Face recognition visual effect color correction strategy triggering system

Abstract

The invention relates to a face recognition visual effect color correction strategy triggering system. The system comprises a state switching mechanism which is used for triggering a visual acquisition mechanism of a set live broadcast room to be switched from a low-definition acquisition state to an ultra-high-definition acquisition state when an anchor face image exists in a live broadcast picture of the set live broadcast room, and otherwise, triggering the visual acquisition mechanism of the set live broadcast room to be switched from the ultra-high-definition acquisition state to the low-definition acquisition state; and the request analysis mechanism is used for requesting to trigger a visual special effect for performing color correction on the face of the anchor when the color mismatching signal is received, otherwise, postponing execution of the visual special effect for performing color correction on the face of the anchor. Through the system, color correction processing on the face of the anchor can be triggered only when it is identified that the face color of the anchor in the set live broadcast room is abnormal, so that the live broadcast visual effect of live broadcast key personnel is guaranteed, and meanwhile, tedious and complex continuous large-area image data processing is avoided.

Description

Technical Field

[0001] This invention relates to the field of image processing, and more specifically, to a color correction strategy triggering system for facial recognition visual effects. Background Technology

[0003] In the production of short videos, the quality of the anchor's image is one of the key factors determining the overall image quality of the short video. However, due to the inherent characteristics of the shooting location and the complexity of the shooting environment, the anchor's face in the shot is prone to color abnormalities, seriously affecting the viewer's experience and the overall image quality of the short video content. Patent publication CN109711306A discloses a method and device for obtaining facial features based on a deep convolutional neural network, including: acquiring facial images of multiple targets using a facial image acquisition device to obtain multiple training models; transmitting the facial images acquired by the facial image acquisition device to a deep convolutional neural network to classify the acquired facial images and detect whether the acquired facial images contain the required complete face; when it is determined that the facial image contains a complete face, segmenting the facial image according to a predetermined elliptical region to obtain an elliptical region image including the facial image; obtaining facial region sub-images; and classifying the facial region sub-images by facial color to obtain facial features. Summary of the Invention

[0004] To address technical issues in related fields, this invention provides a color correction strategy triggering system for facial recognition visual effects. This system issues a color matching signal when the arithmetic mean of the cyan component values ​​corresponding to each pixel on the streamer's face falls within the range of the cyan component values ​​corresponding to the streamer's face; the arithmetic mean of the magenta component values ​​corresponding to each pixel on the streamer's face falls within the range of the magenta component values ​​corresponding to the streamer's face; the arithmetic mean of the yellow component values ​​corresponding to each pixel on the streamer's face falls within the range of the yellow component values ​​corresponding to the streamer's face; and the arithmetic mean of the black component values ​​corresponding to each pixel on the streamer's face falls within the range of the black component values ​​corresponding to the streamer's face. Otherwise, it issues a color mismatch signal. Color correction processing for the streamer's face is only triggered when an abnormal color is detected in the designated live broadcast room. This ensures the visual effect of key personnel during the live broadcast while avoiding tedious and complex continuous processing of large-area image data.

[0005] Therefore, the system includes:

[0006] The state switching mechanism is used to trigger the visual acquisition mechanism of the set live room to switch from low-definition acquisition mode to ultra-high-definition acquisition mode when there is a host's facial image in the live broadcast screen of the set live room, and to trigger the visual acquisition mechanism of the set live room to switch from ultra-high-definition acquisition mode to low-definition acquisition mode when there is no host's facial image in the live broadcast screen of the set live room.

[0007] A dynamic processing device, connected to the state switching mechanism, is used to perform white balance processing based on a dynamic threshold on the received anchor's facial image to obtain and output the corresponding dynamic processed image.

[0008] A combined filtering device, connected to the dynamic processing device, is used to perform combined filtering processing on the received dynamic processing image to obtain and output a corresponding combined filtered image;

[0009] A gamma correction device, connected to the combined filtering device, is used to perform gamma correction processing on the received combined filtered image to obtain and output the corresponding gamma-corrected image.

