tencent cloud

Challenges in Content Moderation

With the rapid development of the internet, smart devices, and various emerging businesses, internet data has been growing explosively and is subject to diverse uncontrollable risks, such as pornographic content and spam ads.

Content involving blood, porn, gambling, and drug use has become an area of focus, which makes content recognition a capability urgently needed by short video, news, and live streaming platforms.

As common non-compliant content in videos is difficult and costly to recognize through traditional means, enterprises face new technical challenges in content operations.

VM Overview

Tencent Cloud Video Moderation System (VM) is a security service with cutting-edge image recognition algorithms, which are trained with massive amounts of data in non-compliant videos to create a model for recognizing videos. It delivers a high recognition accuracy and recall rate, meets the needs for content recognition in multiple dimensions, and has been constantly improved in its recognition standards and capabilities in response to ever-changing risks.

• It can detect various scenes in videos/video streams and accurately recognize content that may be offensive, unsafe, or inappropriate, which effectively reduces the risks of non-compliant content and recognition costs.
• It can accurately recognize pornographic and other non-compliant content to filter videos of custom violation types. The recognition results include three categories: normal, suspected, and non-compliant. We recommend you allow normal videos, manually review suspected videos, and block non-compliant videos, so as to reduce the labor costs and increase the recognition efficiency.
• It provides services and capabilities in the form of open APIs. You can call them to get recognition results and efficiently create a smart business system to increase the efficiency of business operations.