The new version of Intelligent Media Management (IMM) is now available and offers more powerful features than the old version. This topic compares the new and old versions and describes best practices for using the new version.
Benefits of the new version
The new version of Intelligent Media Management (IMM) integrates more practical features to help you better analyze and use your resources stored on services such as Alibaba Cloud Object Storage Service (OSS) and Drive and Photo Service (PDS).
The old version of Intelligent Media Management (IMM) will no longer support the US (Silicon Valley) region starting December 31, 2023.
New features
More powerful data processing
Data processing now supports features such as audio and video processing, document processing, AI-powered image analytics, AI-powered audio and video analytics, file processing, and Content Moderation.
More flexible data management and query
Sets have been upgraded to datasets. Datasets support metadata extraction and metadata queries for all types of files.
New data ingestion methods
IMM provides multiple methods, such as data binding, triggers, and batch processing, to ingest data from various data sources, such as OSS and PDS.
More versatile task management and workflow support
The task list supports filtering by multiple dimensions, such as time and type. Workflows support all data processing capabilities and can be orchestrated for any business scenario.
Experience improvements
Full-featured console
IMM provides a full-featured console that you can use to access all IMM features.
SDKs for multiple languages
IMM provides software development kits (SDKs) for eight mainstream programming languages. You can use the SDKs to access all IMM features.
Multiple billing methods
In addition to the pay-as-you-go billing method, you can purchase data processing resource plans to offset fees incurred by all IMM features.
Higher capacity
IMM provides higher initial concurrency and greater elasticity for all IMM features.
Relationship between the new and old versions
Independent use
The projects and datasets (called sets in the old version) of the new and old versions are completely independent and do not interfere with each other. For example:
In the console, you can switch between the new and old versions to manage the corresponding projects.
SDK/API versions.
The old version uses the old SDK. The API version is 2017-09-06.
The new version uses the new SDK. The API version is 2020-09-30.
Mapping between new and old API operations
Compared with the old API operations (version 2017-09-06), the new API operations (version 2020-09-30) include more practical features and deprecate some previous operations. The following table describes the mapping between the new and old API operations.
Feature | New API | Legacy API | Description |
Service region | None | The new version lets you query a list of regions where Intelligent Media Management is available. The list includes region IDs and localized descriptions. | |
None | The new version no longer provides an API operation to activate Intelligent Media Management. You can activate the service in the Intelligent Media Management console, which is more convenient. For more information, see Activate IMM. Important If you have activated Intelligent Media Management using the old API operation, you can use the new version directly. You do not need to activate the service again. | ||
Project management | When you create a project in the new version, you can set a workflow template and quotas.
| ||
UpdateProject - Updates the CU or service role information of a project | |||
Dataset management | When you create a dataset in the new version, you can set a workflow template and quotas.
| ||
Metadata indexing | The new version supports creating indexes for the metadata of files in a dataset as a batch operation. After the indexes are created, you can perform operations such as data retrieval and face clustering. | ||
IndexVideo - Performs automatic frame splitting, grouping, and recognition on a video | |||
BatchIndexFileMeta - Creates indexes for object metadata in a batch | None | ||
The new version supports obtaining information about all types of files in a dataset. | |||
None | The new version supports retrieving information about all types of files in a dataset as a batch operation. | ||
With the new version, you can delete all types of files from the centralized data storage. | |||
None | The new version lets you batch delete all types of files from the centralized data storage. | ||
UpdateImage - Updates user-defined information about an image in a set | The new version supports updating the metadata of one or more files of all types in a dataset. | ||
None | |||
Query and statistics | The new version supports querying and collecting statistics on all types of files in a dataset. The following query methods are supported:
| ||
Face clustering | CreateFigureClusteringTask - Creates a figure clustering task | - | |
The new version supports merging more than two clusters. | |||
None | The new version supports obtaining a single cluster and its information based on the object ID of the cluster. | ||
The new version supports querying clusters in a dataset based on their creation time or update time. | |||
None | The new version supports retrieving clusters and their information as a batch operation based on the object IDs of the clusters. | ||
- | |||
SearchImageFigureCluster - Queries the cluster to which a face in an image belongs | None | The new version supports searching for the cluster to which a figure in an image belongs. | |
CreateFacesSearchingTask - Creates a task to search for similar-face images | FindSimilarFaces - Finds faces that are similar to a specified face | - | |
Spatio-temporal clustering | None | The new version supports creating spatio-temporal clusters based on metadata such as the shooting time and location of photos in a dataset. You can also query, update, and delete the clusters. | |
Story | None | The new version supports generating a photo story collection from photos in a dataset based on dimensions such as time, location, figures, and tags. | |
