NVIDIA MDX Early Access - Getting Started
Generating business insights and automating infrastructure requires perception and situational awareness combined with higher level reasoning. Detection and classification of objects is just the beginning. Making sense of the spatial and temporal context of streaming sensor metadata, understanding behavior, plotting trends and estimating what might happen next is critical to modern AI applications.
NVIDIA MDX (Metropolis Data Analytics Toolkit) is a complete AI data analytics SDK to help build higher levels of reasoning and understanding into applications leveraging data generated from streaming sensor perception metadata. It can unlock business insights like building occupancy, heatmap, flow of traffic and identify anomalies, then turn them into an action plan.
From Perception to Reasoning
The journey starts with sensors, especially cameras which have been deployed all over businesses and cities. To extract business insights, you start by doing perception - where you convert sensor data to metadata. Data analytics is taking this metadata from 10s/100s of sensors, fusing the data, applying spatio-temporal context for behavior understanding, and creating alerts and data visualization. MDX accelerates the data analytic portion of the solution - taking unstructured sensor data and generating actionable business insights.
MDX is now open for early access by invitation only.
Metropolis Data Analytics Toolkit (MDX) EA
- Reference applications including:
- Analyzing city wide traffic with 100+ cameras, and turning it into searchable traffic behavior
- Measuring pedestrian utilization of space, people counts, heat-maps and behavior
- Includes pre-built rules for creating line crossing and Region-of-interest anomalies
- Extract geo-spatial information with browser based camera calibration
- Analyze object behaviors like speed, distance, direction and anomalies
- Explore raw and filtered data with Kibana dashboards
- Connect to any edge “perception” apps using MDX connectors
- REST APIs to extract and visualize insights
- Cloud-native deployment using Helm Charts and Kubernetes
Getting Started Resources
- Download all the sample apps, configs, scripts - MDX Software package
- Download all the media and metadata to use with reference apps - MDX data package
- Download the packages for K8 based deployment - MDX K8 deployment
- Pull containers from private NGC repositories
- Pull Helm Chart from private NGC repositories
- MDX Developer guide
- Reference apps with MDX
- MDX API Guide
- Behavior Learning
- Analytic Stream
- Release notes
- Post your MDX related questions or feedback in the MDX EA developer forum.
NVIDIA’s platforms and application frameworks enable developers to build a wide array of AI applications. Consider potential algorithmic bias when choosing or creating the models being deployed. Work with the model’s developer to ensure that it meets the requirements for the relevant industry and use case; that the necessary instruction and documentation are provided to understand error rates, confidence intervals, and results; and that the model is being used under the conditions and in the manner intended.