CN108876891B

CN108876891B - Face image data acquisition method and face image data acquisition device

Info

Publication number: CN108876891B
Application number: CN201711204796.2A
Authority: CN
Inventors: 马里千; 高端
Original assignee: Beijing Kuangshi Technology Co Ltd; Beijing Megvii Technology Co Ltd
Current assignee: Beijing Kuangshi Technology Co Ltd; Beijing Megvii Technology Co Ltd
Priority date: 2017-11-27
Filing date: 2017-11-27
Publication date: 2021-12-28
Anticipated expiration: 2037-11-27
Also published as: CN108876891A

Abstract

A face image data acquisition method and a face image data acquisition device are provided. The face image data acquisition method comprises the following steps: acquiring a virtual acquisition scene used for representing a real acquisition scene; obtaining at least one virtual head model data; and respectively putting the at least one piece of virtual head model data into a virtual acquisition scene for rendering so as to obtain at least one piece of face image data. The human face image data acquisition method enables the acquisition work to be no longer limited by the physical position of the person to be acquired, so that the resources and time consumed by the acquisition work can be reduced, and the cost can be saved.

Description

Face image data acquisition method and face image data acquisition device

Technical Field

The embodiment of the disclosure relates to a face image data acquisition method and a face image data acquisition device.

Background

The face image data is widely used in the fields of computer vision, artificial intelligence and the like at present as one of image data, and is used for improving the performance of face detection and recognition, face attribute detection and face key point detection. The construction of the face image acquisition system and the data acquisition itself consume a lot of resources and time and are limited by the physical location of the person to be acquired. Therefore, how to virtualize a real face image acquisition system becomes an important and urgent problem to be solved.

Disclosure of Invention

At least one embodiment of the present disclosure provides a method for acquiring face image data, including: acquiring a virtual acquisition scene used for representing a real acquisition scene; obtaining at least one virtual head model data; and respectively putting the at least one virtual head model data into the virtual acquisition scene for rendering so as to obtain at least one face image data.

For example, in a face image data acquisition method provided by an embodiment of the present disclosure, the acquiring a virtual acquisition scene representing a real acquisition scene includes: acquiring internal parameters of an image acquisition device in the real acquisition scene and position parameters of the image acquisition device in the real acquisition scene; acquiring environment information of the real acquisition scene; establishing an initialized virtual acquisition scene corresponding to the real acquisition scene according to the environmental information of the real acquisition scene and the internal parameters and the position parameters of the image acquisition device; and calibrating the initialized virtual acquisition scene to obtain the virtual acquisition scene.

For example, in the method for acquiring facial image data provided in an embodiment of the present disclosure, the establishing an initialized virtual acquisition scene corresponding to the real acquisition scene according to the environmental information of the real acquisition scene and the internal parameter and the position parameter of the image acquisition device includes: setting a virtual image acquisition device according to the internal parameters and the position parameters of the image acquisition device; and setting a virtual environment according to the environment information of the real acquisition scene to obtain the initialized virtual acquisition scene.

For example, in a method for acquiring face image data provided in an embodiment of the present disclosure, the calibrating the initialized virtual acquisition scene to obtain the virtual acquisition scene includes: acquiring respective real face image data of at least one acquired person in the real acquisition scene acquired by the image acquisition device; acquiring respective head model data of the at least one acquired person; and respectively putting the head model data of the at least one acquired person into the initialized virtual acquisition scene for rendering, and adjusting the brightness parameter of the initialized virtual acquisition scene until the face color distribution error of each face image data obtained by rendering and the corresponding real face image data is less than a preset value, so as to obtain the virtual acquisition scene.

For example, in a face image data acquisition method provided in an embodiment of the present disclosure, the acquiring environmental information of the real acquisition scene includes: and acquiring at least one of light source information, ambient illumination information and background information in the real acquisition scene.

For example, in a face image data acquisition method provided by an embodiment of the present disclosure, the light source information includes at least one of light source position information and light source angle information.

For example, in a method for acquiring face image data provided in an embodiment of the present disclosure, acquiring the light source position information or the light source angle information includes: establishing a three-dimensional coordinate system by taking the head center of the acquired person in the real acquisition scene as an origin, taking the direction of the connecting line of the head of the acquired person and the image acquisition device as a horizontal axis and taking the vertical direction as a vertical axis; and measuring the position or the angle of the light source in the three-dimensional coordinate system to obtain the light source position information or the light source angle information.

For example, in a face image data acquisition method provided in an embodiment of the present disclosure, the acquiring the ambient lighting information includes: and acquiring panoramic image data of the real acquisition scene, and acquiring the environmental illumination information according to the brightness data in the panoramic image data.

For example, in a method for acquiring face image data provided in an embodiment of the present disclosure, the acquiring the background information includes: acquiring a background image of the real acquisition scene acquired by the image acquisition device as background information; or acquiring panoramic image data of the real acquisition scene, and acquiring the background information according to the panoramic image data.

