the above embodiment of the disclosure obtains target image by acquiring unit 501, and then, first is defeated Enter unit 502 and target image is input to target object location model trained in advance, obtains location information set, wherein position Set the position for the target object object that the location information in information aggregate is used to indicate in target image, target object location model For determining the position of target object object that image includes, later, determination unit 503 is based on the position in location information set Information determines the key point information of the target object object in target image, finally, obtained by the first output unit 504 exports Key point information, enrich the method for determination of key point, pass through first determine image in target object object position letter Breath is then based on the mode that location information determines the key point information of target object object in target image, inspection can be improved The accuracy of subject in altimetric image facilitates under the premise of ensuring accuracy in detection, improves the identification speed of key point Degree.

Illustrated with continued reference to FIG. 6, Fig. 6 is shown according to the interactive process of the system for output information of the disclosure Figure.The system includes the first equipment and the second equipment with the communication connection of the first equipment, in which: the first equipment is configured to: being obtained Take target video；Selecting video frame is as target video frame from target video, and executes following sending step: will be selected Target video frame be input in advance trained target object location model, obtain location information, wherein location information is for referring to Show the position of the target object object in video frame, target object location model is for determining the target object pair that video frame includes The position of elephant；In response to obtaining location information, obtained location information is sent to the second equipment；Second equipment is configured At: in response to receiving location information, based on received location information, determine the pass of the target object object in video frame Key point information and output.

As shown in fig. 6, in step 601, the first equipment obtains target video frame.

In the present embodiment, the available target video frame of the first equipment.

Wherein, above-mentioned first equipment can be any electronic equipment.For example, the first equipment can be terminal device or clothes Business device.Above-mentioned target video can be any video, for example, target video can be to the position to target object object therein The video track set.Above-mentioned target object object can be institute in the video shot to any physical entity The object of the physical entity presented.

In step 602, selecting video frame is as target video frame from target video for the first equipment, by selected mesh Mark video frame is input to target object location model trained in advance, obtains location information.

In the present embodiment, the first equipment can selecting video frame, will be selected as target video frame from target video The target video frame taken is input to target object location model trained in advance, obtains location information.Wherein, location information is used for Indicate the position of the target object object in video frame, target object location model is for determining the target object that video frame includes The position of object

In step 603, the first equipment sends location information to the second equipment.

In the present embodiment, the first equipment can be to location information obtained in the second equipment sending step 602.

In step 604, the second equipment determines the target object object in video frame based on received location information Key point information and output.

In the present embodiment, the second equipment can determine the object in video frame based on received location information The key point information and output of body object.

Wherein, the key point information of above-mentioned target object object can serve to indicate that the key point of target object object.It is real It, can be according to specific requirements, to be directed to the position that different key points is arranged in different subjects in trampling.

In some optional implementations of the present embodiment, the first equipment is further configured to: will be located at and be received To location information instruction position at image-region amplify processing, obtain the corresponding enlarged drawing area of location information Domain；Obtained enlarged drawing region is input to key point trained in advance and determines model, obtains the position of location information instruction Set the key point information of the target object object at place.

In some optional implementations of the present embodiment, the second equipment is further configured to: in response to clicking through to key The recognition result that row identification obtains meets preset condition, sends to target control equipment for forbidding the corresponding license plate of the key point The signal that affiliated vehicle is travelled, wherein preset condition includes that recognition result indicating number plate number belongs to default license plate number Code collection is closed.

It should be noted that the present embodiment can also include embodiment corresponding with Fig. 4 in addition to documented content above In identical or corresponding feature, generate identical or corresponding effect, and solves the problems, such as identical or corresponding, no longer go to live in the household of one's in-laws on getting married herein It states.

The system provided by the above embodiment for output information of the disclosure includes the first equipment and leads to the first equipment Believe the second equipment of connection, in which: the first equipment is configured to: obtaining target video；Selecting video frame is made from target video For target video frame, and the following sending step of execution: selected target video frame is input to object trained in advance Body location model, obtains location information, wherein location information is used to indicate the position of the target object object in video frame, mesh The position for the target object object that mark object location model is used to determine that video frame includes；In response to obtaining location information, by institute Obtained location information is sent to the second equipment；Second equipment is configured to: in response to receiving location information, based on being received The location information arrived determines the key point information and output of the target object object in video frame, to enrich key point Method of determination, by first determine image in target object object location information, be then based on location information and determine mesh The mode of the key point information of target object object in logo image, can be improved the accurate of the subject in detection image Property, facilitate under the premise of ensuring accuracy in detection, improves the recognition speed of key point.