[0010] The component processing unit receives the gamma-corrected image, analyzes the cyan, magenta, yellow, and black component values ​​corresponding to each pixel occupied by the anchor's face in the gamma-corrected image, and determines the arithmetic mean of the cyan, magenta, yellow, and black component values ​​corresponding to each pixel occupied by the anchor's face. A color matching signal is issued when the arithmetic mean of the values ​​of each cyan component corresponding to each pixel falls within the range of the values ​​of the cyan component corresponding to the anchor's face, the arithmetic mean of the values ​​of each magenta component corresponding to each pixel occupied by the anchor's face falls within the range of the values ​​of the magenta component corresponding to the anchor's face, the arithmetic mean of the values ​​of each yellow component corresponding to each pixel occupied by the anchor's face falls within the range of the values ​​of the yellow component corresponding to the anchor's face, and the arithmetic mean of the values ​​of each black component corresponding to each pixel occupied by the anchor's face falls within the range of the values ​​of the black component corresponding to the anchor's face; otherwise, a color mismatch signal is issued.

[0011] The request parsing mechanism, connected to the component processing mechanism, is used to request the triggering of a visual effect for color correction of the anchor's face when the color mismatch signal is received, and is also used to postpone the execution of the visual effect for color correction of the anchor's face when the color matching signal is received.

[0012] The facial recognition visual effect color correction strategy triggering system of this invention is compact in design and stable in operation. Because it can trigger color correction processing on the anchor's face only when an abnormality in the anchor's facial color is detected, based on targeted image processing, it avoids tedious and complex continuous large-area image data processing while ensuring the visual effect of key personnel in the live broadcast. Brief description of the attached figures

[0013] Those skilled in the art can better understand the many advantages of the present invention by referring to the accompanying drawings.

[0014] Figure 1 This is a schematic diagram of the structure of a face recognition visual effect color correction strategy triggering system according to the primary embodiment of the present invention.

[0015] Figure 2 This is a schematic diagram of the structure of a face recognition visual effect color correction strategy triggering system according to a secondary embodiment of the present invention.

[0016] Figure 3 This is a schematic diagram of the structure of a face recognition visual effect color correction strategy triggering system according to a further embodiment of the present invention. Detailed Implementation

[0017] Figure 1 This is a schematic diagram of a face recognition visual effect color correction strategy triggering system according to a primary embodiment of the present invention, the system comprising:

[0018] The state switching mechanism is used to trigger the visual acquisition mechanism of the set live room to switch from low-definition acquisition mode to ultra-high-definition acquisition mode when there is a host's facial image in the live broadcast screen of the set live room, and to trigger the visual acquisition mechanism of the set live room to switch from ultra-high-definition acquisition mode to low-definition acquisition mode when there is no host's facial image in the live broadcast screen of the set live room.

[0019] For example, the state switching mechanism is used to trigger the visual acquisition mechanism of the set live room to switch from low-definition acquisition state to ultra-high-definition acquisition state when there is a host's facial image in the live broadcast screen of the set live room, and to trigger the visual acquisition mechanism of the set live room to switch from ultra-high-definition acquisition state to low-definition acquisition state when there is no host's facial image in the live broadcast screen of the set live room, including: performing recognition of the presence of the host's facial image in the live broadcast screen of the set live room based on human facial imaging features.

[0020] A dynamic processing device, connected to the state switching mechanism, is used to perform white balance processing based on a dynamic threshold on the received anchor's facial image to obtain and output the corresponding dynamic processed image.

[0021] A combined filtering device, connected to the dynamic processing device, is used to perform combined filtering processing on the received dynamic processing image to obtain and output a corresponding combined filtered image;

[0022] A gamma correction device, connected to the combined filtering device, is used to perform gamma correction processing on the received combined filtered image to obtain and output the corresponding gamma-corrected image.

[0023] The component processing unit receives the gamma-corrected image, analyzes the cyan, magenta, yellow, and black component values ​​corresponding to each pixel occupied by the anchor's face in the gamma-corrected image, and determines the arithmetic mean of the cyan, magenta, yellow, and black component values ​​corresponding to each pixel occupied by the anchor's face. A color matching signal is issued when the arithmetic mean of the values ​​of each cyan component corresponding to each pixel falls within the range of the values ​​of the cyan component corresponding to the anchor's face, the arithmetic mean of the values ​​of each magenta component corresponding to each pixel occupied by the anchor's face falls within the range of the values ​​of the magenta component corresponding to the anchor's face, the arithmetic mean of the values ​​of each yellow component corresponding to each pixel occupied by the anchor's face falls within the range of the values ​​of the yellow component corresponding to the anchor's face, and the arithmetic mean of the values ​​of each black component corresponding to each pixel occupied by the anchor's face falls within the range of the values ​​of the black component corresponding to the anchor's face; otherwise, a color mismatch signal is issued.

