Released the function of saving churned users as an audience to the retention analysis function. This function enables multi-dimensional examination on churned users, thus contributing to making targeted measures for winning back such users.
Changed Audience analysis to Audience insight that has two submenus: User grouping and User profiling. User grouping allows for segmenting users into different audiences according to different dimensions, and user profiling provides audience features like profiles and attributes to facilitate in-depth user analysis.
Added the Page access in each time segment report to Page analysis. The report compares the numbers of access times and users in different time segments. Such vital information gives you access to your users' product usage preferences and thus helps you seize operations opportunities.
Added the Page access in each time segment report to Page analysis. The report compares the numbers of access times and users in different time segments. Such vital information gives you access to your users' product usage preferences and thus helps you seize operations opportunities.
Debuted the auto rigging capability. Auto rigging can load a preset motion to a 3D model of a biped humanoid, by using the skeleton points on the model. In this way, the capability automatically rigs and animates such a biped humanoid model, lowering the threshold of 3D animation creation and making 3D models appear more interesting.
Added the AR-based real-time guide mode. This mode accurately locates an object, provides real-time image collection guide, and detects key frames. Offering a series of steps for modeling, the mode delivers a fresh, interactive modeling experience.
Offered the auto-smile capability in the fundamental capability SDK. This capability detects faces in the input image and then lightens up the faces with a smile (closed-mouth or open-mouth).
Supplemented the fundamental capability SDK with the object segmentation capability. This AI algorithm-dependent capability separates the selected object from a video, to facilitate operations like background removal and replacement.
Released the interactive biometric verification service. It captures faces in real time and determines whether a face is of a real person or a face attack (like a face recapture image, face recapture video, or a face mask), by checking whether the specified actions are detected on the face. This service delivers a high-level of security, making it ideal in face recognition-based payment scenarios.
Improved the on-device translation service by supporting 12 more languages, including Croatian, Macedonian, and Urdu. Note that the following languages are not yet supported by on-device language detection: Maltese, Bosnian, Icelandic, and Georgian.
Learn more>>
Added the on-cloud REST APIs for the AI dubbing capability, which makes the capability accessible on more types of devices.
Added the asynchronous API for the audio source separation capability. On top of this, a query API was added to maintain an audio source separation task via taskId. This serves as the solution to the issue that in some scenarios, a user failed to find their previous audio source separation task when they exited and re-opened the app, because of the long time taken by an audio source separation task to complete.
Enriched on-device audio source separation with the following newly supported sound types: accompaniment, bass sound, stringed instrument sound, brass stringed instrument sound, drum sound, accompaniment with the backing vocal voice, and lead vocalist voice.
Learn more>>
Added two activity record data types: apnea training and apnea testing in diving, and supported the free diving record data type on the cloud-side service, giving access to the records of more activity types.
Added the sampling data type of the maximum oxygen uptake to the device-side service. Each data record indicates the maximum oxygen uptake in a period. This sampling data type can be used as an indicator of aerobic capacity.
Added the open atomic sampling statistical data type of location to the cloud-side service. This type of data records the GPS location of a user at a certain time point, which is ideal for recording data of an outdoor sport like mountain hiking and running.
Opened the activity record segment statistical data type on the cloud-side service. Activity records now can be collected by time segment, to better satisfy requirements on analysis of activity records.
Added the subscription of scenario-based events and supported the subscription of total step goal events. These fresh features help users set their running/walking goals and receive push messages notifying them of their goals.
Learn more>>
Released the HDR Vivid SDK that provides video processing features like opto-electronic transfer function (OETF), tone mapping, and HDR2SDR. This SDK helps you immerse your users with high-definition videos that get rid of overexposure and have clear details even in dark parts of video frames.
Added the capability for killing the WisePlayer process. This capability frees resources occupied by WisePlayer after the video playback ends, to prevent WisePlayer from occupying resources for too long.
Added the capability to obtain the list of video source thumbnails that cover each frame of the video source — frame by frame, time point by time point — when a user slowly drags the video progress bar, to improve video watching experience.
Added the capability to accurately play video via dragging on the progress bar. This capability can locate the time point on the progress bar, to avoid the inaccurate location issue caused by using the key frame for playback location.
Learn more>>
Added the 3D fluid simulation component. This component allows you to set the boundaries and volume of fluid (VOF), to create interactive liquid sloshing.
Introduced the dynamic diffuse global illumination (DDGI) plugin. This plugin can create diffuse global illumination in real time when the object position or light source in the scene changes. In this way, the plugin delivers a more natural-looking rendering effect.
Learn more>>
For the hms-mapkit-demo sample code: Added the MapsInitializer.initialize API that is used to initialize the Map SDK before it can be used.
Added the public layer (precipitation map) in the enhanced SDK.
- For the hms-sitekit-demo sample code: Updated the Gson version to 2.9.0 and optimized the internal memory. Go to GitHub >>
- For the hms-game-demo sample code: Added the configuration of removing the dependency installation boost of HMS Core (APK), and supported HUAWEI Vision.
Go to GitHub >>
Made necessary updates to other kits. Learn more >>
Top comments (0)