At least one embodiment of the present disclosure further provides a face image data acquisition device, which includes an acquisition unit and a processing unit. The acquisition unit is configured to acquire a virtual acquisition scene representing a real acquisition scene, and to acquire at least one virtual head model data. The processing unit is configured to place the at least one virtual head model data into the virtual acquisition scene for rendering, respectively, to obtain at least one facial image data.

For example, in a face image data acquisition device provided in an embodiment of the present disclosure, the acquisition unit is further configured to acquire internal parameters of an image acquisition device in a real acquisition scene and position parameters of the image acquisition device in the real acquisition scene, and acquire environment information of the real acquisition scene. The processing unit is further configured to establish an initialized virtual acquisition scene corresponding to the real acquisition scene according to the environmental information of the real acquisition scene and the internal parameters and the position parameters of the image acquisition device, and calibrate the initialized virtual acquisition scene to obtain the virtual acquisition scene.

For example, in a facial image data acquisition apparatus provided in an embodiment of the present disclosure, the processing unit, when establishing the initialization virtual acquisition scene, is further configured to: setting a virtual image acquisition device according to the internal parameters and the position parameters of the image acquisition device; and setting a virtual environment according to the environment information of the real acquisition scene to obtain the initialized virtual acquisition scene.

For example, in a facial image data acquisition apparatus provided by an embodiment of the present disclosure, the acquisition unit is further configured to acquire respective real facial image data of at least one subject located in the real acquisition scene acquired by the image acquisition apparatus, and acquire respective head model data of the at least one subject. The processing unit is further configured to respectively place the head model data of the at least one acquired person into the initialized virtual acquisition scene for rendering, and adjust the brightness parameter of the initialized virtual acquisition scene until the face color distribution error of each face image data obtained by rendering and the corresponding real face image data is less than a preset value, so as to obtain the virtual acquisition scene.

At least one embodiment of the present disclosure also provides a storage medium having stored thereon computer instructions adapted to be executed by a processor, the computer instructions, when executed by the processor, performing the following: acquiring a virtual acquisition scene used for representing a real acquisition scene; obtaining at least one virtual head model data; and respectively putting the at least one virtual head model data into the virtual acquisition scene for rendering so as to obtain at least one face image data.

At least one embodiment of the present disclosure further provides a facial image data acquisition apparatus, which includes a processor and a storage medium. The storage medium is configured to store computer instructions that are executable by the processor, and the computer instructions, when executed by the processor, perform the following: acquiring a virtual acquisition scene used for representing a real acquisition scene; obtaining at least one virtual head model data; and respectively putting the at least one virtual head model data into the virtual acquisition scene for rendering so as to obtain at least one face image data.

Drawings

To more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings of the embodiments will be briefly introduced below, and it is apparent that the drawings in the following description relate only to some embodiments of the present disclosure and are not limiting to the present disclosure.

Fig. 1 is a schematic diagram of a face image data acquisition method according to an embodiment of the present disclosure;

FIG. 2 is a schematic diagram of one example of method step S100 shown in FIG. 1;

FIG. 3 is a schematic diagram of one example of method step S130 shown in FIG. 2;

FIG. 4 is a schematic diagram of one example of method step S140 shown in FIG. 2;

FIG. 5 is a schematic diagram of a three-dimensional coordinate system established in an embodiment of the present disclosure;

fig. 6 is a schematic diagram of a specific example of a face image data acquisition method according to an embodiment of the present disclosure;

fig. 7 is a schematic diagram of a face image data acquisition device according to an embodiment of the present disclosure;

FIG. 8 is a schematic diagram of a storage medium provided by an embodiment of the present disclosure; and

fig. 9 is a schematic view of another facial image data acquisition device according to an embodiment of the present disclosure.

Detailed Description

To make the objects, technical solutions and advantages of the embodiments of the present disclosure more clear, the technical solutions of the embodiments of the present disclosure will be described below with reference to the accompanying drawings of the embodiments of the present disclosure. It is to be understood that the described embodiments are only a few embodiments of the present disclosure, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the described embodiments of the disclosure without any inventive step, are within the scope of protection of the disclosure.

Unless otherwise defined, technical or scientific terms used herein shall have the ordinary meaning as understood by one of ordinary skill in the art to which this disclosure belongs. The use of "first," "second," and similar terms in this disclosure is not intended to indicate any order, quantity, or importance, but rather is used to distinguish one element from another. Also, the use of the terms "a," "an," or "the" and similar referents do not denote a limitation of quantity, but rather denote the presence of at least one. The word "comprising" or "comprises", and the like, means that the element or item listed before the word covers the element or item listed after the word and its equivalents, but does not exclude other elements or items. The terms "connected" or "coupled" and the like are not restricted to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "upper", "lower", "left", "right", and the like are used merely to indicate relative positional relationships, and when the absolute position of the object being described is changed, the relative positional relationships may also be changed accordingly.