[0024] The request parsing mechanism, connected to the component processing mechanism, is used to request the triggering of a visual effect to correct the color of the anchor's face when the color mismatch signal is received, and is also used to postpone the execution of the visual effect to correct the color of the anchor's face when the color matching signal is received.

[0025] The analysis of the cyan, magenta, yellow, and black component values ​​corresponding to each pixel in the gamma-corrected image of the anchor's face includes the following: the value of any one of the cyan, magenta, yellow, and black component values ​​corresponding to each pixel is between 0 and 255.

[0026] Figure 2 This is a schematic diagram of the structure of a face recognition visual effect color correction strategy triggering system according to a secondary embodiment of the present invention.

[0027] and Figure 1 different, Figure 2The facial recognition visual effect color correction strategy triggering system may also include the following components:

[0028] The humidity measurement mechanism includes multiple humidity measurement units, which are used to measure the current surface humidity values ​​of the combined filtering device, the gamma correction device, the component processing mechanism, and the request parsing mechanism, respectively.

[0029] The humidity measurement mechanism includes multiple humidity measurement units for measuring the current surface humidity values ​​of the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism, respectively. The multiple humidity measurement units used by the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism are multiple non-contact humidity sensors.

[0030] The humidity measurement mechanism includes multiple humidity measurement units for measuring the current surface humidity values ​​of the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism, respectively. It also includes multiple non-contact humidity sensors with identical internal structures used by the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism, respectively.

[0031] The humidity measurement mechanism includes multiple humidity measurement units for measuring the current surface humidity values ​​of the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism, respectively. It also includes multiple non-contact humidity sensors used by the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism, each having the same upper and lower humidity measurement thresholds.

[0032] The humidity measurement mechanism includes multiple humidity measurement units for measuring the current surface humidity values ​​of the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism, respectively. It also includes multiple non-contact humidity sensors used by the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism, each at an equal distance from the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism.

[0033] Figure 3 This is a schematic diagram of the structure of a face recognition visual effect color correction strategy triggering system according to a further embodiment of the present invention.

[0034] and Figure 1 different, Figure 3 The facial recognition visual effect color correction strategy triggering system may also include the following components:

[0035] The instant notification mechanism is connected to multiple non-contact humidity sensors used in the combined filtering device, the gamma correction device, the component processing mechanism, and the request parsing mechanism, respectively, and is used to execute corresponding humidity alarm actions based on the humidity measurement results of the multiple non-contact humidity sensors used in the combined filtering device, the gamma correction device, the component processing mechanism, and the request parsing mechanism, respectively.

[0036] The instant notification mechanism is connected to multiple non-contact humidity sensors used in the combined filtering device, the gamma correction device, the component processing mechanism, and the request parsing mechanism, respectively. It is used to perform corresponding humidity alarm actions based on the humidity measurement results of the multiple non-contact humidity sensors used in the combined filtering device, the gamma correction device, the component processing mechanism, and the request parsing mechanism. The instant notification mechanism includes a built-in humidity receiving unit, a humidity judgment unit, and a notification execution unit.

[0037] The instant notification mechanism, which is connected to multiple non-contact humidity sensors used by the combined filtering device, the gamma correction device, the component processing mechanism, and the request parsing mechanism, and is used to perform corresponding humidity alarm actions based on the humidity measurement results of the multiple non-contact humidity sensors used by the combined filtering device, the gamma correction device, the component processing mechanism, and the request parsing mechanism, further includes: within the instant notification mechanism, the humidity receiving unit, the humidity judging unit, and the notification execution unit are connected in sequence.

[0038] In addition, in the face recognition visual effect color correction strategy triggering system, the state switching mechanism is used to trigger the visual acquisition mechanism of the set live room to switch from low-definition acquisition state to ultra-high-definition acquisition state when there is a host's facial image in the live broadcast screen of the set live room, and to trigger the visual acquisition mechanism of the set live room to switch from ultra-high-definition acquisition state to low-definition acquisition state when there is no host's facial image in the live broadcast screen of the set live room. This includes: recognizing whether there is a host's facial image by examining the preview image of the live broadcast screen of the set live room.

