Cloud Infinite (CI) provides intelligent analysis and processing capabilities for various data types such as images, audio/video, and documents from COS sources or third-party sources. It is widely applied across multiple business scenarios. Here are some examples:
E-commerce Scenarios
The e-commerce industry faces massive daily demands from different terminals, users with varying demands (merchants & buyers), and various data types (images, videos, text, audio). By utilizing Cloud Infinite's capabilities, users can reduce manual processing workflows, enhance performance and security, and effectively reduce costs.
Image Basic Processing
One original image, multiple display versions, satisfying customers' diverse image needs across various business scenarios.
Image Compression
Intelligent image compression meets the image distribution needs of cost-conscious customers.
Image Moderation/TAS
It helps users effectively identify prohibited content in avatars, product images, and comments. Detection scenarios include pornography detection, illegal and non-compliant content detection (such as gambling and drug abuse), and advertising content detection (such as QR codes and logos), comprehensively safeguarding platform content security.
Product Cutout
Leveraging Tencent Cloud's powerful AI capabilities, it supports identifying the product subject in images and removing background information to generate transparent-background product image assets. This can be applied to features such as intelligent e-commerce poster generation and dark mode adaptation.
Search by Image
Feature extraction, comparison, and search can be performed on images within the library.
Virtual Social Scenarios
Cloud Infinite provides a rich set of audio+video AI capabilities. Using audio and video as a medium, it helps users create various virtual personas for virtual social interactions, enabling users to overcome geographical and time limitations and communicate in a simpler way.
Voice/Text Auditing
It helps users effectively identify prohibited content in audio and text within voice chat rooms. Detection scenarios include pornography detection, illegal and non-compliant content detection (such as gambling and drug abuse), and advertising content detection (such as QR codes and logos), comprehensively safeguarding platform content security.
TTS
Through advanced deep learning technologies, text is converted into natural and smooth speech. Multiple voice options are currently available, with features for adjusting speed, pitch, and volume. This solution is applicable to scenarios such as intelligent customer service, voice interaction, literary audiobooks, and accessibility broadcasting, supporting output in PCM, WAV, and MP3 encoded formats.
Speech Recognition
Cloud Infinite supports recognizing audio files and asynchronously returns the transcribed text. Currently supported languages include Mandarin Chinese, English, and Cantonese. Additionally, it offers processing capabilities for recognition results, such as profanity filtering, filler word removal, and smart conversion of Arabic numerals, meeting diverse ASR requirements.
Video Matting
Automatically identifies human figures in videos and, after comprehensive analysis of the video content, outputs a video highlighting the human figure regions.
Live Streaming Scenario
Industries across the board are investing in video live streaming, TRTC, and related businesses. When deploying live streaming services, enterprises and platforms can use Cloud Infinite's integrated media + moderation + AI live streaming solution, which supports the following features:
Audio and Video Transcoding
Provides transcoding capabilities for media files such as audio and video, which is the process of converting a file stream into another stream. Transcoding can modify parameters of the original stream, including encoding format, resolution, and bitrate, enabling playback adaptation to different devices and network environments. It also reduces file size, thereby minimizing buffering while saving storage space and traffic costs.
Video Watermark
Cloud Infinite supports synchronously adding multiple watermarks to media assets, which enhances brand influence while reducing the risk of media file theft. It also supports various watermark formats, including static and dynamic images as well as text, to meet your watermarking needs across different scenarios.
MAS
It helps users effectively identify prohibited content in video frames. Detection scenarios include pornography detection, illegal and non-compliant content detection (such as gambling and drug abuse), and advertising content detection (such as QR codes and logos), comprehensively safeguarding platform content security.
Video Matting
After analyzing the video content, it outputs a video highlighting human figure regions, ensuring that character visuals remain unobstructed by bullet comments, thereby enhancing the viewing experience.
Cloud Infinite's real-time audio and video stream moderation solution offers the following advantages:
Relaying streams from TRTC to CSS incurs no additional fees.
Compared with pulling streams directly from TRTC, pulling streams through a CSS playback domain incurs lower outbound traffic fees.
Audio and video streams can be saved to COS with custom configurations, making it easy to meet regulatory backup requirements.
Supports automated moderation configuration without manual initiation.
Document Online Management Scenarios
Businesses such as cloud drives, online classrooms, and website transcoding all need to meet the online document browsing needs of users across multiple ends such as PC and App. This can be quickly achieved by using the document processing capabilities provided by Cloud Infinite.
Document Preview
Supports previewing over 30 file types in image or html format, preserving the original document styles to the greatest extent during online browsing, resolving the issue of certain document types being unsupported on different devices, and meeting the online document browsing needs across various terminals.