Facial recognition system

Abstract

The invention relates to a face recognition system which comprises a face snapshot mechanism, a continuous mapping mechanism, a numerical value judgment device, a signal analysis device, a numerical value judgment device and the like. The face recognition system is compact in logic and easy and convenient to operate.

Description

Technical Field

[0001] The present invention relates to the field of live broadcast supervision, and in particular to a facial recognition system. Background Art

[0002] The core competitiveness of live streaming e-commerce is based on industrial advantages, genuine and good products and services. Sixiang Unlimited has cross-border anchor resources, and online celebrities can stimulate consumers' curiosity, thereby bringing high traffic and commodity transaction volume. It also actively explores the integration of live streaming bases with local large, medium, small and micro enterprises' supply chain, management chain, production chain, consumption chain and other value chains to achieve the sustainable development of the live streaming economy, and amplify the driving role of the live streaming economy as an image display end, marketing breakthrough end, commodity trading end, data collection end, and capital financing end. It builds a new paradigm of data-driven live streaming economy and activates the vitality of various market players in the upstream and downstream industrial chains of live streaming e-commerce, including live streaming e-commerce bases, live streaming platforms, MCN agencies, professional service agencies, etc. Summary of the invention

[0003] During the live broadcast process, excellent anchors are also the core competitiveness of live broadcast e-commerce. However, in order to ensure the visual effect, some anchors apply too thick foundation and perform excessive special effects processing, making the overall color of the anchor's face too uniform, which makes the anchor's face unrecognizable and in an overall blurred state. Therefore, it is necessary to test the anchor's facial effect to determine the live broadcast effect. Obviously, the existing technology lacks a targeted processing mechanism.

[0004] In order to solve the technical problems in the related fields, the present invention provides a facial recognition system, which determines each pixel of the host's face in a third mapping picture, obtains each R color component, each G color component and each B color component of each pixel in the RGB color space, determines the number of R color components with repeated values ​​in each R color component, the number of R color components with repeated values ​​in each G color component and the number of R color components with repeated values ​​in each B color component, and determines the distribution uniformity level corresponding to the host's face based on the maximum value of the three numbers, and plays a blur detection signal corresponding to the overall blur of the host's face when the received distribution uniformity level corresponding to the host's face is greater than or equal to a set level limit, and plays a clear detection signal corresponding to the overall recognition of the host's face when the received distribution uniformity level corresponding to the host's face is less than the set level limit, so that when it is analyzed that the overall color of the host's face is too uniform, it is judged that the host's face lacks recognition and is in an overall blurred state, thereby realizing a reliable targeted test of the live broadcast effect of the live broadcast room.

[0005] According to the present invention, there is provided a facial recognition system, the system comprising: A facial capture mechanism is provided in the live broadcast room and is used to capture the host's face before the live broadcast room starts broadcasting, so as to obtain and output a corresponding on-site captured image. The facial capture mechanism has the same working parameters as the video acquisition mechanism when the live broadcast room starts broadcasting; A continuous mapping mechanism, arranged in a live broadcast room and connected to the face capture mechanism, comprises a picture receiving component, a first mapping component, a second mapping component, a third mapping component and a picture output component, wherein the picture receiving component, the first mapping component, the second mapping component, the third mapping component and the picture output component are connected in sequence, the picture receiving component is connected to the face capture mechanism and is used to forward the live captured image to the first mapping component, the first mapping component is used to perform a median filtering process based on a dynamic filtering window on the received live captured image to remove salt and pepper noise in the live captured image to obtain a corresponding first mapping picture, the second mapping component is used to perform an impulse noise elimination process on the received first mapping picture to obtain a corresponding second mapping picture, the third mapping component is used to perform a gamma correction process on the received second mapping picture to obtain a corresponding third mapping picture, and the picture output component is used to output the third mapping picture; A numerical determination device is arranged in the live broadcast room and connected to the continuous mapping mechanism, comprising a block detection unit, a numerical analysis unit and an information determination unit, wherein the numerical analysis unit is respectively connected to the block detection unit and the information determination unit, and the numerical determination mechanism is used to determine each pixel of the host's face in the third mapping picture, obtain each R color component, each G color component and each B color component of each pixel in the RGB color space, determine the number of R color components with repeated values ​​in each R color component, the number of R color components with repeated values ​​in each G color component and the number of R color components with repeated values ​​in each B color component, and determine the distribution uniformity level corresponding to the host's face based on the maximum value of the three numbers, the smaller the value of the maximum value of the three numbers, the lower the distribution uniformity level corresponding to the determined host's face; The signal analysis device is connected to the numerical determination device and is used to play a blur detection signal corresponding to the host's face being too blurred as a whole when the received distribution uniformity level corresponding to the host's face is greater than or equal to the set level limit.

