This topic describes the product and documentation updates for Alibaba Cloud Visual Intelligence API in 2021.
December 2021
Category | Capability Name | Description | Release date | Supported platforms | References |
Offline SDK | Limb keypoint SDK | Provides detection information for 15 key points in authorized human images, such as the nose, eyes, neck, left shoulder, and right shoulder. | 2021-12-30 | Android and iOS | |
Limb action counting SDK | Captures video of human actions using a camera. Detects limb keypoints in real time and automatically counts actions. Supports 15 fitness actions, such as jumping rope, squats, jumping jacks, sit-ups, push-ups, planks, and glute bridges. Custom fitness actions are also supported. | 2021-12-30 | Android and iOS | ||
Limb action counting feedback SDK | Uses AI to detect 15 types of incorrect limb actions in real time and provides immediate feedback. | 2021-12-30 | Android and iOS | ||
Facial landmark SDK | Detects the number of faces and face regions in an image. Returns the face count, coordinates of 106 basic facial landmarks, 134 additional fine-grained landmarks, and 40 pupil landmarks. | 2021-12-30 | Android and iOS | ||
Image enhancement SDK | Enlarges original images by 2× without quality loss. | 2021-12-30 | Android and iOS | ||
Filter SDK | Provides eight filters: Normal, Vibrant, Fresh, Food, Japanese, Beauty, Mint, and Black & White. Applies filters while preserving image quality. | 2021-12-30 | Android and iOS |
November 2021
Category | API Name | Description | Release date | Region | References |
Object detection | Clothing detection | Uses visual AI algorithms, Internet of Things (IoT), and big data analytics to detect whether personnel wear hats, masks, or uniforms in specified areas. Sends real-time alerts for noncompliant attire. | 2021-11-30 | China (Shanghai) | |
Cat and mouse detection | Uses visual AI algorithms, IoT, and big data analytics to detect cats, mice, and other animals in scenes. Sends real-time alerts. | 2021-11-30 | China (Shanghai) |
October 2021
Category | Interface Name | Description | Release date | Region | References |
Face and body | Batch add face data | Adds face data to a specified database in batches. | 2021-10-29 | China (Shanghai) |
September 2021
Category | Interface Name | Description | Release date | Region | References |
Face and body | Smart slimming | Takes a portrait image as input. Detects and analyzes facial features. Generates a slimmer version of the face. Supports up to three faces per image. | September 30, 2021 | China (Shanghai) | |
Smart skin retouching | Takes a portrait image as input. Smooths facial skin and removes blemishes such as acne, acne scars, and freckles. Whitens skin across the entire body. Preserves natural skin texture. Supports multiple people per image. | 2021-09-31 | China (Shanghai) |
August 2021
Category | Interface Name | Description | Release date | Region | References |
Image recognition | Ad creative analysis | Tags people—including celebrities, ordinary people, and computer-generated characters—and scenes in ad images. Supports thousands of content tags with broad coverage. | 2021-08-31 | China (Shanghai) and China (Hohhot) | |
Video understanding | Video content understanding | Added support for China (Hohhot). | 2021-08-31 | China (Shanghai) and China (Hohhot) | |
Object detection | IPC video object detection | Detects objects—such as people, vehicles, and pets—in input videos. | 2021-08-31 | China (Shanghai) |
July 2021
Category | Interface Name | Description | Release date | Region | References |
OCR | VAT invoice roll recognition | Recognizes structured fields on VAT invoice rolls, including total amount including tax, invoice code, invoice number, total tax amount, total amount, password area, issue date, tax rate, buyer identification number, and seller identification number. | 2021-07-31 | China (Shanghai) | |
Fixed-amount invoice recognition | Recognizes structured fields on fixed-amount invoices, including invoice number, invoice code, and invoice amount. | 2021-07-31 | China (Shanghai) | ||
PDF recognition | Performs structured text recognition on PDF files. | 2021-07-31 | China (Shanghai) | ||
Image analysis and processing | Aortic aneurysm and pulmonary hypertension detection | Segments the aorta and pulmonary artery from chest CT DICOM scan data. Extracts centerlines for both vessels. Generates optimal-view Stretch CPR, Cross Section, and Straightened CPR images that wrap around each vessel. Returns the maximum diameter of each vessel. Also returns cross-sectional area at 1 mm intervals along each centerline and the position of each point in the patient coordinate system of the original image. | 2021-07-31 | China (Shanghai) |
June 2021
Category | Capability | Description | Release date | Supported platforms | References |
