AI Model Library

Generic Classifier

Classify objects or regions using natural-language text prompts — no training data required.

Generic Detector

Detect objects straight from a natural-language text prompt, matched against what's in the frame — no custom training required.

Vision Language Model

Run vision-language models like GPT-4o, Claude, Gemini and Llama on a region or object to answer questions and extract structured data from video.

Automobile Color Classification

This model can be used to identify the color of automobiles

Automobile Make Classification

This model can be used to identify the make of automobiles

Vehicle Type Classification

This model can be use to identify the type of vehicles

Face Detection

This model detects faces.

Mask Classification

This model identifies if a person is wearing a mask or not.

Vehicle Detection

This model detect vehicles, bicycles, and pedestrians.

People Detection

This model detects people, bags and faces.

License Plate Detection

Detects (but does not read) license plates.

HardHat Detection

This model can detect if a person is wearing a hardhat or not.

Fire Detection

This model detects a possible fire. Note : The model is not 100% accurate in detecting the presence or absence of fires.

Gender Classification

This model can be connected to a face detector and tries to classify the person gender

Age Classification

This model can be connected to a face detector and tries to classify the person age

Vehicle Detection (Dash cameras)

This model detect vehicles, bicycles, and pedestrians. Its use is intended for cameras fitted on dash

Face Detection (Infrared cameras)

This model detect faces. Its use is intended for infrared cameras

Vehicle Detection (High Vantage)

This model was trained for detect vehicles in cameras with high vantage point.

Hands Detection

Detects hands at short distance (ideally at less than 4 meters from the camera)

People Detection (Crowds)

Detects heads, particularly useful for people counting in crowded environments

Vehicle Detection (Low Vantage)

This model was trained for detect vehicles in cameras with low vantage point.

Forklift Detection Lite

Detects Forklifts

License Plate Recognition

Detects License Plates in images and recognizes text on the license plate.

Face Gear Detection

Detects masks, faces, caps, etc.

People Detection (Crowd Bodies)

Detect full bodies and heads of people in crowds

Pose Detection

Detects human body poses

Vehicle ReIdentification

Reidentify vehicles by generating a unique signature for each vehicle.

Person ReIdentification

Reidentify people by generating a unique signature for each person.

People Detection (360 Camera)

People detection model trained on 360 and fisheye cameras. Reference: 2041-yolov5s-800-rotation

Face ReIdentification

Face Reidentification model

Face Pose

Face Pose estimation model

General Purpose 1 (Small)

Detect 80 different classes of objects with over 60% accuracy (lightweight version)

Yolo-NAS

Barcode Detection

Detects 1D barcodes & QR codes

NSFW Classification

Performs the assessment of offensive content, classifying images as 'neutral', 'sexy' or 'porn'

People Detection (Crowds - High Density )

Detect heads of people in high density crowds such as stadiums and events. Requires custom parser - see High Density Crowd Counting template for example.

Age Classification v2

This model can be connected to a face detector and tries to classify the person age

Logistics (small)

Logistics and warehouse model. Detects 20 different objects typically found in warehouses including people, forklifts, pallets and more. This is the smaller, faster variant.

Logistics (medium)

Logistics and warehouse model. Detects 20 different objects typically found in warehouses including people, forklifts, pallets and more. This is the medium variant, slower than small but more accurate.

PPE Detection for Worker Safety

Detects PPE for worker safety : Hard hats, Safety vests, Gloves, Safety shoes, Persons and Heads

Fire v2

Alternate fire detection model. This model detects a possible fire. Note : The model is not 100% accurate in detecting the presence or absence of fires.

License Plate Recognition (Higher Accuracy)

Higher accuracy License plate recognition model that works better for oblique angles.

General Purpose 1 (Medium)

Detect 80 different classes of objects with over 60% accuracy (medium version)

General Purpose 1 (Large)

Detect 80 different classes of objects with over 60% accuracy (larger version)

Adult Child Classifier

This model identifies if a person is an adult or a child

General Purpose 2 (DF)

Detect 365 different classes of objects with over 60% accuracy (requires JetPack >= 6.1 for Jetson deployments)

Emergency Vehicles Type Classifier

This model classifies vehicles as ambulance, fire truck, police, or regular (non-emergency) vehicles

Weapons Detection

Detects guns, with labels big_gun and small_gun.

General Purpose 2 Simple (DF)

Detect 80 different classes of objects with over 60% accuracy (requires JetPack >= 6.1 for Jetson deployments)

Person Outfit Classifier

This model classifies a person’s outfit by detecting the color and type of their top and bottom clothing

Emergency Vehicles Classifier

This model classifies emergency and non-emergency (regular) vehicles