[0039] According to the embodiments of the present invention, the present invention has at least the following three key technical points:

[0040] The first step is to trigger color correction processing of the anchor's face only when abnormal facial color is detected in the live broadcast room. This ensures the visual effect of key personnel in the live broadcast while avoiding tedious and complex continuous large-area image data processing.

[0041] The second step involves receiving a gamma-corrected image, analyzing the cyan, magenta, yellow, and black component values ​​corresponding to each pixel occupied by the anchor's face in the gamma-corrected image, and determining the arithmetic mean of the cyan, magenta, yellow, and black component values ​​corresponding to each pixel occupied by the anchor's face. This provides a basis for judging whether the anchor's face color is distorted.

[0042] Thirdly: A color matching signal is issued when the arithmetic mean of the values ​​of the cyan components corresponding to each pixel occupied by the anchor's face falls within the range of the values ​​of the cyan components corresponding to the anchor's face, the arithmetic mean of the values ​​of the magenta components corresponding to each pixel occupied by the anchor's face falls within the range of the values ​​of the magenta components corresponding to the anchor's face, the arithmetic mean of the values ​​of the yellow components corresponding to each pixel occupied by the anchor's face falls within the range of the values ​​of the yellow components corresponding to the anchor's face, and the arithmetic mean of the values ​​of the black components corresponding to each pixel occupied by the anchor's face falls within the range of the values ​​of the black components corresponding to the anchor's face; otherwise, a color mismatch signal is issued.

[0043] It should be understood from the above description that the invention and its many accompanying advantages will be apparent, and that various changes can be made in the form, structure, and arrangement of the components without departing from the subject matter or sacrificing all its substantial advantages. The form of the description is merely illustrative.

[0044] Those skilled in the art will recognize that the state of this art can include advancements to points where minor differences may remain between hardware, software, and / or firmware implementations in various aspects of a system; the use of hardware, software, and / or firmware is generally (but not always, in some cases where the choice between hardware and software may become important) a design choice representing a cost-efficiency trade-off. Those skilled in the art will understand that various carriers (e.g., hardware, software, and / or firmware) may exist that can influence the processes and / or systems and / or other technologies described herein, and the preferred carrier will vary depending on the circumstances under which the processes and / or systems and / or other technologies can be deployed. For example, if the implementer determines that speed and accuracy may be of paramount importance, the implementer may choose a primary hardware and / or firmware carrier; or, if flexibility may be of paramount importance, the implementer may choose a primary software implementation; or, alternatively, the implementer may choose some combination of hardware, software, and / or firmware. Therefore, there may be several possible carriers through which the processes and / or systems and / or other technologies described herein can be influenced, none of which is inherently superior to the others. Any carrier to be utilized may be chosen based on the circumstances under which the carrier will be deployed and the specific concerns of the implementer (e.g., speed, flexibility, or predictability), any of which may vary. Those skilled in the art will recognize that the optical aspects of the implementation will typically utilize optically guided hardware, software, and / or firmware.

Claims

1. A face recognition visual effect color correction strategy triggering system, characterized in that, The system includes: The state switching mechanism is used to trigger the visual acquisition mechanism of the set live room to switch from low-definition acquisition mode to ultra-high-definition acquisition mode when there is a host's facial image in the live broadcast screen of the set live room, and to trigger the visual acquisition mechanism of the set live room to switch from ultra-high-definition acquisition mode to low-definition acquisition mode when there is no host's facial image in the live broadcast screen of the set live room. A dynamic processing device, connected to the state switching mechanism, is used to perform white balance processing based on a dynamic threshold on the received anchor's facial image to obtain and output the corresponding dynamic processed image. A combined filtering device, connected to the dynamic processing device, is used to perform combined filtering processing on the received dynamic processing image to obtain and output a corresponding combined filtered image; A gamma correction device, connected to the combined filtering device, is used to perform gamma correction processing on the received combined filtered image to obtain and output the corresponding gamma-corrected image. The component processing unit receives the gamma-corrected image, analyzes the cyan, magenta, yellow, and black component values ​​corresponding to each pixel occupied by the anchor's face in the gamma-corrected image, and determines the arithmetic mean of the cyan, magenta, yellow, and black component values ​​corresponding to each pixel occupied by the anchor's face. A color matching signal is issued when the arithmetic mean of the values ​​of each cyan component corresponding to each pixel falls within the range of the values ​​of the cyan component corresponding to the anchor's face, the arithmetic mean of the values ​​of each magenta component corresponding to each pixel occupied by the anchor's face falls within the range of the values ​​of the magenta component corresponding to the anchor's face, the arithmetic mean of the values ​​of each yellow component corresponding to each pixel occupied by the anchor's face falls within the range of the values ​​of the yellow component corresponding to the anchor's face, and the arithmetic mean of the values ​​of each black component corresponding to each pixel occupied by the anchor's face falls within the range of the values ​​of the black component corresponding to the anchor's face; otherwise, a color mismatch signal is issued. The request parsing mechanism, connected to the component processing mechanism, is used to request the triggering of a visual effect for color correction of the anchor's face when the color mismatch signal is received, and is also used to postpone the execution of the visual effect for color correction of the anchor's face when the color matching signal is received.

