NGC | Catalog
Welcome Guest
CatalogContainersDeepVision - Face Recognition

DeepVision - Face Recognition

For pull tags and more information, please view on a desktop device.
Logo for DeepVision - Face Recognition

Description

Deep Vision AI Inc. applies proprietary advanced computer vision technology to understand images and video automatically, turning visual content into real-time analytics and valuable insights.

Publisher

DeepVision

Latest Tag

onpremise-3.1.2

Modified

September 24, 2020

Compressed Size

4.71 GB

Multinode Support

No

Multi-Arch Support

No

Deep Vision AI

Advanced computer vision technology to understand images and video automatically, turning visual content into real-time analytics and valuable insights. Deep Vision AI platform delivers facial recognition and vehicle recognition analytics used for cutting - edge safety and security, as well as business intelligence solutions. Team of experts in computer vision and machine learning includes contest-winning ImageNet researchers with strong technical competencies and commitment to innovation. Benefits The solution is camera agnostic and can be plugged into existing and new camera infrastructure. Highly optimized AI models with 15 streams performance per GPU, cost effective solution. Supports cloud, on-premise, edge-based or combined deployments scenarios. New custom software modules are built upon request to meet unique customer needs. GDPR compliant

(Admin Console Container)[https://ngc.nvidia.com/catalog/containers/partners:deepvision:admin.console]

Vehicle Recognition 1.1.1_4e408a

Deep Vision AIÕs Vehicle Recognition Model has the unique ability to count and recognize the year, make, model and license plates of vehicles from any angle. Governments and municipalities use vehicle recognition to automatically analyze vehicle flows and send alarms reporting designated vehicles to law enforcement. The model is also used to infer demographics based on vehicle recognition, quantify vehicle flow, and assess changing traffic patterns. Advertisers and brands also use this information to target contextualized ads based on the changing mix of demographics and to understand ROI of outdoor advertising. Types: Vehicle Analytics - Vehicle counting - Year, make and model recognition - License Plate recognition (Vehicle Inspection Container)[https://ngc.nvidia.com/catalog/containers/partners:deepvision:vf.vehicle]

People Counting and Age/ Gender Demographics Version: 2.0.0

Deep Vision AI demographics software module continuously monitors target zones to provide the count, gender, age and unique identification of individuals over time. ItÕs used to understand demographic variations over time for a designated area of the city, or to track customer patterns such as dwell-time spent in lines or waiting areas of retail stores. It also helps brands and advertisers to quantify demographics or to target individuals for advertising and product placement. Types: Age and gender recognition - People counting

(Demographics Container) [https://ngc.nvidia.com/catalog/containers/partners:deepvision:vf.demographics]

Facial Recognition Version: 3.1.2

Recognition software module continuously monitors target zones to provide the count, gender, age and unique identification of individuals over time. The Facial Recognition Model tracks unique individuals and provides facial matches for specified individuals. This allows to improve overall safety and security, extract greater value from traditional security infrastructure, help retailers recognize important customers in real-time or quantify the frequency of visitors. Types: Face Analytics - Watchlist
(Facial recognition Container)[https://ngc.nvidia.com/catalog/containers/partners:deepvision:vf.face.recognition]

DeepVision Helm Chart

[https://ngc.nvidia.com/catalog/helm-charts]

Prerequisites

  • Nvidia EGX stack
  • NVIDIA Tesla T4

License

  1. Get licence_info file. From the cluster node targeted to host the application, run: $ mkdir licence_info; docker run -it --rm --name get_licence_info --privileged -v $PWD/licence_info:/root/licence nvcr.io/metropolis/deepvision/vf.vehicle:onpremise-1.1.1_4e408a sh -c "get_licence_info"

  2. Send the licence_info binary file located in the licence_info directory to .

Quickstart Guide

  1. Log in to your account on the NGC website https://ngc.nvidia.com OR create a new NGC account

  2. Get your NGC API Key Click Setup from the side menu, then click API Key from the Setup page, then Generate API Key

  3. Download the Helm Chart from NGC Helm Registry: Deepvision Helm Chart

  4. Edit the content of deepvision-vehicle/values.yaml to match your deployment

  5. On a terminal connected to EGX Manager: $ kubectl get nodes $ kubectl label nodes nvidia.com/sku=

  6. Deploy Vehicle Recognition $ helm install deepvision-vehicle

  7. Check status $ kubectl get pods | grep vehicle

  8. Connect to User Interface (Admin Console) on a browser http://:

  9. Request license for camera infrastructure: Go to Admin Console Administration Click on Manage Cameras Select all cameras you need to activate Click on Multiple Actions Select Activate Click on Apply Button Fill form, provide email, select time interval You will receive serial numbers for licenses required.

#Terms and Condition DeepVision