[0006] Therefore, the present invention has at least the following three beneficial technical effects: Technical effect 1: determining each pixel of the host's face in the third mapping picture, obtaining each R color component, each G color component and each B color component of each pixel in the RGB color space, determining the number of R color components with repeated values ​​in each R color component, the number of R color components with repeated values ​​in each G color component and the number of R color components with repeated values ​​in each B color component, and determining the distribution uniformity level corresponding to the host's face based on the maximum value of the three numbers, the smaller the value of the maximum value of the three numbers, the lower the distribution uniformity level corresponding to the determined host's face; Technical effect 2: A numerical determination mechanism with a customized structure including a block detection unit, a numerical analysis unit and an information determination unit is introduced. The numerical analysis unit is connected to the block detection unit and the information determination unit respectively, so as to provide reliable hardware resources for analyzing the distribution uniformity level corresponding to the host's face; Technical effect three: The signal analysis device is introduced to play the blur detection signal corresponding to the overall blur of the anchor's face when the received distribution uniformity level corresponding to the anchor's face is greater than or equal to the set level limit. It is also used to play the clear detection signal corresponding to the overall recognition of the anchor's face when the received distribution uniformity level corresponding to the anchor's face is less than the set level limit. Therefore, when it is analyzed that the overall color of the anchor's face is too uniform, it is judged that the anchor's face lacks recognition and is in an overall blurred state, thereby realizing reliable targeted testing of the live broadcast effect of the live broadcast room.

[0007] The facial recognition system of the present invention has compact logic and is easy to operate. By screening multiple targeted visual data related to the face of the broadcasting anchor, a customized detection is performed to determine whether the overall face of the broadcasting anchor is too blurred. When the overall color of the anchor's face is analyzed to be too uniform, it is judged that the anchor's face lacks recognition and is in an overall blurred state, thereby improving the test effect of the live broadcast room. DETAILED DESCRIPTION

[0008] The embodiment of the face recognition system of the present invention will be described in detail below.