As one kind of image data, the face image data is widely used in the fields of computer vision, artificial intelligence, and the like, and is used to improve the performance of face detection and recognition, face attribute detection, and face key point detection, for example. To obtain a large amount of face image data, a face image acquisition system needs to be built and data acquisition needs to be performed, which consumes a large amount of resources and time. For example, in order to enrich the diversity of facial image data, it is usually necessary to acquire facial images of a large number of different acquired persons in many different real acquisition scenes, and at this time, each acquired person needs to go to each real acquisition scene respectively, and then acquire corresponding facial image data in the real acquisition scene. For example, if facial image data of 100 subjects in 100 real capturing scenes needs to be obtained, 100 × 100 capturing operations need to be performed. Therefore, how to virtualize a real face image acquisition system makes the acquisition work not limited by the physical location of the person to be acquired, which becomes an important and urgent problem to be solved.

At least one embodiment of the present disclosure provides a method for acquiring face image data, including: acquiring a virtual acquisition scene used for representing a real acquisition scene; obtaining at least one virtual head model data; and respectively putting the at least one piece of virtual head model data into a virtual acquisition scene for rendering so as to obtain at least one piece of face image data.

At least one embodiment of the present disclosure further provides a facial image data acquisition device.

The face image data acquisition method and the face image data acquisition device provided by the embodiment of the disclosure can enable the face image acquisition work to be no longer limited by the physical position of the person to be acquired, thereby reducing the resources and time consumed by the acquisition work, further saving the cost, and enabling the provision of a large amount of face image data to be possible.

Embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

An embodiment of the present disclosure provides a method for acquiring face image data, as shown in fig. 1, the method for acquiring face image data includes the following operations.

Step S100: acquiring a virtual acquisition scene used for representing a real acquisition scene;

step S200: obtaining at least one virtual head model data; and

step S300: and respectively putting the at least one piece of virtual head model data into a virtual acquisition scene for rendering so as to obtain at least one piece of face image data.

For example, with respect to step S100, in some embodiments, a certain real capture scene may be digitally virtualized to obtain a corresponding digital virtual capture scene, for example, an image of the real capture scene may be obtained by using a panoramic camera during digital virtualization. For another example, in other embodiments, the virtual capture scene representing a certain real capture scene may be directly acquired from, for example, a storage medium, for example, the virtual capture scene stored in the storage medium is acquired in advance. That is, when step S100 is executed, if a virtual captured scene representing a real captured scene is obtained in advance and stored in a storage medium, only direct call is needed in this step, and it is not necessary to perform virtualization operation on the real captured scene.

It should be noted that the real capturing scene in the embodiments of the present disclosure represents a specific real capturing scene, where the specific real capturing scene represents a scene photographed by the image capturing device facing a certain angle at a specific time, a specific place and a specific lighting environment. For example, one real acquisition scenario may be: a scene shot ten am towards south at an office doorway of a company. As another example, a real capture scenario may be: the scene shot towards the true north in the afternoon before the Beijing Temple of heaven. In the embodiment of the present disclosure, the real capture scene includes an indoor scene and also includes an outdoor scene, which is not limited in the embodiment of the present disclosure.

For example, for step S200, in some embodiments, a head model acquisition may be performed on at least one subject to obtain at least one virtual head model data. For another example, in other embodiments, the at least one virtual head model data may be retrieved directly from, for example, a storage medium, for example, where the at least one virtual head model data stored therein is pre-obtained. That is, when step S200 is executed, if at least one piece of virtual head model data is obtained in advance and stored in the storage medium, only direct calling is needed in this step, and there is no need to perform head model collection on at least one subject.

For example, when performing a head model acquisition on an acquired subject, in some embodiments, the head of the acquired subject may be laser scanned to obtain corresponding virtual head model data. For another example, virtual head model data may be obtained using an infrared lens or a dot matrix projector, a plurality of infrared light spots (invisible light) may be formed on a human face by the dot matrix projector, the human face may be photographed by an infrared camera, and a scanned face contour may be three-dimensionally modeled using dot matrix data obtained by the photographing. For another example, in other embodiments, the multi-angle head image of the subject may be obtained first, and then the corresponding virtual head model data may be obtained by using a three-dimensional reconstruction algorithm. The three-dimensional reconstruction algorithm used in the embodiments of the present disclosure may use an existing method, and is not described herein again.

For another example, in step S200, the virtual head model data may also be directly simulated by using a simulation computing device and combining with a simulation algorithm, as long as the algorithm can approximate to a real person. For example, the analog computing device may be implemented as a general purpose computing device such as a computer, or may be implemented as a special purpose computing device, which is not limited by the embodiments of the present disclosure.