Offline SDK | Real-time video segmentation SDK | Uses deep learning frameworks and detection-recognition techniques to deliver high-precision visual segmentation. Segments foreground subjects and scenes at the pixel level in real time. Works well with highly transparent subjects and complex backgrounds. | 2021-06-30 | Android and iOS | |
Offline image segmentation SDK | Uses detection and recognition techniques to perform precise, flawless background removal on user-captured or uploaded images. Delivers high-precision visual segmentation. Supports segmentation and background replacement for complex images. | 2021-06-30 | Android and iOS | ||
Certificate recognition SDK | Uses Alibaba Cloud Visual Intelligence API’s innovative certificate recognition technology for efficient certificate recognition. | 2021-06-30 | Android and iOS | ||
Vehicle recognition SDK | Scans and recognizes all mainland China single-line license plates and VIN codes from video streams. | 2021-06-30 | Android and iOS | ||
General OCR SDK | Performs offline OCR on Android or iOS devices. Has a small package size and delivers sub-second recognition speed. | 2021-06-30 | Android and iOS |
May 2021
Category | API Name | Description | Release date | Region | References |
3D vision | Monocular video depth estimation | Takes a color video as input. Estimates a depth map for each frame. Generates a point cloud. | 2021-05-31 | China (Shanghai) | Offline |
April 2021
Category | API Name | Description | Release date | Region | References |
Face and body | Online proctoring | Detects candidate behavior during online exams. Supports screen chat tool detection and candidate status detection. | 2021-04-30 | China (Shanghai) | |
3D vision | Image-based human reconstruction | Estimates the 3D depth value for each pixel in a single human image. Outputs a local unit mesh model from the corresponding viewpoint. | 2021-04-30 | China (Shanghai) | Offline |
Multi-view 3D reconstruction | Takes multiple color images of the same scene and their corresponding camera positions as input. Reconstructs a 3D model of the main subject or scene. Outputs a 3D point cloud. | 2021-04-30 | China (Shanghai) | Offline |
March 2021
Category | Interface Name | Description | Release date | Region | References |
3D vision | Monocular image depth estimation | Estimates the 3D depth value for each pixel in a single image. Outputs a depth map. | 2021-03-25 | China (Shanghai) | Offline |
Stereo depth estimation | Takes two stereo color images—left and right—as input. Estimates and outputs the disparity map for the left image. | 2021-03-25 | China (Shanghai) | Offline |
February 2021
Category | API Name | Description | Release date | Region | References |
Face and body | Static gesture recognition | Detects gestures in images. | 2021-02-26 | China (Shanghai) | Offline |
January 2021
Category | Interface Name | Description | Release date | Region | References |
Face and body | Face fusion | Fuses one person’s face into another person’s facial features, with proper authorization. | 2021-01-31 | China (Shanghai) | |
Add face fusion template | Uses approved face images as templates for face fusion. | 2021-01-31 | China (Shanghai) | ||
Query face fusion template | Lists existing face templates. | 2021-01-31 | China (Shanghai) | ||
Delete face fusion template | Deletes existing face templates. | 2021-01-31 | China (Shanghai) | ||
Human sketch stylization | Automatically crops the head region from a portrait image and generates a sketch effect. | 2021-01-31 | China (Shanghai) | ||
Server-side identity verification | You can request the user’s image information for identity verification on the server side by providing the verified person’s name and ID number. | 2021-01-31 | China (Shanghai) | ||
Mobile identity verification request | Requests identity verification information on mobile devices using the person’s name and ID number. | 2021-01-31 | China (Shanghai) | ||
Mobile identity verification query | Returns identity verification information after confirming the ID is valid and the name matches. | 2021-01-31 | China (Shanghai) | ||
Object detection | Vehicle illegal parking detection | Detects parked vehicles in target areas of images. | 2021-01-31 | China (Shanghai) | |
Vehicle congestion detection | Determines whether traffic congestion occurs based on vehicles in images. | 2021-01-31 | China (Shanghai) | ||
Image recognition | Food recognition | Identifies food categories and calorie counts in images. | 2021-01-31 | China (Shanghai) | |
Video segmentation | Green screen video segmentation | Removes green screens from videos. Automatically segments foreground subjects from green screen backgrounds. | 2021-01-31 | China (Shanghai) | Offline |
Video understanding | Video content understanding | Analyzes elements in videos, such as celebrities, ordinary people, and game footage. | 2021-01-31 | China (Shanghai) |