[0009] The facial recognition system shown in the host face embodiment according to embodiment A of the present invention includes: A facial capture mechanism is provided in the live broadcast room and is used to capture the host's face before the live broadcast room starts broadcasting, so as to obtain and output a corresponding on-site captured image. The facial capture mechanism has the same working parameters as the video acquisition mechanism when the live broadcast room starts broadcasting; For example, a facial capture mechanism is provided in a live broadcast room, and is used to perform capture processing on the face of the anchor before the live broadcast room starts broadcasting, so as to obtain and output corresponding on-site captured images. The facial capture mechanism has the same working parameters as the video acquisition mechanism when the live broadcast room starts broadcasting, including: the facial capture mechanism includes a timing service unit, which is used to provide a timing service for performing capture processing on the face of the anchor before the live broadcast room starts broadcasting; A continuous mapping mechanism, arranged in a live broadcast room and connected to the face capture mechanism, comprises a picture receiving component, a first mapping component, a second mapping component, a third mapping component and a picture output component, wherein the picture receiving component, the first mapping component, the second mapping component, the third mapping component and the picture output component are connected in sequence, the picture receiving component is connected to the face capture mechanism and is used to forward the live captured image to the first mapping component, the first mapping component is used to perform a median filtering process based on a dynamic filtering window on the received live captured image to remove salt and pepper noise in the live captured image to obtain a corresponding first mapping picture, the second mapping component is used to perform an impulse noise elimination process on the received first mapping picture to obtain a corresponding second mapping picture, the third mapping component is used to perform a gamma correction process on the received second mapping picture to obtain a corresponding third mapping picture, and the picture output component is used to output the third mapping picture; A numerical determination device is arranged in the live broadcast room and connected to the continuous mapping mechanism, comprising a block detection unit, a numerical analysis unit and an information determination unit, wherein the numerical analysis unit is respectively connected to the block detection unit and the information determination unit, and the numerical determination mechanism is used to determine each pixel of the host's face in the third mapping picture, obtain each R color component, each G color component and each B color component of each pixel in the RGB color space, determine the number of R color components with repeated values ​​in each R color component, the number of R color components with repeated values ​​in each G color component and the number of R color components with repeated values ​​in each B color component, and determine the distribution uniformity level corresponding to the host's face based on the maximum value of the three numbers, the smaller the value of the maximum value of the three numbers, the lower the distribution uniformity level corresponding to the determined host's face; A signal analysis device is connected to the value determination device, and is used to play a blur detection signal corresponding to the host's face being too blurred as a whole when the received distribution uniformity level corresponding to the host's face is greater than or equal to the set level limit; The signal analysis device is used to play a clear detection signal corresponding to the overall recognition degree of the host's face when the distribution uniformity level corresponding to the received host's face is less than the set level limit; Among them, obtaining each R color component, each G color component and each B color component of each pixel in the RGB color space, and determining the number of R color components with repeated values ​​in each R color component, the number of R color components with repeated values ​​in each G color component and the number of R color components with repeated values ​​in each B color component include: the value of any color component among the R color component, G color component and B color component of each pixel is between 0-255.

[0010] Different from the embodiment A of the present invention, the facial recognition system in the embodiment B of the present invention may further include the following components: The pressure detection mechanism includes a plurality of pressure detection units, which are used to respectively detect the current surface pressure values ​​of the picture receiving component, the first mapping component, the second mapping component, the third mapping component and the picture output component; The pressure detection mechanism includes a plurality of pressure detection units for respectively detecting the current surface pressure values ​​of the picture receiving component, the first mapping component, the second mapping component, the third mapping component and the picture output component, including: the plurality of pressure detection units respectively used for the picture receiving component, the first mapping component, the second mapping component, the third mapping component and the picture output component are a plurality of contact pressure sensors; The pressure detection mechanism includes a plurality of pressure detection units for respectively detecting the current surface pressure values ​​of the picture receiving component, the first mapping component, the second mapping component, the third mapping component and the picture output component, and further includes: the internal structures of the plurality of contact pressure sensors respectively used for the picture receiving component, the first mapping component, the second mapping component, the third mapping component and the picture output component are the same; The pressure detection mechanism includes a plurality of pressure detection units for respectively detecting the current surface pressure values ​​of the picture receiving component, the first mapping component, the second mapping component, the third mapping component and the picture output component, and further includes: a plurality of non-contact pressure sensors respectively used for the picture receiving component, the first mapping component, the second mapping component, the third mapping component and the picture output component have the same upper pressure measurement threshold and lower pressure measurement threshold; And wherein, the pressure detection mechanism includes multiple pressure detection units, which are used to respectively detect the current surface pressure values ​​of the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component, and also includes: multiple contact pressure sensors respectively used for the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component all perform corresponding current surface pressure value detection based on the difference between the air pressure and the internal pressure of the detected device.