It referring next to Fig. 7, Fig. 7 is shown according to another interactive process of the system for output information of the disclosure It is intended to.The system includes the first equipment and the second equipment with the communication connection of the first equipment, in which: the first equipment is configured to: Obtain target video；Selecting video frame is as target video frame from target video, and executes following sending step: will be selected The target video frame taken is input to target object location model trained in advance, obtains location information, wherein location information is used for Indicate the position of the target object object in video frame, target object location model is for determining the target object that video frame includes The position of object；In response to obtaining location information, obtained location information is sent to the second equipment.Second equipment is configured At: in response to receiving location information, based on received location information, determine the pass of the target object object in video frame Key point information and output.First equipment is further configured to: in response to the target object in the target video frame of choosing this The location information of object is sent, and the video frame of unselected mistake is chosen from target video as target video frame, is continued Execute sending step.

As shown in fig. 7, first equipment obtains target video in step 701.

In the present embodiment, the available target video of the first equipment.

In a step 702, selecting video frame is as target video frame from target video for the first equipment, by selected mesh Mark video frame is input to target object location model trained in advance, generates first location information.

In the present embodiment, the first equipment can selecting video frame, will be selected as target video frame from target video The target video frame taken is input to target object location model trained in advance, generates first location information.

In step 703, the first equipment sends first location information to the second equipment.

In the present embodiment, the first equipment can send first location information to the second equipment.

In step 704, the first equipment generates second location information.

In the present embodiment, second location information can be generated in the first equipment.

In step 705, the second equipment is based on first location information, determines the pass of the target object object in video frame Key point information and output.

In the present embodiment, the second equipment can be based on first location information, determine the target object pair in video frame The key point information and output of elephant.

In step 706, the first equipment sends second location information to the second equipment.

In the present embodiment, the first equipment can send second location information to the second equipment.

In step 707, the first equipment determines the position letter of the target object object in the target video frame that this is chosen Breath is sent.

In the present embodiment, the first equipment can determine the position of the target object object in the target video frame of this selection Confidence breath is sent.

In step 708, the second equipment is based on second location information, determines the pass of the target object object in video frame Key point information and output.

In the present embodiment, the second equipment can be based on second location information, determine the target object pair in video frame The key point information and output of elephant.

In step 709, the first equipment chooses the video frame of unselected mistake as target video frame from target video, Selected target video frame is input to above-mentioned target object location model, obtains third location information.

In the present embodiment, the video frame that the first equipment can choose unselected mistake from target video is regarded as target Selected target video frame is input to above-mentioned target object location model, obtains third location information by frequency frame.

In step 720, the first equipment sends third location information to the second equipment.

In the present embodiment, the first equipment can send third location information to the second equipment.

In step 711, the first equipment generates the location information of the 4th target object object.

In the present embodiment, the location information of the 4th target object object can be generated in the first equipment.

In step 712, the first equipment sends the 4th location information to the second equipment.

In the present embodiment, the first equipment can send the 4th location information to the second equipment.

In step 713, the second equipment is based on third location information, determines the pass of the target object object in video frame Key point information and output.

In the present embodiment, the second equipment can be based on third location information, determine the target object pair in video frame The key point information and output of elephant.

In step 714, the second equipment is based on the 4th location information, determines the pass of the target object object in video frame Key point information and output.

In the present embodiment, the second equipment can be based on the 4th location information, determine the target object pair in video frame The key point information and output of elephant.

In the present embodiment, the second equipment is based on the 4th location information in the following way, determines in video frame The key point information of target object object:

Firstly, the image-region being located at the position that received location information indicates is amplified processing, obtain The corresponding enlarged drawing region of location information.

Then, obtained enlarged drawing region is input to key point trained in advance and determines model, obtain position letter Cease the key point information of the target object object at the position of instruction

It should be appreciated that the execution sequence of above-mentioned steps 701- step 714 is only schematical, for example, step 713 may be used also To be executed before step 712.