2. The face recognition visual effect color correction strategy triggering system as described in claim 1, characterized in that: The analysis of the cyan, magenta, yellow, and black component values ​​corresponding to each pixel in the gamma-corrected image of the anchor's face includes: the value of any one of the cyan, magenta, yellow, and black component values ​​corresponding to each pixel is between 0 and 255.

3. The face recognition visual effect color correction strategy triggering system of claim 2, wherein, The system also includes: The humidity measurement mechanism includes multiple humidity measurement units, which are used to measure the current surface humidity values ​​of the combined filtering device, the gamma correction device, the component processing mechanism, and the request parsing mechanism, respectively. The humidity measurement mechanism includes multiple humidity measurement units for measuring the current surface humidity values ​​of the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism, respectively. The multiple humidity measurement units used by the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism are multiple non-contact humidity sensors.

4. The face recognition visual effect color correction strategy triggering system as described in claim 3, characterized in that: The humidity measurement mechanism includes multiple humidity measurement units for measuring the current surface humidity values ​​of the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism, respectively. It also includes multiple non-contact humidity sensors with identical internal structures used in the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism, respectively.

5. The face recognition visual effect color correction strategy triggering system as described in claim 3, characterized in that: The humidity measurement mechanism includes multiple humidity measurement units for measuring the current surface humidity values ​​of the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism, respectively. It also includes multiple non-contact humidity sensors used by the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism, each having the same upper and lower humidity measurement thresholds.

6. The face recognition visual effect color correction strategy triggering system as described in claim 5, characterized in that: The humidity measurement mechanism includes multiple humidity measurement units for measuring the current surface humidity values ​​of the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism, respectively. It also includes multiple non-contact humidity sensors used by the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism, each at an equal distance from the combined filtering device, the gamma correction device, the component processing mechanism, and the request analysis mechanism.

7. The face recognition visual effect color correction strategy triggering system as described in any one of claims 3-6, characterized in that, The system also includes: The instant notification mechanism is connected to multiple non-contact humidity sensors used in the combined filtering device, the gamma correction device, the component processing mechanism, and the request parsing mechanism, respectively, and is used to execute corresponding humidity alarm actions based on the humidity measurement results of the multiple non-contact humidity sensors used in the combined filtering device, the gamma correction device, the component processing mechanism, and the request parsing mechanism.

8. The face recognition visual effect color correction strategy triggering system as described in claim 7, characterized in that: The instant notification mechanism is connected to multiple non-contact humidity sensors used in the combined filtering device, the gamma correction device, the component processing mechanism, and the request parsing mechanism, respectively. It is used to perform corresponding humidity alarm actions based on the humidity measurement results of the multiple non-contact humidity sensors used in the combined filtering device, the gamma correction device, the component processing mechanism, and the request parsing mechanism, respectively. The instant notification mechanism includes a built-in humidity receiving unit, a humidity judgment unit, and a notification execution unit.

9. The face recognition visual effect color correction strategy triggering system as described in claim 8, characterized in that: The instant notification mechanism, which is connected to multiple non-contact humidity sensors respectively used in the combined filtering device, the gamma correction device, the component processing mechanism, and the request parsing mechanism, and is used to execute corresponding humidity alarm actions based on the humidity measurement results of the multiple non-contact humidity sensors respectively used in the combined filtering device, the gamma correction device, the component processing mechanism, and the request parsing mechanism, further includes: within the instant notification mechanism, the humidity receiving unit, the humidity judging unit, and the notification execution unit are sequentially connected.

Images