Live Subject Analysis System for Application Target Identification

Abstract

The present invention relates to a live subject analysis system for application target identification. The system includes: a target identification device, which is arranged in a live test device and is used for sending a color analysis request when a human target exists in the collected live video, otherwise, sending a stop analysis signal; a third identification device, which is used for determining that the entity color of the human target is reliable when the overall Y-channel value, the overall U-channel value, and the overall V-channel value of the distribution pattern corresponding to the determined human target are respectively within the Y-channel value range, the U-channel value range, and the V-channel value range corresponding to the human target, otherwise, determining that the entity color of the human target is abnormal and triggering a human target color adjustment request. Through this system, it is possible to effectively identify whether the entity color of a human target is abnormal by customizing a color analysis mechanism based on multiple visualization data, thereby improving the test accuracy and efficiency of the live video.

Description

Technical Field

[0001] The present invention relates to the field of live broadcast testing, and more specifically, to a live broadcast subject analysis system for application target identification. Background Art

[0002] With the improvement of the marginal benefit brought by the continuous enrichment of the content and form of network live broadcasts, people are increasingly used to using live broadcasts to chat with people, learn makeup, interact with celebrities, and understand product information. Live broadcasts rely on intuitive video images to connect to communication systems that mainly rely on text and pictures, such as WeChat and Weibo. People can more intuitively contact the real other party through live broadcasts, which is a new platform and new space for network interpersonal communication.

[0003] An ideal live broadcast picture needs to continuously correct the live broadcast environment and facilities in the live broadcast room to obtain, for example, whether the physical color of the anchor in the live broadcast room meets the expected requirements. Obviously, this continuous debugging process is too cumbersome and complex. The key lies in the lack of an effective identification mechanism for whether the physical color of the human target is abnormal. Therefore, an effective identification mechanism for whether the physical color of the human target is abnormal is needed to verify whether the physical color of the anchor in the live broadcast room meets the expected requirements. Summary of the Invention

[0004] To solve the technical problems in related fields, the present invention provides a live broadcast subject analysis system for application target identification, which can, based on a number of visualized data selected specifically, perform high-precision identification of whether the physical color of the human target is abnormal through a customized color analysis mechanism, and complete on-site verification of whether the physical color of the anchor in the live broadcast room meets the expected requirements, thereby improving the test accuracy and efficiency of the live broadcast picture.

[0005] The present invention needs to have at least the following important inventive points:

[0006] The first point: Identify the distribution pattern of the human target in the finally optimized image received, obtain the Y-channel value, U-channel value, and V-channel value of each pixel point in the distribution pattern in the YUV space, and determine the overall Y-channel value, overall U-channel value, and overall V-channel value of the distribution pattern corresponding to the human target;

[0007] The second point: When the overall Y-channel value, overall U-channel value, and overall V-channel value of the distribution pattern corresponding to the determined human target are respectively within the Y-channel value range, U-channel value range, and V-channel value range corresponding to the human target, determine that the physical color of the human target is reliable; otherwise, determine that the physical color of the human target is abnormal and trigger a human target color adjustment request;

[0008] Third: A targeted image signal processing link including a band-stop filtering device, a curve processing device, and a morphological operation device is adopted to complete the targeted optimization processing of the live broadcast screen with human targets, providing higher-quality basic data for the subsequent color analysis of human targets.

[0009] According to the present invention, a live broadcast subject analysis system for application target identification is provided. The system includes:

[0010] A target identification device, arranged in a live broadcast test device, used for sending a color analysis request when there is a human target in the collected live broadcast screen, and also used for sending a stop analysis signal when there is no human target in the collected live broadcast screen;

[0011] A band-stop filtering device, arranged in the live broadcast test device and connected to the target identification device, used for performing band-stop filtering processing on the received live broadcast screen when receiving the color analysis request to obtain and output a corresponding band-stop filtered image;

[0012] A curve processing device, connected to the band-stop filtering device, used for adjusting the maximum curvature of each curve in the received band-stop filtered image to be lower than or equal to a preset curve maximum curvature threshold to obtain and output a corresponding curvature adjustment image;

