Pretrained Models


The OpenVINO™ toolkit includes two sets of optimized models that can expedite development and improve image processing pipelines for Intel® processors. Use these models for development and production deployment without the need to search for or to train your own models.

Public Model Set

Download and incorporate some of the most popular models created by the open developer community using the included Model Downloader. Add these models directly to your environment and accelerate your development.

Find the downloader in this toolkit folder: \deployment_tools\model_downloader.

 

Free Model Set

Discover the capabilities of Intel® software and silicon with a fully functioning set of pretrained models. These models provide common vision use cases and reduce development time and cost. Documentation for each model includes links to public data.

Set Up a Machine for Running Models

Minimum System Requirements

Age & Gender Recognition

This neural network-based model provides age and gender estimates with enough accuracy to help you focus your marketing efforts.

Find the model in this toolkit folder: \deployment_tools\intel_models\age-gender-recognition-retail-0013.

Face Detection

Standard Model

Identify faces for a variety of uses, such as observing if passengers are in a vehicle or counting indoor pedestrian traffic. Combine it with a person detector to identify who is coming and going.

Find the model in this toolkit folder: \deployment_tools\intel_models\face-detection-adas-001.

Enhanced Model

While similar to the standard model, this model performs better in a wider range of lighting conditions. The detector backbone is SqueezeNet light (half-channels) with a single-shot detector (SSD) for shooting indoor and outdoor scenes with a front-facing camera.

Find the model in this toolkit folder: \deployment_tools\intel_models\face-detection-retail-0004.

Retail Environment Model

Different colored bounding boxes simultaneously detect a head and an entire person. Based on a backbone similar to MobileNetV2, the model includes depth-wise convolutions that reduce computation for a 3 x 3 convolution block.

Find the model in this toolkit folder: \deployment_tools\intel_models\face-person-detection-retail-0002.

Head Position

This model shows the position of the head, and provides guidance on what caught the subject's attention.

Find the model in this toolkit folder: \deployment_tools\intel_models\head-pose-estimation-adas-0001.

Note: This model does not capture the subject's gazing direction.

Emotion Recognition

Identify neutral, happy, sad, surprise, and angry emotions.

Find the model in this toolkit folder: \deployment_tools\intel_models\emotions-recognition-retail-0003.

five identified face landmarks

Landmark Detection

This lightweight regressor model identifies five facial landmarks: two eyes, a nose, and two lip corners. The model is best suited for smart classroom use cases.

Find the model in this toolkit folder: \deployment_tools\intel_models\landmarks-regression-retail-0001.

example of identifying faces

Face Re-Identification

Use this lightweight network for face re-identification in smart classroom scenarios. For best results, use a frontally oriented and aligned input face.

Find the model in this toolkit folder: \deployment_tools\intel_models\face-reidentification-retail-0001

Human Detection

Eye-Level Detection

View the number of people in a frame at any given time. This model performs best when the camera angle is approximately at eye level and is based on the hyper-feature (R-FCN) backbone.

Find the model in this toolkit folder: \deployment_tools\intel_models\face-detection-retail-0001.

Note: This model does not work with the Intel® Movidius™ Neural Compute Stick.

High-Angle Detection

Use this model for cameras mounted at higher vantage points to count the people in a frame.

Find the model in this toolkit folder: \deployment_tools\intel_models\person-detection-retail-0013.

Detect People, Vehicles, & Bikes

Distinguish between people, people riding bikes, bikes alone, and vehicles. A variety of lighting conditions in this model improve accuracy in daylight, darkness, and variations in the weather.

Find the model in this toolkit folder: \deployment_tools\intel_models\person-vehicle-bike-detection-crossroad-0078.

Pedestrian Detection

Distinguish between people and objects in public using a network that is based on an SSD framework with a tuned MobileNetV1 as a feature extractor.

Find the model in this toolkit folder: \deployment_tools\intel_models\pedestrian-detection-adas-0002.

Pedestrian & Vehicle Detection

Identify people and vehicles by using a network that is based on an SSD framework with a tuned MobileNetV1 as a feature extractor.

Find the model in this toolkit folder: \deployment_tools\intel_models\pedestrian-and-vehicle-detector-adas-0001.

Identify Someone in Different Videos

Use whole-body images to identify one person in different video streams.

Standard Model

Find the model in this toolkit folder: \deployment_tools\intel_models\person-reidentification-retail-0076.

Faster Model

Find the model in this toolkit folder: \deployment_tools\intel_models\person-reidentification-retail-0079.

Ultra-Small & Ultra-Fast Model

Achieve a further performance boost with the model in this toolkit folder: \deployment_tools\intel_models\person-reidentification-retail-0031

Pedestrian Attributes

Identify key attributes of a person crossing the road: gender, hat, long sleeves, long pants, long hair, coat, and jacket.

Find the model in this toolkit folder: \deployment_tools\intel_models\person-attributes-recognition-crossroad-0031.

individual action detection

Action Detection for a Smart Classroom

This model recognizes poses that include sitting, standing, and raising a hand. Use this action detector for a smart classroom scenario based on the RMNet backbone with depthwise convolutions.