For example, in step S300, at least one piece of virtual head model data acquired in step S200 may be respectively placed in the virtual captured scene acquired in step S100 for rendering to obtain at least one piece of face image data. For example, the virtual capture scene may be built into a rendering engine. For example, the rendering engine may employ an Unreal3D illusion engine, and the rendering engine employed in the embodiment of the present disclosure is not limited as long as it can implement establishment of a 3D rendering environment.

As described above, for example, if facial image data of 100 acquired persons in 100 real acquisition scenes need to be obtained, with the facial image data acquisition method provided in the embodiments of the present disclosure, virtual acquisition scenes for representing the 100 real acquisition scenes may be respectively obtained, head model data of the 100 acquired persons may be respectively obtained, and finally, the 100 head model data are respectively placed into the 100 virtual acquisition scenes for rendering, so as to obtain 100 × 100 facial image data. By adopting the method for acquiring the face image data, the acquisition work is not limited by the physical position of the acquired person any more, so that the resources and time consumed by the acquisition work can be reduced, and the cost can be saved.

In an embodiment of the present disclosure, the head model data includes a three-dimensional shape of the head, and may further include a texture of the head. The texture represents the color of each minimum area unit of the face area, and is a picture which can be expanded into a plane and contains color information. Texture may present skin details such as pores, wrinkles, lips, and the like.

It should be noted that, in the embodiment of the present disclosure, the virtual head model data represents the head model data put into the virtual acquisition scene at the time of virtual acquisition.

For example, as shown in fig. 2, in one example, the step S100 may include the following operations.

Step S110: acquiring internal parameters of an image acquisition device in a real acquisition scene and position parameters of the image acquisition device in the real acquisition scene;

step S120: acquiring environment information of a real acquisition scene;

step S130: establishing an initialized virtual acquisition scene corresponding to the real acquisition scene according to the environmental information of the real acquisition scene and the internal parameters and the position parameters of the image acquisition device; and

step S140: and calibrating the initialized virtual acquisition scene to obtain the virtual acquisition scene.

For example, in step S110, in the case that the image capturing device employs a camera, the internal parameters of the image capturing device may include, for example: aperture, exposure time, focal length, distance of the focal plane to the camera, and internal photosensitive element size, among others. For example, in the case that the image capturing device adopts other devices, the internal parameters may also include other parameters, which is not limited by the embodiments of the present disclosure.

For example, the internal parameters of the image capturing device may be pre-stored in the storage medium, and only need to be directly called from the storage medium when step S110 is executed.

In the embodiment of the present disclosure, the image capturing device includes, but is not limited to, a camera, a video camera, and the like, as long as it can capture an image in a real capturing scene (for example, including a human face image, a background image in the real capturing scene, and the like).

For a certain real acquisition scene, a three-dimensional coordinate system can be artificially established. For example, as shown in fig. 5, a three-dimensional coordinate system may be established with the head center of an object 10 in a real capturing scene as an origin O, the direction of the line connecting the head of the object 10 and the image capturing device 20 being an X axis (horizontal axis), the vertical direction being a Y axis (vertical axis), and the direction perpendicular to the X axis in the horizontal plane being a Z axis. For example, a left-hand coordinate system is shown in FIG. 5. It should be noted that the embodiment of the present disclosure is not limited to the manner of establishing the three-dimensional coordinate system shown in fig. 5, and for example, a cylindrical coordinate system, a spherical coordinate system, or the like may also be adopted.

For example, on the basis of the three-dimensional coordinate system, in step S110, the position parameters of the image capturing device in the real capturing scene may be directly measured by using a measuring tool such as a tape measure or a tape measure, or may be measured by using a distance sensor, or may be calculated by using a reference scale in the image and then using the captured image itself.

For example, in one example, step S120 includes acquiring at least one of light source information, ambient lighting information, and background information in the real capture scene. For example, in performing step S120, only a part of the environmental information, such as light source information, in the real capturing scene may be captured, and other environmental information, such as ambient lighting information, may adopt a default value or a preset value.

It should be noted that the light source described in the embodiments of the present disclosure includes a natural light source and an artificial light source. For example, natural light sources include sunlight and the like. For example, artificial light sources include incandescent lamps, LED lamps, and the like. For example, in an indoor setting, when an incandescent light bulb is turned on and sunlight is transmitted through a window, the light sources include the sun and the incandescent light bulb. For another example, in an outdoor scene, when artificial lighting is not employed, the light source includes only sunlight.

For example, the light source information includes at least one of light source position information and light source angle information. For example, on the basis of the three-dimensional coordinate system established as described above, the light source position information or the light source angle information may be obtained by measuring the position and angle of the light source in the three-dimensional coordinate system.

For example, when the light source only comprises an artificial light source, only the position information of the light source needs to be obtained, for example, in an indoor scene, when only one incandescent lamp works, the position information of the incandescent lamp can be obtained only by measuring on the basis of the established three-dimensional coordinate system and by using a measuring tool. For example, when the light source only includes a natural light source, only the light source angle information needs to be obtained, for example, in an indoor or outdoor scene, when there is no artificial illumination but sunlight, the light source angle information can be obtained only by measuring with a measuring tool on the basis of the established three-dimensional coordinate system, for example, the light source angle information can be measured according to the projection of an object.