The first equipment and the second equipment in the system provided by the above embodiment for output information of the disclosure can be with Cooperatively the target object object in video is positioned and (determines the key point of the target object object in video frame Information and output), when the number of parameters for the target object location model that the first equipment is run is more, (such as number of parameters reaches To million, millions), also, the key point that is run of the second equipment determines the number of parameters of model less (such as key point is true The number of parameters of cover half type is 1/10th of the number of parameters of target object location model) when, pass through the first equipment and second Equipment is fitted to each other, and time difference of the two in moving model can be made full use of, for example, determining model in key point When determining the key point information of the target object object in a upper video frame, target object location model can determine next video The location information of target object object in frame.It is possible thereby to improve the knowledge of key point under the premise of ensuring accuracy in detection Other speed.

Below with reference to Fig. 8, it illustrates the electronic equipment that is suitable for being used to realize embodiment of the disclosure, (example is as shown in figure 1 Server or terminal device) 800 structural schematic diagram.Terminal device in embodiment of the disclosure can include but is not limited to all As mobile phone, laptop, digit broadcasting receiver, PDA (personal digital assistant), PAD (tablet computer), PMP are (portable Formula multimedia player), the mobile terminal and such as number TV, desk-top meter of car-mounted terminal (such as vehicle mounted guidance terminal) etc. The fixed terminal of calculation machine etc..Terminal device/server shown in Fig. 8 is only an example, should not be to the implementation of the disclosure The function and use scope of example bring any restrictions.

As shown in figure 8, electronic equipment 800 may include processing unit (such as central processing unit, graphics processor etc.) 801, random access can be loaded into according to the program being stored in read-only memory (ROM) 802 or from storage device 808 Program in memory (RAM) 803 and execute various movements appropriate and processing.In RAM 803, it is also stored with electronic equipment Various programs and data needed for 800 operations.Processing unit 801, ROM 802 and RAM803 are connected with each other by bus 804. Input/output (I/O) interface 805 is also connected to bus 804.

In general, following device can connect to I/O interface 805: including such as touch screen, touch tablet, keyboard, mouse, taking the photograph As the input unit 806 of head, microphone, accelerometer, gyroscope etc.；Including such as liquid crystal display (LCD), loudspeaker, vibration The output device 807 of dynamic device etc.；Storage device 808 including such as tape, hard disk etc.；And communication device 809.Communication device 809, which can permit electronic equipment 800, is wirelessly or non-wirelessly communicated with other equipment to exchange data.Although Fig. 8 shows tool There is the electronic equipment 800 of various devices, it should be understood that being not required for implementing or having all devices shown.It can be with Alternatively implement or have more or fewer devices.Each box shown in Fig. 8 can represent a device, can also root According to needing to represent multiple devices.

Particularly, in accordance with an embodiment of the present disclosure, it may be implemented as computer above with reference to the process of flow chart description Software program.For example, embodiment of the disclosure includes a kind of computer program product comprising be carried on computer-readable medium On computer program, which includes the program code for method shown in execution flow chart.In such reality It applies in example, which can be downloaded and installed from network by communication device 809, or from storage device 808 It is mounted, or is mounted from ROM802.When the computer program is executed by processing unit 801, the implementation of the disclosure is executed The above-mentioned function of being limited in the method for example.

It is situated between it should be noted that computer-readable medium described in embodiment of the disclosure can be computer-readable signal Matter or computer readable storage medium either the two any combination.Computer readable storage medium for example can be with System, device or the device of --- but being not limited to --- electricity, magnetic, optical, electromagnetic, infrared ray or semiconductor, or it is any more than Combination.The more specific example of computer readable storage medium can include but is not limited to: have one or more conducting wires Electrical connection, portable computer diskette, hard disk, random access storage device (RAM), read-only memory (ROM), erasable type are programmable Read-only memory (EPROM or flash memory), optical fiber, portable compact disc read-only memory (CD-ROM), light storage device, magnetic are deposited Memory device or above-mentioned any appropriate combination.In embodiment of the disclosure, computer readable storage medium, which can be, appoints What include or the tangible medium of storage program that the program can be commanded execution system, device or device use or and its It is used in combination.And in embodiment of the disclosure, computer-readable signal media may include in a base band or as carrier wave The data-signal that a part is propagated, wherein carrying computer-readable program code.The data-signal of this propagation can be adopted With diversified forms, including but not limited to electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal is situated between Matter can also be any computer-readable medium other than computer readable storage medium, which can be with It sends, propagate or transmits for by the use of instruction execution system, device or device or program in connection.Meter The program code for including on calculation machine readable medium can transmit with any suitable medium, including but not limited to: electric wire, optical cable, RF (radio frequency) etc. or above-mentioned any appropriate combination.