[0011] Different from the A embodiment of the present invention, the facial recognition system in the C embodiment of the present invention may further include the following components: A real-time warning mechanism is connected to the plurality of contact pressure sensors respectively used by the picture receiving component, the first mapping component, the second mapping component, the third mapping component and the picture output component, and is used to perform corresponding pressure warning operations based on the pressure measurement results of the plurality of contact pressure sensors respectively used by the picture receiving component, the first mapping component, the second mapping component, the third mapping component and the picture output component; Wherein, the real-time warning mechanism is respectively connected with a plurality of contact pressure sensors respectively used by the picture receiving component, the first mapping component, the second mapping component, the third mapping component and the picture output component, and is used to perform corresponding pressure warning operations based on the pressure measurement results of the plurality of contact pressure sensors respectively used by the picture receiving component, the first mapping component, the second mapping component, the third mapping component and the picture output component, including: the real-time warning mechanism includes a timing analysis unit and a warning operation unit; And wherein the real-time warning mechanism is respectively connected to a plurality of contact pressure sensors respectively used for the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component, and is used to perform corresponding pressure warning operations based on the pressure measurement results of the plurality of contact pressure sensors respectively used for the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component, further comprising: within the real-time warning mechanism, the timing analysis unit is connected to the warning operation unit; By way of example, the real-time warning mechanism is respectively connected to a plurality of contact pressure sensors respectively used for the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component, and is used to perform corresponding pressure warning operations based on the pressure measurement results of the plurality of contact pressure sensors respectively used for the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component, and also includes: the real-time warning mechanism is a photoelectric warning mechanism.

[0012] In addition, in the facial recognition system, obtaining each R color component, each G color component, and each B color component of each pixel in the RGB color space, and determining the number of R color components with repeated values ​​in each R color component, the number of R color components with repeated values ​​in each G color component, and the number of R color components with repeated values ​​in each B color component further comprises: completing simulation and testing of a data processing process of obtaining each R color component, each G color component, and each B color component of each pixel in the RGB color space, and determining the number of R color components with repeated values ​​in each R color component, the number of R color components with repeated values ​​in each G color component, and the number of R color components with repeated values ​​in each B color component in a numerical simulation mode; Specifically, obtaining each R color component, each G color component, and each B color component of each pixel in the RGB color space, determining the number of R color components with repeated values ​​in each R color component, the number of R color components with repeated values ​​in each G color component, and the number of R color components with repeated values ​​in each B color component also includes: the values ​​of the R color component, G color component, and B color component of each pixel in the RGB color space are all between 0 and 255.

[0013] From the foregoing it will be appreciated that although specific embodiments of the invention have been described herein for purposes of illustration, various modifications may be made without departing from the spirit and scope of the invention. Accordingly, the present invention is limited only by the appended claims.

Claims

1. A facial recognition system, characterized in that: The system comprises: A facial capture mechanism is provided in the live broadcast room and is used to capture the host's face before the live broadcast room starts broadcasting, so as to obtain and output a corresponding on-site captured image. The facial capture mechanism has the same working parameters as the video acquisition mechanism when the live broadcast room starts broadcasting; A continuous mapping mechanism, arranged in a live broadcast room and connected to the face capture mechanism, comprises a picture receiving component, a first mapping component, a second mapping component, a third mapping component and a picture output component, wherein the picture receiving component, the first mapping component, the second mapping component, the third mapping component and the picture output component are connected in sequence, the picture receiving component is connected to the face capture mechanism and is used to forward the live captured image to the first mapping component, the first mapping component is used to perform a median filtering process based on a dynamic filtering window on the received live captured image to remove salt and pepper noise in the live captured image to obtain a corresponding first mapping picture, the second mapping component is used to perform an impulse noise elimination process on the received first mapping picture to obtain a corresponding second mapping picture, the third mapping component is used to perform a gamma correction process on the received second mapping picture to obtain a corresponding third mapping picture, and the picture output component is used to output the third mapping picture; A numerical determination device is arranged in the live broadcast room and connected to the continuous mapping mechanism, comprising a block detection unit, a numerical analysis unit and an information determination unit, wherein the numerical analysis unit is respectively connected to the block detection unit and the information determination unit, and the numerical determination mechanism is used to determine each pixel of the host's face in the third mapping picture, obtain each R color component, each G color component and each B color component of each pixel in the RGB color space, determine the number of R color components with repeated values ​​in each R color component, the number of R color components with repeated values ​​in each G color component and the number of R color components with repeated values ​​in each B color component, and determine the distribution uniformity level corresponding to the host's face based on the maximum value of the three numbers, the smaller the value of the maximum value of the three numbers, the lower the distribution uniformity level corresponding to the determined host's face; The signal analysis device is connected to the numerical determination device and is used to play a blur detection signal corresponding to the host's face being too blurred as a whole when the received distribution uniformity level corresponding to the host's face is greater than or equal to the set level limit.