For example, in one example, the acquiring of the ambient lighting information in the real capture scene includes: acquiring panoramic image data of a real acquisition scene, and acquiring environmental illumination information according to brightness data in the panoramic image data. For example, a wide-angle camera may be used to capture images in positive and negative directions of the X axis, the Y axis, and the Z axis, respectively, with the origin O shown in fig. 5 as the center, for example, six images may be captured, the captured results may be recorded as a 16-bit RAW file, and a panorama may be synthesized using a panorama stitching algorithm. Since the obtained panorama data includes the luminance data, the ambient illumination information of the real capture scene can be obtained from the luminance data in the panorama data.

In the embodiment of the present disclosure, the number of images captured by the wide-angle camera and the format of the recording file are not limited. In addition, the ambient lighting information described in the embodiments of the present disclosure represents brightness information of an environment in a real capture scene, and for example, the illumination provided by an incandescent lamp or sunlight may cause brightness to be different at different positions in the environment through multiple diffuse reflections in the real capture scene. The obtained ambient lighting information is used to characterize the brightness at different positions in the real captured scene.

In addition, the panorama data of the real captured scene may be stored in advance in, for example, a storage medium, and may be called directly when necessary.

For example, in one example, the acquiring the background information in the real capture scene includes: and acquiring a background image of a real acquisition scene acquired by the image acquisition device as background information. For example, in executing step S140, it is necessary to acquire real face image data of the person to be acquired in the real acquisition scene, for example, as shown in fig. 5, the image acquisition device 20 is used to acquire the real face image data of the person to be acquired under certain position parameters and internal parameters. When the capturing is completed, the captured person may be removed from the real capturing scene, and a background image of the real capturing scene is captured as background information without changing the position parameters and internal parameters of the image capturing apparatus 20. It should be noted that the background image of the real captured scene may also be stored in advance in, for example, a storage medium, and may be called directly when needed.

For example, in another example, the acquiring the context information in the real capture scene may further include: and acquiring panoramic image data of a real acquisition scene, and acquiring background information according to the panoramic image data. Since the panoramic image data already includes image data of six directions in the positive and negative directions of the X axis, the Y axis, and the Z axis in the real captured scene, a background image of the image capturing apparatus toward the direction of the captured person can be extracted from the panoramic image data. Reference may be made to the corresponding description above for how to obtain panorama data, which is not repeated here.

For example, as shown in fig. 3, in one example, step S130 may include the following operations.

Step S131: setting a virtual image acquisition device according to the internal parameters and the position parameters of the image acquisition device; and

step S132: and setting a virtual environment according to the environment information of the real acquisition scene to obtain an initialized virtual acquisition scene.

For example, in step S130, an initialization virtual capture scene may be established in one rendering engine. For example, the rendering engine may employ an Unreal3D illusion engine, and the rendering engine employed in the embodiment of the present disclosure is not limited as long as it can implement establishment of a 3D rendering environment.

For example, in step S131, the virtual image capturing device may be set in the rendering engine according to the internal parameters and the position parameters of the image capturing device obtained in step S110, that is, a virtual image capturing device similar to the real image capturing device is simulated in the rendering engine for performing the virtual capturing operation subsequently.

For example, in step S132, a virtual environment may be set in the rendering engine according to the environment information in the real capture scene acquired in step S120. For example, a virtual light source in the virtual environment may be set in the rendering engine according to the acquired light source information in the real acquisition scene; setting the ambient illumination of the virtual environment according to the acquired ambient illumination information of the real acquisition scene; and setting the background of the virtual environment according to the acquired background information of the real acquisition scene.

For example, as shown in fig. 4, in one example, step S140 may include the following operations.

Step S141: acquiring respective real face image data of at least one acquired person in a real acquisition scene acquired by an image acquisition device;

step S142: acquiring respective head model data of at least one acquired person; and

step S143: and respectively putting the respective head model data of at least one acquired person into the initialized virtual acquisition scene for rendering, and adjusting the brightness parameter of the initialized virtual acquisition scene until the face color distribution error of each face image data obtained by rendering and the corresponding real face image data is less than a preset value, so as to obtain the virtual acquisition scene.

For example, in step S141, as shown in fig. 5, for example, respective real face image data of at least one subject 10 located in a real capturing scene may be captured using the image capturing apparatus 20. It should be noted that the real face image data in step S141 may be stored in advance in, for example, a storage medium, and may be called directly when necessary.

In step S142, respective head model data of at least one subject in step 141 is acquired. For example, the head model data may be obtained by head model acquisition of respective heads of at least one subject. For another example, the respective head model data of the at least one subject may be directly obtained from, for example, a storage medium, for example, the respective head model data of the at least one subject stored in the storage medium is obtained by pre-acquisition.