[0013] A morphological operation device, connected to the curve processing device, used for performing a morphological operation of dilation first and then erosion on the received curve processing image to obtain and output a corresponding final optimized image;

[0014] A first recognition device, arranged in the live broadcast test device and connected to the morphological operation device, used for recognizing the distribution pattern of the human target in the received final optimized image, and obtaining the Y-channel value, U-channel value, and V-channel value of each pixel point in the YUV space in the distribution pattern;

[0015] A second recognition device, connected to the first recognition device, used for determining the overall Y-channel value, overall U-channel value, and overall V-channel value of the distribution pattern corresponding to the human target. The determined overall Y-channel value of the distribution pattern corresponding to the human target is the arithmetic mean of the respective Y-channel values corresponding to each pixel point in the distribution pattern corresponding to the human target, the determined overall U-channel value of the distribution pattern corresponding to the human target is the arithmetic mean of the respective U-channel values corresponding to each pixel point in the distribution pattern corresponding to the human target, and the determined overall V-channel value of the distribution pattern corresponding to the human target is the arithmetic mean of the respective V-channel values corresponding to each pixel point in the distribution pattern corresponding to the human target;

[0016] A third recognition device, connected to the second recognition device, is configured to determine that the entity color of the human target is reliable when the overall Y-channel value, the overall U-channel value, and the overall V-channel value of the distribution pattern corresponding to the determined human target are respectively within the Y-channel value range, the U-channel value range, and the V-channel value range corresponding to the human target; otherwise, it determines that the entity color of the human target is abnormal and triggers a human target color adjustment request.

[0017] The live subject analysis system for application target identification of the present invention operates stably and is intelligently designed. Since it can perform high-precision identification of whether the entity color of the human target is abnormal through a customized color analysis mechanism based on multiple visual data selected specifically, and complete the on-site verification of whether the entity color of the anchor in the live broadcast room meets the expected requirements, thereby improving the test accuracy and efficiency of the live broadcast screen and avoiding getting stuck in a complex and cumbersome test process. Brief Description of the Drawings

[0018] Those skilled in the art can better understand the numerous advantages of the present invention by referring to the accompanying drawings, in which:

[0019] Figure 1 is a schematic structural diagram of a live subject analysis system for application target identification according to the primary embodiment of the present invention.

[0020] Figure 2 is a schematic structural diagram of a live subject analysis system for application target identification according to the secondary embodiment of the present invention.

[0021] Figure 3 is a schematic structural diagram of a live subject analysis system for application target identification according to the tertiary embodiment of the present invention. Detailed Embodiments

[0022] Figure 1 is a schematic structural diagram of a live subject analysis system for application target identification according to the primary embodiment of the present invention, and the system includes:

[0023] A target identification device, disposed in a live test device, is configured to issue a color analysis request when a human target exists in the captured live broadcast screen, and is further configured to issue a stop analysis signal when no human target exists in the captured live broadcast screen;

[0024] Exemplarily, a target identification device, disposed in a live test device, is configured to issue a color analysis request when a human target exists in the captured live broadcast screen, and is further configured to issue a stop analysis signal when no human target exists in the captured live broadcast screen, including: identifying whether a human target exists in the captured live broadcast screen based on the gray imaging characteristics corresponding to the human body;

[0025] For example, a target identification device is provided inside the live test device. When a human target exists in the captured live video, it issues a color analysis request. When no human target exists in the captured live video, it issues a stop analysis signal. It further includes: when the occupied area of the human body recognized in the captured live video is less than or equal to the set area value, it determines that there is no human target in the captured live video; otherwise, there is a human target in the captured live video.

[0026] A band-stop filtering device is provided inside the live test device and is connected to the target identification device. When it receives the color analysis request, it performs band-stop filtering on the received live video to obtain and output the corresponding band-stop filtered image.

[0027] A curve processing device is connected to the band-stop filtering device. It adjusts the maximum curvature of each curve in the received band-stop filtered image to be lower than or equal to the preset curve maximum curvature threshold to obtain and output the corresponding curvature adjusted image.

