In the Internet industry which emphasizing the rapid iteration, more and more platforms are docking with third parties to shorten the R&D cycle in addition to self-developed technologies. Based on this, ZEGOCLOUD is centered on Real-Time Audio and Video tech, and cooperates with upstream and downstream technology providers to provide an integrated solution of “Audio and Video +”. Allowing Customers to integrate multiple functions at one time. This article will introduce content review.
Advantages of the “Audio and Video +” Integrated Solution.
1. Multi-party selection, complete capabilities, stable and reliable Technology
ZEGOCLOUD’s “Audio and Video +” solution can be quickly selected according to the needs of industry and customers, and flexibly connected to third parties. It can also be connected to multiple suppliers as an alternative for a certain function and achieve stable and high availability of technology and services.
2. Turn the tripartite joint modulation to two-party docking, make the integrated access faster and more efficient.
Customers only need to access ZEGOCLOUD services to achieve Real-Time Audio and Video functions, without multiple selections and repeated tests. Efficient and Convenient to reduce the cost of business, R&D and maintenance.
3. Multi-party integration to reduce repetitive billing and user costs.
ZEGOCLOUD constructs the connection between customer and the third party to help customers optimize the product structure, avoid repeated “wheel-making” in development, and reduce multiple billings and the user’s cost of use.
ZEGOCLOUD’s “Audio and Video +” integrated solution has accomplished the integration of content review, beautification stickers, interactive whiteboard and other functions. The following specifically introduces the content review function.
The Overall of “Audio and Video + Content Review”
Voice dating and video social networking which more lively and interesting interactive communication methods are widely popular among young people. But the feature of voice and video social platforms like strong interaction, heavy sharing, and high usage rate also give some criminals a chance to take advantage. Under increasingly strict supervision, many voice social apps have been rectified, and removed from app store.
Previously, platforms such as chat house, live video, and pan-entertainment social networking platforms, often needed to connect with third-party content review providers to help the platform filter out illegal content and reduce operational risks after integrated ZEGOCLOUD’s SDKs.
Now, through the integration of ZEGOCLOUD’s “Audio and Video + Content Review” integrated solution, users only need to “call the interface, initiate an authentication request, and wait for the callback result” three steps to achieve real-time voice and interactive video content review function.
1. Real-time voice content review function
ZEGOCLOUD’s real-time voice content review program can comprehensively detect breathless and other special voices, as well as political, pornographic, advertising and other illegal content through breathlessness recognition technology and RNN voice transfer technology.
- Docking with multiple technical reviewers, and the service is guaranteed. ZEGOCLOUD has docked with a number of third-party audit technology suppliers to improve the accuracy of detection in multiple dimensions and ensure high stability of services.
- Use the room as the audit unit to reduce costs. ZEGOCLOUD real-time voice review program is based on the room as a unit, and effectively reduces the cost of user content review in the scenario of multi-anchor interaction.
- Mature voice pre-processing to optimize the voice recognition effect. Based on ZEGOCLOUD audio processing engine, through noise suppression, volume gain and other algorithms, it reduces background sound and noise interference and improves the speech recognition rate.
2. Interactive video content review function
ZEGOCLOUD interactive video content function, which adopts the method of cutting frames to determine the video results, and supports the recognition of multiple sub-scenarios such as pornography, violence and terrorism, sensitive faces, and bad scenes.
- Content review + picture storage. In the video content review function, the “algorithm + manual” double review can effectively improve the accuracy, while the manual review needs to be reviewed through the intercepted pictures, which involves the storage function of the pictures. In ZEGOCLOUD function, the integration of “review + storage” has been achieved, and customers no longer need to connect to a third party for image storage.
- Review by stream, and lock specific accounts/hosts. ZEGOCLOUD video content review function is identified by stream as the unit, and the corresponding anchor can be found according to the stream ID, and accurate banning and investigation can be carried out.
ZEGOCLOUD’s “Audio and Video +” integrated solution, based on the open API interface, so that the product architecture can be freely extended, to achieve the docking of multiple functions. Against the background of current regulatory pressures and the ravages of black production, ZEGOCLOUD’s full set of content review programs architecture have integrated multiple identification algorithms, allowing customers to achieve the efficient integration of “Audio and Video + Content Review”.