In step S143, the head model data of each of the at least one subject acquired in step S142 may be placed in the initialized virtual acquisition scene established in step S130 and rendered, and the face image data corresponding to each head model data may be obtained through the rendering. Then, the brightness parameter of the initialized virtual captured scene in the rendering engine is gradually adjusted until the face color distribution error of each piece of face image data obtained by rendering and the corresponding real face image data (obtained in step S141) is smaller than a preset value, so as to obtain the virtual captured scene.

For example, when comparing the face image data obtained by rendering with the corresponding real face image data in step S143, the sum of squared errors on a pixel-by-pixel basis may be made smaller than a preset value, for example, a percentage with respect to a certain reference value. For example, the preset value may be specifically set according to actual conditions, and the preset value may be different in different real acquisition scenarios.

It should be noted that, when calibrating the initialized virtual capturing scene in step S140, the calibration may be performed by using the real face image data and the head model data of one captured person, or may be performed by using the real face images and the respective head model data of two, three, or more captured persons, which is not limited in this embodiment of the disclosure. Compared with the calibration by only using the data of one acquired person, the accuracy of the established virtual acquisition scene can be improved by using the data of a plurality of acquired persons for calibration.

In one example of the present disclosure, as shown in fig. 6, a specific example of a face image data acquisition method is provided. The method includes the operational steps shown in fig. 6.

For example, steps S11, S12, S13, S14, S15, and S16 are first performed using a real-world acquisition system (e.g., including an image acquisition device, a head model acquisition device, a measurement tool, etc.).

For example, in step S11, the real face image data of the person to be captured in the real capturing scene may be captured using a camera. In step S12, after the person to be captured is removed from the real capturing scene, the background image of the real capturing scene is captured without changing the internal parameters and the position parameters of the camera. In step S13, the position parameters of the camera in the real capture scene may be measured using the measurement tool. In step S14, the position or angle information of the light source may be measured using a measuring tool. In step S15, a wide-angle camera may be employed to capture a panorama of the real capture scene. In step S16, the head model acquisition device may be used to acquire head model data of the person to be acquired, for example, the head model acquisition device may include a 3D laser scanning device, a matrix camera, and the like.

For example, in step S17, an initialization virtual capture scene may be established in the rendering engine according to the data obtained in steps S12, S13, S14, S15. After the initialization virtual capture scene is established, the initialization virtual capture scene may be calibrated according to the data obtained in steps S11 and S16 in step S18, so that the virtual capture scene may be obtained.

For example, in step S20, the virtual head model data obtained in step S19 may be placed in the virtual captured scene obtained in step S18, and rendered to obtain face image data.

By adopting the method for acquiring the face image data, the acquisition work is not limited by the physical position of the acquired person any more, so that the resources and time consumed by the acquisition work can be reduced, and the cost can be saved.

An embodiment of the present disclosure also provides a facial image data acquisition apparatus 10, as shown in fig. 7, the facial image data acquisition apparatus 10 includes an acquisition unit 100 and a processing unit 200.

For example, the acquisition unit 100 is configured to acquire a virtual acquisition scene representing a real acquisition scene, and to acquire at least one virtual head model data. That is, the acquisition unit 100 is configured to execute step S100 and step S200 in the above-described embodiment. For example, the virtual capture scene and the virtual head model data may be obtained in advance and stored in a storage medium, and the obtaining unit 100 may directly call from the storage medium. For example, the storage medium may be disposed in the face image data acquisition apparatus 10 itself, or may be disposed elsewhere, and the embodiment of the disclosure is not limited thereto.

For example, the processing unit 200 is configured to place at least one virtual head model data into the virtual acquisition scene for rendering, respectively, to obtain at least one face image data. That is, the processing unit 200 is configured to execute step S300 in the above-described embodiment. For example, the processing unit 200 may include a rendering engine in which a virtual capture scene may be established.

The facial image data acquisition device provided by the embodiment of the disclosure can enable the acquisition work to be no longer limited by the physical position of the acquired person, thereby reducing the resources and time consumed by the acquisition work and further saving the cost.

For example, in one example of the embodiment of the present disclosure, the obtaining unit 100 is further configured to obtain internal parameters of the image capturing device in the real capturing scene and position parameters of the image capturing device in the real capturing scene, and obtain environment information of the real capturing scene. That is, the acquisition unit 100 may also be configured to perform step S110 and step S120 in the above-described embodiments. For the internal parameters and the position parameters of the image capturing device and the environment information of the real capturing scene, reference may be made to the corresponding descriptions in the above embodiments, which are not described herein again.

The processing unit 200 is further configured to establish an initialized virtual collecting scene corresponding to the real collecting scene according to the environment information of the real collecting scene and the internal parameters and the position parameters of the image collecting device, and calibrate the initialized virtual collecting scene to obtain the virtual collecting scene. That is, the processing unit 200 may also be configured to execute step S130 and step S140 in the above-described embodiments.