2. The facial recognition system according to claim 1, characterized in that: The signal analysis device is used to play a clear detection signal corresponding to the overall presence recognition of the host's face when the distribution uniformity level corresponding to the received host's face is less than the set level limit; Among them, obtaining each R color component, each G color component and each B color component of each pixel in the RGB color space, and determining the number of R color components with repeated values ​​in each R color component, the number of R color components with repeated values ​​in each G color component and the number of R color components with repeated values ​​in each B color component include: the value of any color component among the R color component, G color component and B color component of each pixel is between 0-255.

3. The facial recognition system according to claim 2, characterized in that The system further comprises: The pressure detection mechanism includes a plurality of pressure detection units, which are used to respectively detect the current surface pressure values ​​of the picture receiving component, the first mapping component, the second mapping component, the third mapping component and the picture output component; Among them, the pressure detection mechanism includes multiple pressure detection units, which are used to respectively detect the current surface pressure values ​​of the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component, including: the multiple pressure detection units respectively used for the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component are multiple contact pressure sensors.

4. The facial recognition system as claimed in claim 3, characterized in that: The pressure detection mechanism includes multiple pressure detection units, which are used to respectively detect the current surface pressure values ​​of the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component. It also includes: the internal structures of multiple contact pressure sensors respectively used for the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component are the same.

5. The facial recognition system as claimed in claim 3, characterized in that: The pressure detection mechanism includes multiple pressure detection units, which are used to respectively detect the current surface pressure values ​​of the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component. It also includes: multiple non-contact pressure sensors respectively used for the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component have the same pressure measurement upper limit threshold and pressure measurement lower limit threshold.

6. The facial recognition system as claimed in claim 5, characterized in that: The pressure detection mechanism includes multiple pressure detection units, which are used to respectively detect the current surface pressure values ​​of the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component. It also includes: multiple contact pressure sensors respectively used for the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component all perform corresponding current surface pressure value detection based on the difference between the air pressure and the internal pressure of the detected device.

7. The facial recognition system according to claim 2, characterized in that: The system further comprises: The real-time warning mechanism is respectively connected to multiple contact pressure sensors used by the picture receiving component, the first mapping component, the second mapping component, the third mapping component and the picture output component, and is used to perform corresponding pressure warning operations based on the pressure measurement results of the multiple contact pressure sensors used by the picture receiving component, the first mapping component, the second mapping component, the third mapping component and the picture output component.

8. The facial recognition system as claimed in claim 7, characterized in that: The real-time warning mechanism is respectively connected to multiple contact pressure sensors respectively used for the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component, and is used to perform corresponding pressure warning operations based on the pressure measurement results of the multiple contact pressure sensors respectively used for the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component, including: the real-time warning mechanism includes a timing analysis unit and a warning operation unit.

9. The facial recognition system as claimed in claim 8, characterized in that: The real-time warning mechanism is respectively connected to multiple contact pressure sensors used by the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component, and is used to perform corresponding pressure warning operations based on the pressure measurement results of the multiple contact pressure sensors used by the screen receiving component, the first mapping component, the second mapping component, the third mapping component and the screen output component, and also includes: within the real-time warning mechanism, the timing analysis unit is connected to the warning operation unit.