Above-mentioned computer-readable medium can be included in above-mentioned electronic equipment；It is also possible to individualism, and not It is fitted into the electronic equipment.Above-mentioned computer-readable medium carries one or more program, when said one or more When a program is executed by the electronic equipment, so that the electronic equipment: obtaining target image；Target image is input to preparatory training Target object location model, obtain location information set, wherein the location information in location information set is used to indicate target The position of target object object in image, the position for the target object object that target object location model is used to determine that image includes It sets；Based on the location information in location information set, the key point information of the target object object in target image is determined；Output Obtained key point information.

The behaviour for executing embodiment of the disclosure can be write with one or more programming languages or combinations thereof The computer program code of work, described program design language include object oriented program language-such as Java, Smalltalk, C++ further include conventional procedural programming language-such as " C " language or similar program design language Speech.Program code can be executed fully on the user computer, partly be executed on the user computer, as an independence Software package execute, part on the user computer part execute on the remote computer or completely in remote computer or It is executed on server.In situations involving remote computers, remote computer can pass through the network of any kind --- packet It includes local area network (LAN) or wide area network (WAN)-is connected to subscriber computer, or, it may be connected to outer computer (such as benefit It is connected with ISP by internet).

Flow chart and block diagram in attached drawing are illustrated according to the system of the various embodiments of the disclosure, method and computer journey The architecture, function and operation in the cards of sequence product.In this regard, each box in flowchart or block diagram can generation A part of one module, program segment or code of table, a part of the module, program segment or code include one or more use The executable instruction of the logic function as defined in realizing.It should also be noted that in some implementations as replacements, being marked in box The function of note can also occur in a different order than that indicated in the drawings.For example, two boxes succeedingly indicated are actually It can be basically executed in parallel, they can also be executed in the opposite order sometimes, and this depends on the function involved.Also it to infuse Meaning, the combination of each box in block diagram and or flow chart and the box in block diagram and or flow chart can be with holding The dedicated hardware based system of functions or operations as defined in row is realized, or can use specialized hardware and computer instruction Combination realize.

Being described in unit involved in embodiment of the disclosure can be realized by way of software, can also be passed through The mode of hardware is realized.Described unit also can be set in the processor, for example, can be described as: a kind of processor Including acquiring unit, the first input unit, determination unit and the first output unit.Wherein, the title of these units is in certain feelings The restriction to the unit itself is not constituted under condition, for example, acquiring unit is also described as " obtaining the list of target image Member ".

Above description is only the preferred embodiment of the disclosure and the explanation to institute's application technology principle.Those skilled in the art Member is it should be appreciated that invention scope involved in the disclosure, however it is not limited to technology made of the specific combination of above-mentioned technical characteristic Scheme, while should also cover in the case where not departing from foregoing invention design, it is carried out by above-mentioned technical characteristic or its equivalent feature Any combination and the other technical solutions formed.Such as features described above has similar function with (but being not limited to) disclosed in the disclosure Can technical characteristic replaced mutually and the technical solution that is formed.

Claims

1. a kind of method for output information, comprising:

Obtain target image；

The target image is input to target object location model trained in advance, obtains location information set, wherein described Location information in location information set is used to indicate the position of the target object object in the target image, the object The position for the target object object that body location model is used to determine that image includes；

Based on the location information in the location information set, the key point of the target object object in the target image is determined Information；

Export obtained key point information.

2. according to the method described in claim 1, wherein, the location information based in the location information set determines The key point information of target object object in the target image, comprising:

Image-region at the position of location information instruction in the location information set is amplified into processing, obtains position The corresponding enlarged drawing region of information；

Obtained enlarged drawing region is sequentially input to key point trained in advance and determines model, obtains being located at location information The key point information of target object object at the position of instruction.

