It can position the 21 core critical points of the human body, including five sense organs, neck, and four limbs. Also, it supports complicated scenarios involving multiple persons and a big amplitude of motion.
It can detect all human bodies in an image and recognize over 20 attributes of the human body, including gender, age, clothing and jewelry, and behaviors and actions.
It can calculate the number of humans in an image and predict its movement trend. It can calculate the target people group's statistical quantity without capturing the face and full-body portraits.
It can position 21 main skeleton nodes of the hand. Also, it can be used to customize such scenarios as hand gesture detection, AR special effect, and man-machine interaction.
It can recognize 24 common gestures, including OK, fist, hand heart, thumb-up, and number gestures. It supports selfies, shooting by others, and various shooting angles.
It can recognize the contour of the human body and separate the portrait from the background. It is applicable for such scenarios as shooting background replacement, photosynthesis, and special effect on the human body.
For the driving scenario, it can recognize such drivers' illegal behaviors as mobile phones, smoking, wearing the seatbelt and mask, and fatigued driving.
For the monitoring video snippet, it can recognize common dangerous behaviors, such as tumbling, smashing, and fighting. At present, it supports single-person and double-person scenarios.
For the software-based human body recognition model service that can be deployed to the local server, you can configure the mainstream Linux system.
It can position and track the human bodies in real-time and perform the population statistics analysis from multiple dimensions. Also, it can monitor the population and give a warning for such potential safety hazards as illegal intrusion and over-crowding in some areas. It can recognize dangerous and illegal behaviors, such as running and jumping in a public area, tumbling, smoking, and making possible the population management and control timely to avoid safety accidents.
It can calculate the number of customers entering and exiting a shopping mall or store, allowing you to perform customer diversion timely. Furthermore, it can be used for business analysis and site selection, and planning. It can recognize the customers' attribute characteristics and pass by the store, collect customers' portraits, and help you perform precise marketing and store site selection. Additionally, it allows you to analyze the track of customers' behaviors and the hot area in your store and provides you with personalized recommendations to optimize the goods display.
For the operation vehicles such as taxis, cars for online hailing, coaches, buses, and trucks, it can recognize the driver‘s dangerous behaviors, including smoking, use of mobile phone, fatigue, and distraction, and no wearing of the seatbelt and mask, and give a warning timely to reduce the accident rate and enhance the driving safety. It can monitor the getting-on and -off for many passengers in the car, analyze the passenger traffic at a stop and over-load of a vehicle, and provide a precise reference basis for the route planning and stop design.
It can analyze the human body gestures and movement trajectory and angle based on the information on the key points of the human body. Also, it assists the athletes in exercise and daily fitness to improve teaching and study efficiency. In the live video platform, you can use such interaction forms as body props, special effects for hand gestures, motion-sensing games, and background replacement, enriching the entertainment experience.
It has 9 human body recognition capabilities and adopts an industry-leading algorithm. In addition, it makes possible the continuous optimization of the Description on the basis of the application feedback and the rapid iteration of the scenario data.
It offers stable and ease-to-use online APIs, offline SDKs, and private deployment package. It is applicable for various terminal access requirements.
It offers such capabilities as face attribute recognition, key point positioning, 1:1 face comparison, 1:N face recognition, and liveness detection.
On the basis of the deep learning and large-scale image training, it can recognize such comprehensive information as object classification, position, level of confidence precisely in an image.
It offers multi-scenario, multi-language and high-precision optical character detection and recognition services. Also, it has several indexes ranked at the first position worldwide.