[0028] A morphological operation device is connected to the curve processing device. It performs a morphological operation of dilation first and then erosion on the received curve processed image to obtain and output the corresponding final optimized image.

[0029] A first identification device is provided inside the live test device and is connected to the morphological operation device. It is used to identify the distribution pattern of the human target in the received final optimized image, and obtain the Y-channel value, U-channel value, and V-channel value of each pixel point in the YUV space in the distribution pattern.

[0030] A second identification device is connected to the first identification device. It is used to determine the overall Y-channel value, overall U-channel value, and overall V-channel value of the distribution pattern corresponding to the human target. The determined overall Y-channel value of the distribution pattern corresponding to the human target is the arithmetic average of the respective Y-channel values corresponding to each pixel point in the distribution pattern corresponding to the human target. The determined overall U-channel value of the distribution pattern corresponding to the human target is the arithmetic average of the respective U-channel values corresponding to each pixel point in the distribution pattern corresponding to the human target. And the determined overall V-channel value of the distribution pattern corresponding to the human target is the arithmetic average of the respective V-channel values corresponding to each pixel point in the distribution pattern corresponding to the human target.

[0031] A third identification device, connected to the second identification device, is configured to determine that the physical color of the human target is reliable when the overall Y-channel value, overall U-channel value, and overall V-channel value of the distribution pattern corresponding to the determined human target are respectively within the Y-channel value range, U-channel value range, and V-channel value range corresponding to the human target; otherwise, it determines that the physical color of the human target is abnormal and triggers a human target color adjustment request.

[0032] Among them, the target identification device is set inside the live test device and is configured to send a color analysis request when there is a human target in the captured live video, and is also configured to send a stop analysis signal when there is no human target in the captured live video, including: completing the identification process of whether there is a human target in the captured live video based on the gray value distribution range corresponding to the human target.

[0033] Figure 2 It is a schematic structural diagram of a live video analysis system for application target identification according to a secondary embodiment of the present invention.

[0034] And Figure 1 different, Figure 2 The live video analysis system for application target identification in

[0035] The humidity measurement mechanism includes multiple humidity measurement units and is configured to respectively measure the current surface humidity values of the morphological operation device, the first identification device, the second identification device, and the third identification device.

[0036] Among them, the humidity measurement mechanism includes multiple humidity measurement units and is configured to respectively measure the current surface humidity values of the morphological operation device, the first identification device, the second identification device, and the third identification device, including: the multiple humidity measurement units respectively adopted for the morphological operation device, the first identification device, the second identification device, and the third identification device are multiple non-contact humidity sensors.

[0037] Among them, the humidity measurement mechanism includes multiple humidity measurement units and is configured to respectively measure the current surface humidity values of the morphological operation device, the first identification device, the second identification device, and the third identification device, and further includes: the internal structures of the multiple non-contact humidity sensors respectively adopted for the morphological operation device, the first identification device, the second identification device, and the third identification device are the same.

[0038] Among them, the humidity measurement mechanism includes multiple humidity measurement units for respectively measuring the current surface humidity values of the morphological operation device, the first identification device, the second identification device, and the third identification device. It further includes: multiple non-contact humidity sensors respectively adopted for the morphological operation device, the first identification device, the second identification device, and the third identification device having the same humidity measurement upper limit threshold and humidity measurement lower limit threshold;

[0039] And among them, the humidity measurement mechanism includes multiple humidity measurement units for respectively measuring the current surface humidity values of the morphological operation device, the first identification device, the second identification device, and the third identification device. It further includes: the distances from the multiple non-contact humidity sensors respectively adopted for the morphological operation device, the first identification device, the second identification device, and the third identification device to the morphological operation device, the first identification device, the second identification device, and the third identification device are equal.

[0040] Figure 3 It is a schematic structural diagram of a live subject analysis system for application target identification according to a secondary embodiment of the present invention.

