Human body analysis _ human body recognition scheme based on deep learning

Accurately recognize the human body related information in the image, and provide human body detection and tracking, key point positioning, pedestrian flow statistics, attribute recognition, behavior analysis, portrait segmentation, gesture recognition, fingertip detection and other capabilities. Support private deployment of public cloud, local server or proprietary cloud server and multiple product forms of EdgeBoard software and hardware integration to ensure the privacy and security of business data

case The exhibition hall can be more "smart"! Baidu Brain AI Smart Stadium Landing in China Science and Technology Museum >

Product List

Application case

Characteristic advantages

Usage

Support communication

Related recommendations

Product List

Human body key point recognition

Locating 21 core key points of human body, including facial features, neck, limbs, etc., supporting complex scenes such as multiple people and big movements

Learn more

People flow statistics

Count the number of human bodies and flow trend in the image, and take the head and shoulder as the target to identify the number of people, without face and full body photos

Learn more

Human body detection and attribute recognition

Detect all the human bodies in the image, and identify more than 20 attributes of the human body, including gender, age, clothing, behavior, etc

Learn more

Portrait segmentation

Identify the contour range of the human body and separate it from the background, which is suitable for taking pictures of background replacement, photo synthesis, human special effects and other scenes

Learn more

gesture recognition

Recognize 24 common gestures, such as OK, fist, heart comparison, like, and number, and support self taking, others taking, and various shooting angles

Learn more

Hand key point recognition

Locate 21 main bone nodes of the hand, which can be used to customize gesture detection, AR special effects, human-computer interaction and other scenes

Learn more

fingertip detection

Accurately locate the coordinate position of the index finger tip, especially suitable for children's learning machine reading scene, to achieve point reading, question searching, automatic tracking, etc

Learn more

Driving behavior analysis

Identify drivers' violations such as using mobile phones, smoking, not wearing seat belts, not wearing masks, fatigue, etc

Learn more

Human privatization deployment package

The "pure software" human body recognition model service that can be deployed to the local server is suitable for mainstream Linux systems

Learn more

Human body analysis software and hardware integration scheme

Based on EdgeBoard edge AI computing box, with built-in body analysis algorithm, it can be deployed offline locally, with good software hardware integration effect

Learn more

Application case

Security monitoring

Driving monitoring

Sports entertainment

Smart retail

Security monitoring

Real time positioning and tracking of human body for multi-dimensional population statistical analysis. Monitor the flow of people and warn of potential safety hazards such as illegal intrusion of personnel and too dense crowd in local areas; Identify dangerous, illegal and other abnormal behaviors, such as running, jumping, falling, smoking in public places, timely control, and avoid safety accidents

Cooperation cases

Characteristic advantages

Leading technology

9 human recognition capabilities, leading algorithm effect in the industry, and fast iteration based on application feedback and scene data, continuous optimization

In depth scenario

For vehicle scenarios, combined with industry attributes, special training driving behavior analysis model, more vertical services continue to expand

Flexible and easy to use

Provide stable and easy to use online APIs, offline SDKs, and privatization deployment packages to adapt to various terminal access requirements

Usage

Online interface

Provide online human body key point recognition, human body detection and attribute recognition, traffic statistics, portrait segmentation, gesture recognition, and driving behavior analysis interfaces, which can be used at the console for free

API documentation

Use Now

Privatization deployment

Deploy to the local server (GPU resources are required) to realize functions such as human detection, key point positioning, attribute identification, and pedestrian flow statistics, which are suitable for security monitoring, park management, and other scenarios

Scheme introduction

Apply for free trial

Software and hardware integration scheme

The scheme uses the EdgeBoard edge AI computing box as the hardware carrier, which can realize model reasoning offline at the edge side, effectively saving the server hardware and bandwidth transmission costs, and can also achieve high-performance real-time detection and identification in a network free and weak network environment

Buy Now

Working with Documents

Support communication

Novice Guide

Basic introduction of access to AI services, including account authentication, application creation and management, authentication mechanism, FAQs, etc

View details>

Server SDK download

The standard encapsulated HTTP SDK supports six languages, including Java and PHP, and quickly integrates API calls on the server side

Download Now>

Community communication

Rich technical sections and essence content, communicate with tens of thousands of developers online, and get started quickly

Join now>