3. according to the method described in claim 2, wherein, the number of parameters and the key point of the target object location model Determine that the ratio of the number of parameters of model is greater than 10.

4. method described in one of -3 according to claim 1, wherein the number of parameters of the target object location model is greater than one Million.

5. method described in one of -3 according to claim 1, wherein the object of target object object instruction is vehicle, key point Position where the key point indicating number plate of information instruction.

6. according to the method described in claim 5, wherein, the method also includes:

Meet preset condition in response to the recognition result identified to key point, sends to target control equipment for prohibiting The only signal that vehicle belonging to the corresponding license plate of the key point is travelled, wherein the preset condition includes that recognition result refers to Show that license plate number belongs to default license plate number set.

7. method described in one of -3 according to claim 1, wherein the target image is the video in predetermined video Frame；And

The method also includes:

Video frame in the video after the target image is input to the target object location model, again To location information set；

For the location information in the location information set that obtains again, it is based on the location information, determines and is located at the location information The key point information and output of target object object at the position of instruction.

8. a kind of method for output information, comprising:

First equipment obtains target video；

First equipment selecting video frame from the target video is walked as target video frame, and following send of execution It is rapid: selected target video frame being input to target object location model trained in advance, obtains location information, wherein position The position for the target object object that confidence breath is used to indicate in video frame, the target object location model is for determining video frame The position for the target object object for including；In response to obtaining location information, obtained location information is sent to the second equipment；

Second equipment is in response to receiving location information, based on received location information, determines the mesh in video frame Mark the key point information and output of subject.

9. according to the method described in claim 8, wherein, the method also includes:

First equipment has been sent in response to the location information of the target object object in the target video frame of choosing this Finish, the video frame of unselected mistake is chosen from the target video as target video frame, continues to execute the sending step.

10. it is described based on received location information according to the method described in claim 8, wherein, it determines in video frame The key point information and output of target object object, comprising:

The image-region being located at the position that received location information indicates is amplified into processing, obtains location information pair The enlarged drawing region answered；

Obtained enlarged drawing region is input to key point trained in advance and determines model, obtains the position of location information instruction Set the key point information of the target object object at place.

11. according to the method described in claim 10, wherein, the number of parameters and the key of the target object location model Point determines that the ratio of the number of parameters of model is greater than 10.

12. the method according to one of claim 8-11, wherein the number of parameters of the target object location model is greater than 1000000.

13. the method according to one of claim 8-11, wherein the object of target object object instruction is vehicle, crucial Position where the key point indicating number plate of point information instruction.

14. according to the method for claim 13, wherein the method also includes:

Second equipment meets preset condition in response to the recognition result identified to key point, sets to target control Preparation send the signal for forbidding vehicle belonging to the corresponding license plate of the key point to be travelled, wherein the preset condition packet It includes recognition result indicating number plate number and belongs to default license plate number set.

15. a kind of device for output information, comprising:

Acquiring unit is configured to obtain target image；

First input unit is configured to for the target image being input to target object location model trained in advance, obtains Location information set, wherein the location information in the location information set is used to indicate the object in the target image The position of body object, the position for the target object object that the target object location model is used to determine that image includes；

Determination unit is configured to determine the mesh in the target image based on the location information in the location information set Mark the key point information of subject；

First output unit is configured to export obtained key point information.

16. a kind of system for output information is communicated to connect the system comprises the first equipment and with first equipment Second equipment, in which:

First equipment is configured to: obtaining target video；Selecting video frame is as target video from the target video Frame, and execute following sending step: selected target video frame is input to target object location model trained in advance, Obtain location information, wherein location information is used to indicate the position of the target object object in video frame, and the target object is fixed The position for the target object object that bit model is used to determine that video frame includes；In response to obtaining location information, by obtained position Confidence breath is sent to the second equipment；

Second equipment is configured to: in response to receiving location information, based on received location information, determining video The key point information and output of target object object in frame.

17. a kind of electronic equipment, comprising:

One or more processors；

Storage device is stored thereon with one or more programs,

When one or more of programs are executed by one or more of processors, so that one or more of processors are real The now method as described in any in claim 1-7.

18. a kind of computer-readable medium, is stored thereon with computer program, wherein real when described program is executed by processor The now method as described in any in claim 1-7.