[0041] Different from Figure 1 the live subject analysis system for application target identification in Figure 3 may further include the following components:

[0042] An instant notification mechanism, respectively connected to the multiple non-contact humidity sensors respectively adopted for the morphological operation device, the first identification device, the second identification device, and the third identification device, and used for performing corresponding humidity alarm actions based on the humidity measurement results of the multiple non-contact humidity sensors respectively adopted for the morphological operation device, the first identification device, the second identification device, and the third identification device;

[0043] Among them, the instant notification mechanism, respectively connected to the multiple non-contact humidity sensors respectively adopted for the morphological operation device, the first identification device, the second identification device, and the third identification device, and used for performing corresponding humidity alarm actions based on the humidity measurement results of the multiple non-contact humidity sensors respectively adopted for the morphological operation device, the first identification device, the second identification device, and the third identification device includes: the instant notification mechanism internally includes a humidity receiving unit, a humidity judgment unit, and a notification execution unit;

[0044] And among them, the instant notification mechanism is respectively connected to a plurality of non-contact humidity sensors respectively adopted for the morphological operation device, the first identification device, the second identification device, and the third identification device, and is used to perform corresponding humidity alarm actions based on the humidity measurement results of the plurality of non-contact humidity sensors respectively adopted for the morphological operation device, the first identification device, the second identification device, and the third identification device. It further includes: within the instant notification mechanism, the humidity receiving unit, the humidity judging unit, and the notification execution unit are connected in sequence.

[0045] In addition, in the live subject analysis system for application target identification, the target identification device is arranged in the live test device and is used to send a color analysis request when a human target exists in the captured live video, and is also used to send a stop analysis signal when no human target exists in the captured live video. It further includes: the gray value distribution range corresponding to the human target is defined by the upper gray value threshold and the lower gray value threshold corresponding to the human target, and the upper gray value threshold is greater than the lower gray value threshold.

[0046] Although examples of the embodiments of the present invention have been shown and described, those skilled in the art should understand that various other modifications can be made without departing from the true scope of the present invention, and equivalents can be substituted. In addition, many modifications can be made without departing from the inventive concept described herein so as to adapt a particular situation to the teachings of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed, but the present invention includes all embodiments falling within the scope of the appended claims.

Claims

1. A live broadcast subject analysis system for application target identification, characterized in that: The system comprises: The target identification device is arranged in the live broadcast test device, and is used to issue a color analysis request when a human target exists in the collected live broadcast picture, and is also used to issue a stop analysis signal when a human target does not exist in the collected live broadcast picture; A band-stop filter device, arranged in the live broadcast test device and connected to the target identification device, is used to perform band-stop filtering processing on the received live broadcast picture when receiving the color analysis request, so as to obtain and output a corresponding band-stop filter image; A curve processing device, connected to the band-stop filtering device, for adjusting the maximum curvature of each curve in the received band-stop filtering image to be lower than or equal to a preset maximum curvature threshold of the curve, so as to obtain and output a corresponding curvature adjustment image; A morphological operation device, connected to the curve processing device, for performing a morphological operation of first dilation and then erosion on the received curvature adjustment image to obtain and output a corresponding final optimized image; A first recognition device is provided in the live broadcast test device and connected to the morphological operation device, and is used to recognize a distribution pattern of a human target in the received final optimized image, and obtain a Y channel value, a U channel value, and a V channel value of each pixel point in the distribution pattern in a YUV space; a second identification device, connected to the first identification device, and used to determine an overall Y channel value, an overall U channel value, and an overall V channel value of a distribution pattern corresponding to a human target, wherein the overall Y channel value of the distribution pattern corresponding to the human target determined is the arithmetic mean of each Y channel value corresponding to each pixel point in the distribution pattern corresponding to the human target, the overall U channel value of the distribution pattern corresponding to the human target determined is the arithmetic mean of each U channel value corresponding to each pixel point in the distribution pattern corresponding to the human target, and the overall V channel value of the distribution pattern corresponding to the human target determined is the arithmetic mean of each V channel value corresponding to each pixel point in the distribution pattern corresponding to the human target; A third identification device is connected to the second identification device, and is used to determine that the physical color of the human target is reliable when the overall Y channel value, the overall U channel value and the overall V channel value of the distribution pattern corresponding to the determined human target are respectively in the Y channel value interval, the U channel value interval and the V channel value interval corresponding to the human target; otherwise, it is determined that the physical color of the human target is abnormal and a human target color adjustment request is triggered.

