Live broadcast is in the limelight, how should the platform do a good job of content review to avoid risks?

China's Internet content has undergone two major changes, from text to video, and finally to live broadcast.

It has to be said that content and content creators are closer to users, and real-time is stronger. However, another risk also arises ---

How to ensure content security?

The text to the video are uploaded first and then reviewed, and then facing the user after the review. Such a [delay] can ensure that the content received by the user must be legal and safe.

However, the emergence of live broadcast is naturally contrary to [Delay].

As a result, major domestic platforms have begun to introduce second-level lens processing to meet the low-latency network security needs.

In order to facilitate the majority of corporate users, the Programmer Inn has selected the following ten powerful companies from hundreds of content review companies through systematic screening and sorting, and recommended them to everyone (in no particular order).

1. Baidu Brain  https://ai.baidu.com/solution/censoring

 

Baidu Brain Security Platform implements custom image and text audit strategy configuration through interface operation, and at the same time updates the audit strategy preset in the iterative system according to the audit requirements of the supervisory unit, for video, live broadcast, social networking, e-commerce, O2O, games And other industries provide more comprehensive, real-time, and refined audit capabilities.

Its characteristic functions mainly have the following four points: First, update the regulatory requirements at the daily level. 2. A library of millions of public figures. 3. More comprehensive and fine review capabilities. Fourth, more flexible and free audit configuration. Ability to adjust the tightness of the audit model by itself to achieve audit capabilities that meet business needs.

 

2. Tencent Security  https://s.tencent.com/

Tencent's content security is based on Tencent's massive data support and Tencent's deep learning recognition technology, which can efficiently and accurately identify the illegal content of multimedia such as text, pictures and videos, helping customers reduce pornography, violence, politics, advertising and other customer-defined violations Risk, solve the problem of identifying violations of website and APP developers. While saving the cost of manual review, it greatly improves the identification efficiency and protects the healthy development of the business.

Tencent has a top-level machine learning research and development team, which has accumulated complete solutions and leading technology in the fields of face, image, music, voice, and machine learning. The FDDB evaluation set a new world record, and the detection accuracy rate exceeded 99%.

 

3. Alibaba Cloud  https://www.aliyun.com/product/lvwang

Based on deep learning technology, Alibaba Cloud Content Security provides intelligent identification services for multimedia content risks such as pictures, videos, voices, texts, etc., which can not only help users reduce the risks of pornography, violence, terrorism, and political-related violations, but also greatly reduce the cost of manual review .

It has the following four characteristics: 1. High cost performance, automatic and accurate identification, saving 90% of review manpower, and returning millisecond results; 2. Rich functions, providing content detection in the form of videos and pictures, covering violence, terrorism, politics, pornography and other violations Risks; 3. Practical testing, serving leading customers in industries such as live broadcast and video, with rich sample accumulation and control experience; 4. Personalized customization, one-to-one algorithm operation, support for personalized effect adjustment, and rapid iteration through data return.

 

4. Huawei Cloud  https://www.huaweicloud.com/product/moderation.html

Huawei Cloud Content Moderation, similarly, is based on image, text, and video detection technology, which can automatically detect pornography, advertising, political violence, and politically sensitive people, helping customers reduce the risk of business violations.

Product advantages: 1. Accurate detection, based on deep learning technology and a large number of sample libraries, to help customers quickly and accurately detect illegal content and maintain content security; 2. Rich functions, provide graphic and video content detection, covering pornographic, advertising, and related content Content detection for various violation risks such as violence and politics, as well as the detection of image clarity and composition quality; third, stable and reliable, content review services have been successfully applied to various scenarios, based on the long-term practice of Huawei and other corporate customers, Pass the test of complex scenarios; Fourth, simple and efficient, provide RESTful standardized API interface, and service SDK to facilitate customer use and integration; help customers reduce labor costs and save business expenses.