For example, in one example of an embodiment of the present disclosure, the processing unit 200, when establishing the initialization virtual acquisition scenario, is further configured to: and setting a virtual image acquisition device according to the internal parameters and the position parameters of the image acquisition device, and setting a virtual environment according to the environment information of the real acquisition scene to obtain an initialized virtual acquisition scene. That is, the processing unit 200 may also be configured to execute step S131 and step S132 in the above-described embodiments.

For example, in one example of the embodiment of the present disclosure, the obtaining unit 100 is further configured to obtain respective real face image data of at least one subject located in a real collecting scene, collected by the image collecting device, and obtain respective head model data of the at least one subject. That is, the acquisition unit 100 may also be configured to perform step S141 and step S142 in the above-described embodiments.

The processing unit 200 is further configured to respectively place the head model data of at least one acquired person into the initialized virtual acquisition scene for rendering, and adjust the brightness parameter of the initialized virtual acquisition scene until the face color distribution error of each face image data obtained by rendering and the corresponding real face image data is smaller than a preset value, so as to obtain the virtual acquisition scene. That is, the processing unit 200 may also be configured to execute step S143 in the above-described embodiment.

It should be noted that the acquiring unit 100 and the processing unit 200 in the facial image data collecting apparatus 10 provided in the embodiment of the present disclosure may be implemented to include an application specific integrated circuit, hardware (circuit), firmware, or any other combination manner to implement the desired function of each unit, for example, may be implemented as a digital signal processor, etc.

Alternatively, the acquisition unit 100 and the processing unit 200 in the face image data acquisition apparatus 10 may be implemented to include a processor and a storage medium configured to store computer instructions that are suitable for execution by the processor, and when executed by the processor, may implement the desired functions of each unit. The embodiments of the present disclosure are not limited in this regard.

An embodiment of the present disclosure also provides a storage medium 500, as shown in fig. 8, where the storage medium 500 stores thereon computer instructions 501 that may be adapted to be executed by a processor, and when the computer instructions 501 are executed by the processor, the following operations may be implemented:

acquiring a virtual acquisition scene used for representing a real acquisition scene;

obtaining at least one virtual head model data; and

and respectively putting the at least one piece of virtual head model data into a virtual acquisition scene for rendering so as to obtain at least one piece of face image data.

That is, the processor may perform step S100, step S200, and step S300 in the above-described embodiments.

For example, in one example, the storage medium 500 may be provided in a computing device, which may be configured to interface with a display device. For example, the computing device may also include a processor that may invoke computer instructions 501 stored in the storage medium 500.

An embodiment of the present disclosure also provides a human face image data acquisition device 10, as shown in fig. 9, the image processing device 10 includes a processor 600 and a storage medium 500. The storage medium 500 has stored thereon computer instructions 501 that may be adapted to be executed by the processor 600, and when executed by the processor 600, the computer instructions 501 may perform the following operations:

obtaining at least one virtual head model data; and

That is, the processor 600 may perform step S100, step S200, and step S300 in the above-described embodiments.

In the embodiments of the present disclosure, the processor may be implemented by a general-purpose integrated circuit chip or an application-specific integrated circuit chip, for example, the integrated circuit chip may be disposed on a motherboard, for example, the motherboard may also be disposed with a memory, a power circuit, and the like; further, a processor may also be implemented by circuitry, or in software, hardware (circuitry), firmware, or any combination thereof. In embodiments of the present disclosure, a processor may include various computing structures, such as a Complex Instruction Set Computer (CISC) structure, a Reduced Instruction Set Computer (RISC) structure, or one that implements a combination of instruction sets. In some embodiments, the processor may also be a microprocessor, such as an X86 processor or an ARM processor, or may be a Digital Signal Processor (DSP), or the like.

In the embodiments of the present disclosure, a storage medium may be provided on the motherboard, for example, and the storage medium may store instructions and/or data executed by the processor. For example, a storage medium may include one or more computer program products that may include various forms of computer-readable memory, such as volatile memory and/or non-volatile memory. The volatile memory may include, for example, Random Access Memory (RAM), cache memory (cache), and/or the like. The non-volatile memory may include, for example, Read Only Memory (ROM), magnetic disks, optical disks, semiconductor memory (e.g., flash memory, etc.), and so forth. On which one or more computer program instructions may be stored that a processor may execute to implement the desired functionality (implemented by the processor) in embodiments of the disclosure.

The above description is only a specific embodiment of the present disclosure, but the scope of the present disclosure is not limited thereto, and the scope of the present disclosure should be subject to the scope of the claims.

