Security and Compliance
Huaxiyun and its Xinjiang media customers needed a faster, more reliable way to review huge volumes of broadcast and video content because manual review was expensive, slow, and prone to missing risky or noncompliant material.
Reported impact
Huaxiyun improved content-review speed and safety for Xinjiang media workflows, reduced manual moderation effort, and created a cloud-based moderation model that can scale across many broadcast channels.
Huaxiyun introduced Baidu AI Cloud's VCR media-content-review capability into the Shiliu Media Cloud used for Xinjiang radio and television operations. The system applies deep-learning-based content moderation to images and video so it can detect politically sensitive material, pornography, violence, terrorism, commercial spam, and other problem categories that human reviewers may miss at scale. It also evaluates visual quality attributes such as image clarity and aesthetics, allowing the media platform to classify content by service level instead of only blocking risk. Architecturally, the moderation engine is combined with Huaxiyun's broader cloud-media environment, which is built on distributed cloud-computing and big-data capabilities to support media services across Xinjiang.
Video and image content enters the Shiliu Media Cloud, where VCR analyzes frames for risky categories and image-quality attributes. Review results flow back into the content-review platform so broadcasters can intercept unsafe material quickly and move approved content through the publishing pipeline faster.
Huaxiyun introduced Baidu AI Cloud's media content review VCR into the Shiliu Media Cloud build to solve the content-review problem (original: 华栖云在石榴融媒云的设计建设中，首次引入了百度智能云的人工智能技术：媒体内容审核VCR).
The platform also evaluates image aesthetics and clarity and is combined with cloud-computing and big-data technology to deliver media cloud services across the Xinjiang region (original: VCR还能从美观和清晰等维度对图像进行筛选…该平台结合先进的云计算、大数据技术覆盖广，能够为全疆提供融媒云服务).
VCR can identify pornography, violence and terror, political sensitivity, advertising, disgusting content, and more, including details that human eyes may miss under fatigue (original: VCR基于深度学习的智能内容审核方案，能够准确识别图片和视频中的涉黄、涉暴涉恐、政治敏感、微商广告、恶心等内容，让很多人眼在疲劳下极易忽略的细节都难逃机器法眼).
Xinjiang broadcasting deployed a VCR intelligent review system with Baidu AI Cloud, using multidimensional AI recognition to intercept political and pornographic risks in seconds while improving review efficiency 90% and covering 100+ channels (original: 新疆广电联合百度智能云部署VCR智能审核系统，通过AI多维识别技术实现涉政涉黄等20类风险秒级拦截，审核效率提升90%且零漏检，覆盖全疆100+频道).
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Databricks Sample Jul 7, 2026 Bally's Bally's needed a better way to support omnichannel gaming personalization, compliant analytics, and faster market launches because legacy technology fragmented customer and operational data, which slowed… 10x serverless compute Databricks Data + AI Platform · Delta Lake · Lakeflow Jobs Databricks Sample Mar 14, 2025 Scatter Lab However, as a startup, building their own LLM presented a number of practical constraints. 1.5 Million Users Secured in just nine months after launching Agent Bricks 
Microsoft Sample Jan 30, 2025 SHUEISHA Numerous SaaS apps and digital tools are used in the creative pursuits of writers, creators, and editors, but issues surround the protection of intellectual property. Meanwhile, working styles have been changing rapidly… One might expect that creating a report for that number of individuals would take a tremendous amount of time,… Microsoft 365 · Microsoft 365 E5 · Microsoft Security Copilot 
