Free Online AI Inference to Accelerate the Proof-of-Concept Procedures
Aetina AI is a free no-code web-based edge hardware evaluation and recommendation engine that provides system recommendations based on the requirements of specific AI models. The service takes the guesswork out of matching AI models with the correct hardware, helping developers and enterprises accelerate proof-of-concept procedures and redefine the development of AI projects.
Instant Access to Many AI Models
No-risk free trial gives you a simple way to test AI model performance online.
Smart City
Smart Factory
Smart Retail
Smart Healthcare
FP32
Intersection Safety
AI Model：PeopleNet
People Detection
The model detects one or more people in a crowd. The people detection AI model is a high-precision pre-trained AI model based on FP32.
Scenarios
For non-intersection safety scenarios, it can be used for access control, flow analysis, and people counting in factories or retail, as well as intrusion detection for restricted areas or public area safety.
For traffic such as urban crossroads or intersections, it can detect whether pedestrians run red lights, jaywalk, etc., and give timely warnings to avoid accidents.
The model detects one or more cars from a moving vehicle within the video or an image. The inference runs on a pruned model at INT8 precision and intended to be used with DeepStream SDK or TensorRT for efficient AI-based video analysis deployments.
Scenarios
To detect vehicles on the road, meanwhile detect illegal behaviors such as speeding, illegal occupancy, U-turns, entering prohibited sections, etc., to assist law enforcement and improve traffic management efficiently with computer vision and artificial intelligence.
The model detects the American license plate in an image and returns a bounding box and the license plate number.
Scenarios
Carry out license plate recognition for vehicles at the entrances or exits in parking lots, buildings, or construction sites to improve the efficiency of vehicle management.
The model aims to detect and predict one or multi-person human pose estimation by detecting the person's skeleton in an image, consisting of keypoints and the connections in between.
Scenarios
Motion Recognition: Recognize the movement, pose, and behavior of people, which can be applied to fall detection, intrusion detection, and further motion detection suitable in healthcare, security, and smart retail.
It also can be applied to human-machine cooperation in robotic factories to detect person pose and behavior by setting virtual fences around the robotic arms. It is set to slow down or stop the operation of robotic arms when workers head inside the fence to enhance safety and avoid collision.
The model detects face masks in an image with computer vision and deep learning using TensorFlow.
Scenarios
Enforcement of the wearing of face masks, no matter whether in hospitals, medical environments, laboratories, construction sites, or workplaces, the wearing of face masks can be effectively monitored.
The model detects one or more people in a crowd. The people detection AI model is a high-precision pre-trained AI model based on FP32.
Scenarios
For non-intersection safety scenarios, it can be used for access control, flow analysis, and people counting in factories or retail, as well as intrusion detection for restricted areas or public area safety.
For traffic such as urban crossroads or intersections, it can detect whether pedestrians run red lights, jaywalk, etc., and give timely warnings to avoid accidents.
The model detects one or more people in a crowd. The people detection AI model is a high-precision pre-trained AI model based on FP32.
Scenarios
For non-intersection safety scenarios, it can be used for access control, flow, and people counting in the factories or retails, as well as intrusion detection for restricted areas or public area safety.
For traffic such as urban crossroads or intersections, it can detect whether pedestrians run red lights, jaywalk, etc., and give timely warnings to avoid accidents.
The model aims to detect and classify human emotion into 6 categories, including Neutral, Happy, Surprise, Squint, Disgust, and Scream.
Scenarios
Facial emotion recognition based on computer vision can assist in analyzing the attractiveness of ads or the shopping experience, helping retailers to optimize consumer satisfaction.
Furthermore, assisting doctors to be aware of the diseases through the facial manifestations, assisting teachers in observing students’ concentration in learning, or giving warning signs of fatigue driving.
The AI model provided here aims to detect and classify 3 different categories of fruit: Apple, Banana, and Orange.
Scenarios
When it comes to non-packaged items, like fruits and vegetables, use the image-recognition based on powerful computer vision and artificial intelligence to accelerate the self-checkout and improve the in-store shopping experience.
The model aims to detect and predict one or multi-person human pose estimation by detecting the person's skeleton in an image, consisting of keypoints and the connections in between.
Scenarios
Motion Recognition: Recognize the movement, pose and behavior of people, which can be applied to fall detection, intrusion detection, and further motion detection suitable in healthcare, security, and smart retail.
It also can be applied to human-machine cooperation in robotic factories to detect person pose and behavior by setting virtual fences around the robotic arms. It is set to slow down or stop the operation of robotic arms when workers head inside the fence to enhance safety and avoid collision.
The model detects face masks in an image with computer vision and deep learning using TensorFlow
Scenarios
Enforcement of the wearing of face masks, no matter whether in hospitals, medical environments, laboratories, construction sites, or workplaces, the wearing of face masks can be effectively monitored.
The model aims to detect and classify human emotion into 6 categories, including Neutral, Happy, Surprise, Squint, Disgust, and Scream.
Scenarios
Facial emotion recognition based on computer vision can assist in analyzing the attractiveness of ads or the shopping experience, helping retailers to optimize consumer satisfaction.
Furthermore, assisting doctors to be aware of the diseases through the facial manifestations, assisting teachers in observing students’ concentration in learning, or giving warning signs of fatigue driving.
Access pre-trained models and learn inference in minutes directly in your web browser. No more tedious software installation or code imports.
Rapid Hardware Evaluation
Free test of instant inference on an edge device remotely and benchmark the edge hardware before you buy.
Confidence to Scale
Speed up the orchestration of custom projects from POC to production with seamless collaboration across Aetina and AI ecosystem partners.
Aetina AI
Aetina AI
Aetina AI brings innovative solutions that closely meet the needs of Artificial Intelligence and Edge Computing and usher them from concept to the real world. We help clients and developers grow their AI business with end-to-end AI management services, a wide range of AI computing systems, and application-oriented customization services. Aetina AI brings together AI partners and their technical expertise from hardware and software to sensors and services, to create a one-stop platform to accelerate your go-to-market projects.
↑
This website uses Cookies to improve the website and user experience. By continuing to browse this website, you confirm and agree to our privacy policy (including the use of Cookies). For the latest and more related content, please refer to ourPrivacy Policy.
FP32
INT8
FP32
FP32
FP32
FP32
FP32
FP16
FP32
FP32
FP32
FP32