Claims

1. A face image data acquisition method comprises the following steps:

obtaining at least one virtual head model data;

respectively putting the at least one virtual head model data into the virtual acquisition scene for rendering so as to obtain at least one face image data;

wherein the acquiring a virtual capture scenario representing a real capture scenario comprises:

acquiring internal parameters of an image acquisition device in the real acquisition scene and position parameters of the image acquisition device in the real acquisition scene;

acquiring environment information of the real acquisition scene;

establishing an initialized virtual acquisition scene corresponding to the real acquisition scene according to the environmental information of the real acquisition scene and the internal parameters and the position parameters of the image acquisition device;

and calibrating the initialized virtual acquisition scene to obtain the virtual acquisition scene.

2. The method for acquiring facial image data according to claim 1, wherein the establishing an initialized virtual acquisition scene corresponding to the real acquisition scene according to the environmental information of the real acquisition scene and the internal parameters and the position parameters of the image acquisition device comprises:

setting a virtual image acquisition device according to the internal parameters and the position parameters of the image acquisition device;

and setting a virtual environment according to the environment information of the real acquisition scene to obtain the initialized virtual acquisition scene.

3. The method for acquiring facial image data according to claim 1 or 2, wherein the calibrating the initialized virtual acquisition scene to obtain the virtual acquisition scene comprises:

acquiring respective real face image data of at least one acquired person in the real acquisition scene acquired by the image acquisition device;

acquiring respective head model data of the at least one acquired person;

and respectively putting the head model data of the at least one acquired person into the initialized virtual acquisition scene for rendering, and adjusting the brightness parameter of the initialized virtual acquisition scene until the face color distribution error of each face image data obtained by rendering and the corresponding real face image data is less than a preset value, so as to obtain the virtual acquisition scene.

4. The method for acquiring facial image data according to claim 1 or 2, wherein the acquiring environmental information of the real acquisition scene comprises:

and acquiring at least one of light source information, ambient illumination information and background information in the real acquisition scene.

5. The method for acquiring facial image data according to claim 4, wherein the light source information comprises at least one of light source position information and light source angle information.

6. The method of claim 5, wherein the obtaining the light source position information or light source angle information comprises:

establishing a three-dimensional coordinate system by taking the head center of the acquired person in the real acquisition scene as an origin, taking the direction of the connecting line of the head of the acquired person and the image acquisition device as a horizontal axis and taking the vertical direction as a vertical axis;

and measuring the position or the angle of the light source in the three-dimensional coordinate system to obtain the light source position information or the light source angle information.

7. The facial image data collection method of claim 4, wherein the obtaining the ambient lighting information comprises:

and acquiring panoramic image data of the real acquisition scene, and acquiring the environmental illumination information according to the brightness data in the panoramic image data.

8. The method for acquiring facial image data according to claim 4, wherein the acquiring the background information comprises:

acquiring a background image of the real acquisition scene acquired by the image acquisition device as background information; or

And acquiring panoramic image data of the real acquisition scene, and acquiring the background information according to the panoramic image data.

9. A human face image data acquisition device comprises an acquisition unit and a processing unit, wherein,

the acquisition unit is configured to acquire a virtual acquisition scene representing a real acquisition scene, and to acquire at least one virtual head model data;

the processing unit is configured to place the at least one virtual head model data into the virtual acquisition scene for rendering respectively to obtain at least one face image data;

the acquiring unit is further configured to acquire internal parameters of an image acquiring device in a real acquiring scene and position parameters of the image acquiring device in the real acquiring scene, and acquire environment information of the real acquiring scene;

the processing unit is further configured to establish an initialized virtual acquisition scene corresponding to the real acquisition scene according to the environmental information of the real acquisition scene and the internal parameters and the position parameters of the image acquisition device, and calibrate the initialized virtual acquisition scene to obtain the virtual acquisition scene.

10. The facial image data collection apparatus of claim 9, wherein the processing unit, in establishing the initialization virtual capture scene, is further configured to:

setting a virtual image acquisition device according to the internal parameters and the position parameters of the image acquisition device; and

11. The facial image data collection apparatus according to claim 9 or 10,

the acquisition unit is further configured to acquire respective real face image data of at least one acquired person in the real acquisition scene acquired by the image acquisition device and acquire respective head model data of the at least one acquired person;

the processing unit is further configured to respectively place the head model data of the at least one acquired person into the initialized virtual acquisition scene for rendering, and adjust the brightness parameter of the initialized virtual acquisition scene until the face color distribution error of each face image data obtained by rendering and the corresponding real face image data is less than a preset value, so as to obtain the virtual acquisition scene.

12. A storage medium having stored thereon computer instructions adapted to be executed by a processor, the computer instructions when executed by the processor performing the following:

obtaining at least one virtual head model data;

acquiring environment information of the real acquisition scene;

13. A human face image data acquisition device comprises a processor and a storage medium, wherein,

the storage medium is configured to store computer instructions that are executable by the processor, and the computer instructions, when executed by the processor, perform the following:

obtaining at least one virtual head model data;

acquiring environment information of the real acquisition scene;