{version release announcement}HMS Core 6.6.0 is coming

Analysis service

◆ Retention analysis support ¬ Save lost users as audiences, developers will be more targeted when formulating relevant user recall strategies through stratification and multi-dimensional analysis of lost populations;

◆ The original "Audience Analysis" was renamed "Crowd Insights", with "User Grouping" and "User Portrait" modules under it. Audience segmentation in different dimensions can be achieved through user grouping, and user portraits can further analyze the characteristics of the corresponding segmented groups such as crowd portraits and attributes;

◆ The original "Uninstallation Analysis" has subdivision modules "Uninstallation Overview" and "Uninstallation Insights". Among them, Uninstall Insights adds a new uninstall flow report, which can help developers deeply locate the root cause of uninstall through the flow analysis before and after the user's application is uninstalled;

◆ Added access period distribution to page analysis. Developers can fully understand the user's product usage preferences and grasp the appropriate operation timing by comparing and analyzing the number of visits and the number of visiting users in different time periods;

◆ Supports the return of "Next Day Retention, 3 Day Retention, and 7 Day Retention" to HUAWEI Ads, which further enriches the evaluation indicators of advertising delivery effect.

View details>>

3D modeling service

◆ Added automatic bone binding capability. After the user uses an ordinary mobile phone to model a bipedal humanoid object, it automatically adapts according to the skeleton points, realizes the automatic skeleton binding of the bipedal humanoid object, and completes the custom action through the dynamic driving of the model, which reduces the threshold for ordinary users to create 3D animation and enhances the fun of the model. sex;

◆ Added AR capture guide mode. It supports accurate positioning of objects, real-time collection guidance, and key frame detection, helping users to complete modeling according to the steps and realizing a new interactive modeling experience.

View details>>

Video Editing Services

◆ The Atomic Ability SDK adds a one-click smile ability to the AI algorithm. Modify the expression by detecting the portrait in the photo, from not smiling to a pursed smile, or a toothy smile, helping to improve the happiness of the portrait expression;

◆ The Atomic Capability SDK adds target segmentation capability to the AI algorithm. The target object in the video is segmented by the AI algorithm to achieve the effect of quickly removing the video background. It is often used for background removal or replacement in various usage scenarios.

View details>>

Machine Learning Services

◆ Added motion detection service. Supports real-time capture of faces, and can determine whether it is a real living body according to the user's cooperative actions, avoiding non-living attacks such as remake pictures, remake videos, masks, etc., improve the accuracy of face recognition, and can be applied to payment-level face recognition scenarios;

◆ The offline text translation service has added 12 languages including Croatian, Macedonian, and Urdu (Malta, Bosnian, Icelandic, and Georgian do not support offline language detection).

View details>>

Audio Editing Services

◆ Added singing voice synthesis capability on the cloud side. It can intelligently generate high-fidelity singing voices through lyrics and tunes, combined with different music styles, support word-level input lyrics for phoneme conversion, generate singing voices corresponding to lyrics, and can adjust detailed parameters such as pitch, glide, breath, vibrato, etc. , make the singing more natural, this ability can be widely used in audio and video creative production, music education, virtual idols and other fields;

◆ Added AI dubbing interface on the cloud side. Support Restful interface, which can cover various terminal devices and meet the multi-device integration requirements of different developers;

◆ A new audio source separation asynchronous interface has been added on the cloud side. Add a query interface, maintain the full life cycle of audio separation tasks through taskId, and solve the problem that in some scenarios of audio separation, when the task takes too long, the user cannot find the previous task when he exits and re-enters the application;

◆ Side-to-side sound source separation Added accompaniment, bass, orchestra, drum, accompaniment with accompaniment, lead vocals and strings to meet more music separation usage scenarios.

View details>>

Sports Health Services

◆ Added data types for diving apnea training and diving apnea test exercise records, and new data types for free diving exercise records on the cloud side, supporting data recording in various sports scenarios;

◆ Added VO2 max exercise sampling data type on the terminal side. Each piece of data represents the maximum oxygen uptake during a certain exercise period, which can be used to monitor the indicators of aerobic exercise capacity in endurance sports;

◆ The new location atomic sampling statistics data type on the cloud side is open. Support to record the GPS location information of the user at every moment, which can be used for data recording of outdoor sports scenes such as mountain climbing and running;

◆ The newly added sports record segment statistics data type on the cloud side is open. Support statistical exercise records by time period, to meet the data record analysis of daily exercise training;

◆ Added scene event subscription capability on the cloud side. Support subscribing to the total number of steps target event, help users to formulate running and walking goals and provide push reminder function.

View details>>

video service

◆ Added HDR Vivid SDK, which provides OETF, Tonemapping, HDR2SDR and other capabilities of HDR Vivid video images. Realize the picture effect of brightness without exposure and details in the dark, creating an ultra-high-definition video playback experience;

◆ Added process exit capability. Release resources after the video is played to solve the problem of long-time occupation of process resources;

◆ Added the ability to obtain the thumbnail list of playback sources. Help developers to obtain each frame picture of the video playback source. When the user slowly slides the progress bar, the current frame thumbnail picture can be displayed according to the time point to optimize the playback experience;

◆ Added precise sliding playback capability. It can accurately locate the time point of the user's sliding playback, and solve the problem of inaccurate sliding playback caused by keyframe positioning.

View details>>

graphics engine service

◆ Added atomization interface 3D fluid component. Developers can use this component to customize the fluid boundary range, set the fluid volume, and finally achieve a 3D fluid sloshing effect that can interact with users;

◆ Added dynamic diffuse global illumination plug-in. Diffuse global illumination can be generated in real time in dynamic scenes and lighting to achieve more natural lighting rendering effects.

View details>>