As numerous excellent Mini Programs emerge in the content operations field, content security risks gradually become prominent. In recent years, apps in popular fields such as new media, live streaming, and gaming are frequently criticized and punished publicly due to content security issues, and even forcibly removed from the platform.
For developers and operators, content review is not an easy job. It requires not only the mastery of information security technologies, but also the investment of human review resources. The platform provides the content review capabilities by packaging existing information security capabilities into APIs and making them available to the public via the Mini Programs Open Platform. This helps Mini Program content developers detect illegal porn-related and political-related information in images and texts and save them the energy and cost they invest in content security.
# Capabilities Provided by the Content Security API
# Intelligent porn detection
Feature description: When users send images to interact with the platform, the images may contain porn features, regardless of whether they are the content uploaded after being photographed, custom profile photos, or the images used to comment on articles. The intelligent porn detection API provides the automatic image recognition capability, which determines whether the image is involved in pornography violations based on the degree of porn quantization and the detected porn features.
Advantages: Intelligent porn detection comes with various advanced recognition technologies, including porn feature hinting, strong recognition for local areas, image and text recognition, and other recognition modes.
The system determines whether images contain pornographic features from multiple dimensions and outputs the results. This feature allows developers to determine the pornographic elements in the images and prevent propagation of pornographic images in Mini Programs.
Use cases: Detects photos taken by camera-related apps (such as photo-editing apps and selfie apps) and uploaded by users, images of e-commerce commodities, and images in media articles.
# Recognition of sensitive faces
Feature description: Some users post images of sensitive people via the upload feature to spread appalling comments. This will lead to the spread of adverse opinions and affect the normal operation of Mini Programs. Therefore, the content security API provides the capability of recognizing sensitive faces. This API can extract the face information from the images, find and determine similar sensitive faces, and give developers a feedback about whether the images contain sensitive people, making developers more capable of discovering illegal content.
Advantages: In addition to standard front face recognition, this feature also supports side-face detection, interference filtering, and multi-face recognition.
Use cases: Detects profile photos, images in media articles, and photos uploaded by social network users.
# Junk text detection
Feature description: Users interact with the platform or with each other by using text more often than images. Text may contain a wider range and a larger number of potential illegal elements than images, and there are also various text variations. In this case, simple keyword filtering cannot meet the requirements of intercepting all junk text. The content security API provides full capabilities of junk text detection to recognize porn, illegal, political-related, and abusing text information that violates the regulations and block it in a timely manner.
Advantages: Interfering text and countless illegal text variations can also be recognized accurately. This feature provides anti-interference capabilities, including but not limited to writing variation, spelling mixing, symbol interference, and sticker interference.
The junk text detection feature enables developers to focus more on Mini Program content development, without spending a large amount of time and efforts in content review. This can improve the development efficiency and make it easy for developers to create Mini Programs of higher quality.
Use cases: Detects infringing words in the personal information of users, articles and comments posted by media news users, and assets (for example, questions and answers uploaded by quiz Mini Game users) edited and uploaded by game users.
# Access Guide
Text Content Detection API Documentation: <a href="https://developers.weixin.qq.com/miniprogram/dev/api/msgSecCheck.html"target="_blank">msgSecCheck
Image Content Detection API Documentation: <a href="https://developers.weixin.qq.com/miniprogram/dev/api/imgSecCheck.html"target="_blank">imgSecCheck