2. The live broadcast subject analysis system for application target identification according to claim 1, characterized in that: The target identification device is arranged in the live broadcast test device, and is used to issue a color analysis request when there is a human target in the collected live broadcast picture, and is also used to issue a stop analysis signal when there is no human target in the collected live broadcast picture, including: completing the identification processing of whether there is a human target in the collected live broadcast picture based on the grayscale value distribution range corresponding to the human target.

3. The live broadcast subject analysis system for application target identification according to claim 2, characterized in that: The system further comprises: A humidity measuring mechanism, comprising a plurality of humidity measuring units, for respectively measuring current surface humidity values ​​of the morphological operation device, the first identification device, the second identification device, and the third identification device; Among them, the humidity measuring mechanism includes multiple humidity measuring units, which are used to respectively measure the current surface humidity values ​​of the morphological operation device, the first identification device, the second identification device and the third identification device, including: the multiple humidity measuring units respectively used for the morphological operation device, the first identification device, the second identification device and the third identification device are multiple non-contact humidity sensors.

4. The live broadcast subject analysis system for application target identification according to claim 3, characterized in that: The humidity measuring mechanism includes a plurality of humidity measuring units for respectively measuring the current surface humidity values ​​of the morphological operation device, the first identification device, the second identification device and the third identification device. The mechanism also includes: the internal structures of the plurality of non-contact humidity sensors respectively used for the morphological operation device, the first identification device, the second identification device and the third identification device are the same.

5. The live broadcast subject analysis system for application target identification as claimed in claim 3, characterized in that: The humidity measuring mechanism includes a plurality of humidity measuring units for respectively measuring the current surface humidity values ​​of the morphological operation device, the first identification device, the second identification device and the third identification device, and further includes: a plurality of non-contact humidity sensors respectively used for the morphological operation device, the first identification device, the second identification device and the third identification device have the same humidity measurement upper limit threshold and humidity measurement lower limit threshold.

6. The live broadcast subject analysis system for application target identification according to claim 5, characterized in that: The humidity measuring mechanism includes a plurality of humidity measuring units for respectively measuring the current surface humidity values ​​of the morphological operation device, the first identification device, the second identification device and the third identification device, and further includes: a plurality of non-contact humidity sensors respectively used for the morphological operation device, the first identification device, the second identification device and the third identification device are at equal distances from the morphological operation device, the first identification device, the second identification device and the third identification device.

7. The live broadcast subject analysis system for application target identification according to any one of claims 3 to 6, characterized in that: The system further comprises: The instant notification mechanism is respectively connected to multiple non-contact humidity sensors respectively used for the morphological operation device, the first identification device, the second identification device and the third identification device, and is used to perform corresponding humidity alarm actions based on the humidity measurement results of the multiple non-contact humidity sensors respectively used for the morphological operation device, the first identification device, the second identification device and the third identification device.

8. The live broadcast subject analysis system for application target identification according to claim 7, characterized in that: The instant notification mechanism has a built-in humidity receiving unit, a humidity judging unit and a notification executing unit.

9. The live broadcast subject analysis system for application target identification according to claim 8, characterized in that: An instant notification mechanism is respectively connected to a plurality of non-contact humidity sensors respectively used for the morphological operation device, the first identification device, the second identification device and the third identification device, and is used to perform corresponding humidity alarm actions based on the humidity measurement results of the plurality of non-contact humidity sensors respectively used for the morphological operation device, the first identification device, the second identification device and the third identification device. The instant notification mechanism also includes: within the instant notification mechanism, the humidity receiving unit, the humidity judgment unit and the notification execution unit are connected in sequence.

Images