Computer Vision Enablement with AWS Panorama
Easily deploy ML models to on-premise cameras, to reduce time to market for CV projects, and achieve at-scale real-time inference in the cloud with AWS Panorama
Fast-Track Computer
Vision on the Edge
of Innovation
Industries as diverse as manufacturing, healthcare, and retail are looking to take advantage of their existing on-premise cameras to drive value with AI & Computer Vision. On-premise cameras and the systems they connect to are often poorly equipped to run CV models. IT teams need a solution to quickly build and deploy ML models to on-premise cameras, to enable real-time inference and process automation at scale. Amazon Web Services (AWS) has resolved this problem by releasing AWS Panorama.
What Is AWS Panorama?
AWS Panorama is a fully managed service that makes it easier to deploy machine learning models to on-premise cameras, enabling companies to bring flexibility to their computer vision projects through at-scale real-time inference and process automation.
AWS Panorama Is Used For
Quality
Control
Part
Identification
Operational
Compliance
Workplace
Safety
Barcode
Scanning
Epidemiological
Safety
Package
Inspection
Tracking
& Tracing
AWS Panorama Components
AWS Panorama consists of two components: A hardware component — AWS Panorama Appliance, and a software component — AWS Panorama SDK.
AWS Panorama Appliance
A machine learning appliance installed on an enterprise network to discover, connect to, and manage networked cameras* and run simultaneous ML models on multiple concurrent video streams. The appliance also serves as a flexible CV application management system to improve ML models over time, to meet evolving application needs.
AWS Panorama SDK
A console that makes it easier to inspect and manage network devices supported by or compatible with AWS Panorama. It also allows for building and deployment of AWS Panorama apps to devices in the AWS Panorama network, and creation or modification of AWS Lambda functions for model inference or data analysis. It supports a wide range of AWS services.
*The appliance can be integrated with less capable industrial cameras without any built-in machine learning capabilities. It supports up to 8 connected cameras with a resolution of 1080p, at a frame rate of 30fps.
How It Works
Computer Vision solutions enabled by AWS Panorama are connected to the AWS cloud, and process video streams from cameras within an AWS Panorama Appliance network, with results delivered to external business processes.
Provectus — An AWS Panorama Launch Partner
Provectus is proud to be an AWS Panorama launch partner. We are committed to helping our customers leverage AWS Panorama to build computer vision solutions that ensure product quality, enhance the efficiency of industrial processes, and enforce worker and epidemiological safety.
What We Offer
POC in a Box is a joint program by Provectus and Amazon Web Services, geared to creating use case-based, deployment-ready POCs on AWS Panorama, in only four days of hands-on engagement.
POC in a Box Framework
Day 1–3
POC Planning
Provectus defines the engagement’s goals and success criteria with customer stakeholders, assesses their data readiness, and checks compatibility of on-premise cameras with AWS Panorama.
POC Workshop
Provectus explores potential use cases, builds initial models, and evaluates the metrics. All active models are evaluated to outline the path to production.
Day 4–5
A/B Testing
Provectus runs the model to test how it performs on live video streams captured by AWS Panorama-enabled cameras. The model is fine-tuned to improve performance.
POC Evaluation
Provectus evaluates A/B test results against initially defined goals and success criteria. The next steps required to move POC to production are provided.
Machine Learning Competency Partner
Provectus is an AWS Premier Consulting Partner with a strong track record of CV solutions. AWS Data & Analytics, DevOps, and Machine Learning Competency Partner.