Image clustering | None | The new version supports categorizing and querying images in a dataset based on similarity. This feature can be used in scenarios such as removing duplicate similar images, for example, burst shots, and selecting the best image. | |
Data binding | None | The new version supports creating a binding between a dataset and an OSS bucket. After the binding is created, IMM performs a full scan of the existing data in the OSS bucket to extract and index the object metadata. After the full scan is complete, IMM performs real-time incremental scans to extract and index the metadata of new files in the OSS bucket. | |
Image processing | - | ||
GetImageQuality - Obtains the overall quality score of an image | - | ||
The new version supports barcode detection. | |||
DetectImageFaces - Detects faces and face information in an image using an AI model | DetectImageFaces - Detects faces and face information in an image | - | |
GetImageCroppingSuggestions - Obtains image cropping suggestions (cropping box position) | - | ||
None | The new version supports adding mosaics, Gaussian blurs, or solid color blocks to a specific area of an image for privacy protection. | ||
CreateImageToPDFTask - Creates a task to convert images to a PDF file | None | The new version supports converting multiple images into a single PDF file. | |
None | The new version supports splicing multiple images into a single image. | ||
CompareImageFaces - Compares the similarity of faces in images | - | ||
- | |||
None | The new version supports detecting vehicle information in an image. | ||
ApsaraVideo Media Processing | CreateVideoCompressTask - Creates a task to compress and store a video | The new version supports operations such as audio extraction and video watermarking for audio and video files. | |
- | |||
CreateVideoLabelClassificationTask - Creates a video label detection task | None | The new version supports detecting the labels that correspond to the content in a video. | |
GetVideoLabelClassificationResult - Obtains a video label detection task | |||
Document processing | The new version provides an improved document preview experience. | ||
RefreshOfficePreviewToken - Refreshes a document preview credential | |||
RefreshWebofficeToken - Refreshes a WebOffice access credential | |||
CreateOfficeConversionTask - Creates a document conversion task | ConvertOfficeFormat - Converts the format of a document (such as an Office document) | - | |
CreateOfficeConversionTask - Creates a conversion task for a single document | |||
None | The new version supports extracting text from a document. | ||
File processing | None | The new version supports compressing and decompressing files, and compressing point cloud files. | |
Content Moderation | None | The new version supports compliance detection for text, images, and videos to reduce the risk of violations. | |
Task management | The new version supports obtaining information about all asynchronous tasks by task ID, including the task start time, running status, and end time. | ||
ListOfficeConversionTask - Lists the document format conversion tasks that are created in a project | The new version supports obtaining information about all asynchronous tasks based on the task running status and the time range when the tasks start or end. | ||
None | The new version supports creating triggers. After a trigger is created, you can use event sources such as OSS to trigger IMM actions. You can select a data processing template to process media files, such as images, videos, and documents. | ||
None | The new version supports creating batch processing jobs. After a job is created, you can specify existing files in storage and perform specified operations on them as a batch operation, such as transcoding and format conversion. | ||
Asynchronous document conversion requests | None | DeleteOfficeConversionTask - Deletes an asynchronous document format conversion task | In the new version, task information is saved for only seven days after a task starts. After seven days, the information can no longer be retrieved. You can retrieve task information in a timely manner using multiple methods. For more information, see the description of each task creation operation. |
Managing media set tags | None | You can use the new SimpleQuery operation to collect statistics on tags in a dataset. | |
Blind image watermark | The new version offers significant improvements in speed and attack resistance compared to the old version. It is also forward-compatible and can decrypt images that were encrypted using the old version. | ||
CreateDecodeBlindWatermarkTask - Creates a task to parse a blind watermark |
Billing differences
The new version uses the pay-as-you-go billing method. You can also purchase resource plans to offset fees. For more information, see Billing overview.
Migrate to the new version
Why migrate?
New features and optimizations are available only in the new version of IMM. The old version is being phased out. To ensure the best user experience, we recommend that you migrate to the new version as soon as possible.
How to migrate
Without the media set feature
Follow these steps to migrate:
Create a project in the new version. You can use the new console or the new API.
Test the corresponding operations for the same features in the new version.
Switch your business to the new API operations.
Using the media set feature
For new data, see Create a dataset.
For historical data, re-index the files from the set to a dataset, and then use the dataset.
If you need help, contact us.
FAQ
Can I still use projects created in the old version?
Yes, you can. Make sure to switch to the correct console version. When you call API operations, ensure that you use the correct SDK and API versions.
We strongly recommend that you migrate to the new version as soon as possible.
Why can't I find my project in the console?
Check whether you are using the correct console version (new or old). Also, check the region that is selected at the top of the console to ensure it is the region where your project is located.
Can I use the new and old versions together?
Yes, you can. The projects for the two versions are completely independent and are billed separately. We strongly recommend that you migrate to the new version as soon as possible.