 

5. NetEase Easy Shield  https://dun.163.com/

NetEase Easy Shield is a security product under NetEase, a content security & business security service provider, responsible for NetEase mailbox anti-spam and UGC product anti-spam. It has 18 years of anti-spam expertise accumulated in text filtering, image recognition, voice analysis, and video Rich practical experience in testing and other aspects. Real-time filtering of harmful information such as advertisements, obscenity, terrorism, and contraband on the server and mobile terminals to reduce operational risks and improve product user experience.

Based on NetEase’s rich security experience, using artificial intelligence and big data technology, Yidun’s intelligent business risk control engine integrates security modules such as detection of illegal and bad information, verification code & anti-cheating, application reinforcement, game anti-plugin, and public opinion monitoring. Drive business security and efficiently reduce content and business risks for more and more enterprise systems.

 

6. Tuputech  https://www.tuputech.com/home

Different from the natural comprehensive advantages of BAT and other large companies, Tupu Technology is a national high-tech enterprise focusing only on artificial intelligence image recognition technology. It relies on advanced artificial intelligence and computer vision technology to integrate media content auditing and new retail on the Internet. Business intelligence management and pan-security are widely used in three major areas.

In April 2018, Tupu Technology won the world's first result in the MegaFace Global Million Face Challenge Test with its excellent technical advantages. At present, the average daily processing volume of image recognition by Tupu Technology is 2 billion times, and it has processed more than hundreds of billions of images. It is a well-known third-party independent recognition cloud platform in China.

With stable and reliable services and products that exceed user expectations, it has won numerous awards including Xunlei, Xiaomi, Meitu, Miaopai, Inke, Kugou, Aihuishou, OPPO, MINISO, Midea, Jinyi Jewelry, Li Ning, etc. The stable cooperation of hundreds of well-known enterprises and government-enterprise institutions is an early enterprise that successfully commercialized artificial intelligence and is unique in the field of artificial intelligence.

However, some well-known cloud service providers in China, which were originally to be included in this article for recommendation, also use the content review service of Tupu Technology.

 

7. Rongyun  https://www.rongcloud.cn/

Rongyun can be said to be the first professional instant messaging cloud service provider in China, focusing on providing basic instant messaging capabilities and cloud services for Internet and mobile Internet developers, and providing dynamic, easily scalable and often virtualized resources through the Internet.

In terms of content security, Rongyun also has a one-stop content review cloud service product and more than ten years of anti-spam and content security experience. The main functions are text filtering, image filtering and audio and video filtering.

However, its advantage lies in the direction of IM instant messaging, focusing on Internet communication cloud services.

 

8. Know that Chuangyu Cloud Security https://www.yunaq.com/

Know that Chuangyu is a professional free cloud website protection platform in China, providing free hacker attack cloud protection and cloud acceleration services for user websites.

It has provided very professional security consulting services and excellent solutions for many companies, and has made great contributions to information security.

 

9. Xiaodun Security https://sec.xiaodun.com/

Xiaodun Security is a subsidiary of Hangzhou Xiaodun Security Technology Co., Ltd. It is a new business security independent brand under Tongdun Technology. It is a leading full-stack business security solution expert in China, with the mission of "maintaining the security and clarity of the digital society". Driven by business attack intelligence, business data assets as raw materials, network security technology as the foundation, artificial intelligence algorithm as the core, relying on self-developed real-time computing, decision engine, knowledge graph, terminal situational awareness and other technologies, combined with advanced The "semi-supervised and unsupervised" machine learning algorithm provides customers with a full-stack business security solution based on a deep understanding of the customer's business model, builds an enterprise digital infrastructure security facility, and escorts the rapid growth of customer business.

 

10. Shumei Technology  https://www.ishumei.com/

Shumei Technology was established on June 1, 2015, and is committed to using artificial intelligence technology and massive data to solve the widespread fraud problems in finance and the Internet.


Its business scope is also very wide, and hundreds of companies cooperate with it. Based on big data and artificial intelligence, it solves multi-scenario fraud problems and provides professional and reliable anti-fraud solutions for Internet companies.

Guess you like

Origin blog.csdn.net/BinSTD/article/details/109466839