Intelligent Media Management:Functions and features

Feature set

Feature

Description

References

Basic operations

Project operations

To use IMM to process data, you must first create a project. You can create different projects to group processing capabilities in your application based on your business requirements.

-

Dataset operations

A dataset is a container for metadata. In most cases, the metadata of correlated media files is stored in the same dataset to facilitate data queries. After you create a dataset, you can create a metadata index for files that are stored in services such as Object Storage Service (OSS) and Photo and Drive Service. When you create a metadata index, IMM collects metadata and stores the obtained metadata in the metadata storage engine. This enables robust data querying, statistical analysis, and effective management capabilities.

-

Task management

Trigger

A trigger initiates data processing tasks on incremental objects in an OSS bucket that meet the specified conditions. For example, you can create a trigger that converts incremental DOCX objects in a specific directory of a bucket into PDF objects or transcodes incremental MOV objects into MP4 objects. Unlike a batch processor, a trigger initiates tasks only on incremental objects in a bucket, not on existing objects in the bucket.

-

Batch processor

A batch processor initiates tasks on existing objects in an OSS bucket that meet the specified conditions. For example, you can use a batch processor to convert existing DOCX objects in a specific directory of a bucket into PDF objects, or transcode all MOV objects into MP4 objects.

-

Task query

You can query the progress information of asynchronous tasks and list all matching tasks.

-

Feature set

Feature

Description

References

Image detection and recognition

Image label detection

IMM can detect image content such as scenes, objects, and events, and can automatically label images. Image label detection supports thousands of labels in more than 30 categories.

-

Face detection

Enabled by computer vision, facial recognition can detect and locate individual faces in images and videos. Face detection can be used in a variety of scenarios, such as authentication, public surveillance, intelligent album management, and customer behavior analysis.

-

Human body detection

IMM can detect the presence and positions of human bodies in images, then return confidence scores. This feature can be used in various fields, such as public safety and crowd monitoring.

-

Vehicle detection

Vehicle detection can be used to detect vehicles in images. This technology is widely used in traffic monitoring, intelligent parking systems, autonomous driving, urban traffic management, electronic toll collection systems, and safety and rescue services.

-

Face similarity comparison

Face similarity comparison compares a collected face image with the faces that are recorded in your system and provides a similarity score, which is then used to determine whether the face matches a specific person. This feature can be used in face-based applications, such as authentication, identify verification, and face recognition.

-

QR code recognition

IMM can detect the positions and content of one or more QR codes or barcodes in image files such as photos and screenshots, and then return the position and text information that the codes convey.

-

Image editing and analysis

Image blurring

You can use mosaics, Gaussian blurs, or solid color shapes to blur a specific area of an image for privacy protection.

-

Image-to-PDF conversion

IMM allows you to convert multiple images into a single PDF file, thus facilitating image searches.

-

Image concatenation

Image concatenation allows you to create a wide-perspective image by concatenating two or more spatially overlapped images captured from different viewpoints and perspectives, and at different times.

-

Image cropping suggestions

The image cropping suggestions feature can return the suggested cropping frame for an image and the aesthetic score of the cropping solution based on a specified cropping ratio. If you specify multiple cropping ratios for an image, cropping solutions are provided for each cropping ratio.

-

Image quality assessment

The image quality assessment feature gives an overall visual quality score of an image based on metrics such as sharpness, noise, distortion, color saturation, and exposure. Image quality assessment can be used to select a high-quality cover image for articles and a thumbnail for videos, and filter out low-quality images.

-

Blind watermarking

IMM can add an image or text blind watermark to an image. A blind watermark is invisible. If you want to recover a blind watermark, use the blind watermark decoding feature. Blind watermarking is useful for image copyright protection.

-

Feature set

Feature

Description

References

Media recognition and detection

Video label detection

The video label detection feature allows you to perform intelligent analysis on a video and obtain labels of the video. IMM provides a comprehensive set of video labels. Video label detection provides high accuracy, effectiveness, and value. You can classify and retrieve videos based on these labels to enhance accuracy and efficiency in video management.

-

Reverse geocoding

With the reverse geocoding feature of IMM, you can detect geographical information in media data.

-

Media editing and processing

Media transcoding

Media transcoding is a media processing feature used for processing multimedia data. It provides a cost-effective, easy-to-use, elastic, and highly-scalable method to convert audio and video objects stored in OSS into formats suitable for playback on PCs, TVs, and mobile terminals.

-

Media metadata collection

IMM collects media metadata, such as the resolution, bitrate, frame rate, and encoding protocol. You can then use the media metadata to retrieve media files, display media information during playback, and intelligently manage the media files. This improves the efficiency of media file management.

-

Live transcoding

Unlike transcoding, which enables video playback only after the entire video is transcoded, live transcoding transcodes only the video segments that need to be played. This allows playback to start immediately after the original video file is uploaded.

-

Feature set

Feature

Description

References

Online document services

Online document editing

Edit a document in OSS or Photo and Drive Service online.

-

Preview files online

Preview a document in OSS or Photo and Drive Service online.

-

Document format conversion

Convert a document from one format to another, such as from DOC to PDF.

-

Document content processing

Content extraction

Automatically extract plain text from documents in various formats and languages.

-

Feature set

Feature

Description

References

File processing

File compression

The compression feature enables you to package objects in OSS into formats such as ZIP, facilitating efficient and streamlined data management in the cloud.

-

File decompression

IMM allows you to extract the specified files from a ZIP, RAR, or 7z package to the specified directory, or decompress the entire package.

-

Special data processing

Point cloud compression

A point cloud is a set of data points in a three-dimensional coordinate system. Point cloud data consumes significant storage space and bandwidth during data transfer. You can use point cloud compression to analyze and process spatial-temporal information stored in point clouds. This greatly reduces the size and storage costs of point cloud data, and helps you implement high-quality and real-time point cloud data encoding and decoding solutions.

-

Feature set

Feature

Description

References

Retrieval-oriented processing

Semantic retrieval

Semantic retrieval uses a vector retrieval method to retrieve information based on the content meaning or semantics, such as "a bird's-eye view of a forest," "a snow-covered city," or "grassland of last summer." You can use the semantic retrieval feature to retrieve data stored in OSS and Photo and Drive Service.

-

Face clustering

The face clustering feature allows you to group multiple images that contain similar faces in a dataset. The feature is suitable for scenarios such as face albums in cloud drives, stranger detection in home surveillance, and customer management in the New Retail industry. After you perform face clustering, you can query all images that contain the face of a specific person in a group.

-

Face search

After you index image metadata into a dataset, you can use face search to search for a specific number of images that are most similar to a given face image or face ID. Face search can be used in face-based business scenarios, such as customer identification.

-

Spatiotemporal clustering

Spatiotemporal clustering allows you to group photos based on temporal and spatial metadata.

-

Image clustering

Image clustering classifies similar images into a group. You can use image clustering to filter and classify images that are taken in continuous shooting mode.

-

Generative processing

Story generation

The story feature uses artificial intelligence (AI) algorithms to group images by time or person into an image album with a quality cover. You can use this feature to create image albums that tell your memorable life stories.

-