Find the model in this toolkit folder: \deployment_tools\intel_models\person-detection-action-recognition-0001

Vehicle Feature Recognition

Vehicle Detection

Identify vehicles by applying an SSD framework that uses a tuned MobileNetV1 to extract features.

Find the model in this toolkit folder: \deployment_tools\intel_models\vehicle-detection-adas-0002.

License Plate Detection: Small-Footprint Network

This model has a small-footprint network that's trained to recognize (but not read) a variety of Chinese license plates.

Find the model in this toolkit folder: \deployment_tools\intel_models\license-plate-recognition-barrier-0001.

a recognized license plate

License Plate Detection: Front-Facing Camera

This MobileNetV2 and SSD-based vehicle and license plate detector recognizes Chinese license plates from a front-facing camera. This model is useful for security barriers that require front license plate detection.

Note: This model replaces the previous version and runs faster while maintaining the same accuracy.

Find the model in this toolkit folder: \deployment_tools\intel_models\vehicle-license-plate-detection-barrier-0106.

Vehicle Metadata

Conduct an initial analysis and present back-key metadata for faster sorting and searching in the future. The average color accuracy for the model is over 82 percent for red, white, black, green, yellow, gray, and blue. Its average vehicle-type attribution is over 87 percent for cars, vans, trucks, and buses.

Find the model in this toolkit folder: \deployment_tools\intel_models\vehicle-attributes-recognition-barrier-0039.

Identify Roadside Objects

Classify objects as roads, curbs, painted lines, or backgrounds.

Find the model in this toolkit folder: \deployment_tools\intel_models\road-segmentation-adas-0001.

Advanced Roadside Identification

Classify objects as roads, sidewalks, buildings, walls, fences, poles, traffic lights, traffic signs, vegetation, terrain, sky, people, passengers, cars, trucks, buses, trains, motorcycles, bicycles, or electric vehicles.

Find the model in this toolkit folder: \deployment_tools\intel_models\semantic-segmentation-adas-0001.

Supported Samples

Reference this table for components that support the pretrained models.

Pretrained Model Supported Samples CPU Integrated Graphics FPGA VPU
face-detection-adas-0001 Interactive face detection  
age-gender-recognition-retail-0013 Interactive face detection
head-pose-estimation-adas-0001 Interactive face detection
emotions-recognition-retail-0003 Interactive face detection  
vehicle-license-plate-detection-barrier-0007 Security barrier camera
vehicle-attributes-recognition-barrier-0039 Security barrier camera
license-plate-recognition-barrier-0001 Security barrier camera
road-segmentation-adas-0001 Image segmentation  
semantic-segmentation-adas-0001 Image segmentation  
person-detection-retail-0001 Object detection  
person-attributes-recognition-crossroad-0031 Crossroad  
pedestrian-detection-adas-0002 Any SSD-based sample  
pedestrian-and-vehicle-detector-adas-0001 Any SSD-based sample  
person-detection-retail-00013 Any SSD-based sample  
face-detection-retail-0004 Any SSD-based sample
face-person-detection-retail-0002 Any SSD-based sample  
person-vehicle-bike-detection-crossroad-0078 Any SSD-based sample  
vehicle-detection-adas-0002 Any SSD-based sample  

Note: FPGA support comes via heterogeneous execution, which means that post-processing occurs on the CPU.

Face Detector

This algorithm deploys a convolutional neural network (CNN) model that's pretrained to detect people's faces in different positions within the camera view. The face detector can run on Intel® processors, integrated graphics from Intel, and the Intel® Movidius™ Neural Compute Stick.

Age & Gender Recognizer

This algorithm helps deploy a deep learning-based face analyzer that accurately recognizes people's ages and genders. Supported targets include Intel processors, integrated graphics from Intel, and the Intel® Movidius™ Neural Compute Stick.

Camera-Tampering Detector

This algorithm is intended to recognize malicious attacks on a camera. It detects camera-tampering events such as occlusion, defocusing, or displacement using classical computer vision approach. The supported deployment target is an Intel processor.

Emotions Recognizer

Readily deploy a CNN model that's pretrained for emotion classification on a single image or a batch of images. This component supports Intel processors and integrated graphics from Intel.

Person Re-Identification

This algorithm deploys a CNN-based pretrained model that re-identifies a person in multiple images. Its supported targets include Intel processors and integrated graphics from Intel.

Crossroad Object Detection

Successfully detect, identify, and re-identify persons and vehicles at crossroads. Supported targets include Intel processors and integrated graphics from Intel.

Minimum System Requirements 

CPU:

  • 6th to 8th generation Intel® Core™ and Intel® Xeon® processors
  • Intel® Pentium® processor N4200/5, N3350/5, N3450/5 with Intel® HD Graphics

Graphics:

  • 6th to 8th generation Intel Core processor with Iris® Pro graphics and Intel HD Graphics
  • 6th to 8th generation Intel Xeon processor with Iris Pro graphics and Intel HD Graphics (excluding the e5 product family, which does not have graphics)