background_AI_2-1200x800.jpg

EdgeFirst Vision Resources 

Hotline Consultant

DeepView Support

To access additional information around our DeepView Zoo's, DeepView VisionPack, DeepView ModelPack, Demo Applications and to submit a support ticket please click the link below.

Minimalist Square Photo Instagram Post (33).png

DeepView Dataset Zoo

The ModelPack Dataset Zoo provides various example datasets which can be opened directly into eIQ Portal to train a ModelPack model.  Each dataset also provides pre-trained models which can be used directly by VisionPack and full checkpoints to evaluate results in eIQ Portal without the need to re-train.

Minimalist Square Photo Instagram Post.png

DeepView Vision Application Zoo

A collection of EdgeFirst vision applications demos powered by our DeepView VisionPack and DeepView ModelPack.
 

Real-world, easy to run examples you can perform at home or in the office.

DeepView Middleware Evaluation Trials

Go from idea to concept to production faster and with more confidence using our DeepView AI Middleware. We’ve optimized runtime and model performance so you can focus on your vision application.

Click the link below for signup for an evaluation trial of our DeepView VisionPack and DeepView Model Pack.

DeepView VisionPackProduct Brief
Accelerated vision pipeline for NPU,CPU, GPU

DeepView ModelPackProduct Brief
State of the art detection models 

Download the latest version of the eIQ Portal to try out our DeepView ModelPack Evaluation Trial

 
6xYHts3.webp
 

DeepView Vision Application Zoo

Minimalist Square Photo Instagram Post.png

Head Pose | NPU, GPU, CPU

The Head Pose application showcases real-time face detection and pose estimation. The results are graphically overlaid over the live camera feed to show the head yaw, pitch roll vectors.

Minimalist Square Photo Instagram Post (22).png

Face Detection | NPU, GPU, CPU

The Face Detection application uses a custom AI model to find human faces in real-time and generate bounding boxes on screen. Works reliably from 1 - 25+ meters

Minimalist Square Photo Instagram Post (11).png

Face Tracking | NPU, GPU, CPU

Minimalist Square Photo Instagram Post (5).png

Body Pose | NPU, GPU, CPU

The Face Tracking application allows to find human faces in real-time and track their movement. The results are graphically overlaid with bounding boxes over the live camera feed and show the last few seconds of track history as trails.

The Body Pose estimation application showcases real-time body detection and pose estimation. This type of application can be used for fitness or gaming applications and other types of human machine interfacing use cases. 

Minimalist Square Photo Instagram Post (3).png

Face Blurring | NPU, GPU, CPU

This application shows a live video streaming remote camera which anonymizes faces to provide privacy.  This allows crowd monitoring while alleviating privacy concerns.

Minimalist Square Photo Instagram Post (24).png

Visit DeepView.Zendesk for more EdgeFirst vision applications

Video Library

For More Videos
Visit our YouTube Channel
 

Blogs & Whitepapers

Picture1- safety (1).jpg

Monitoring PPE in Hazardous Work Spaces

In this article we present a framework for using Computer Vision at the edge for monitoring hazardous workplaces. Our framework can run in real-time (from 30 fps to 60 fps) on an embedded target device (EVK, Maivin) without sacrificing accuracy.

Minimalist Square Photo Instagram Post (23).png

DNN-Based Object Detectors

In this blog, we provide a technical introduction to deep-neural-network-based object detectors. We explain how these algorithms work, and how they have evolved in recent years, utilizing examples of popular object detectors.

Minimalist Square Photo Instagram Post (35).png

ML at the Edge: Visual Intelligence with a low cost MCU

A usable machine learning model required diverse and robust dataset. We will be going over a step-by-step breakdown of the process of the dataset creation as well as how different image conditions can increase the accuracy of the model.

 

Use Case Examples

Modern Minimal Technology Background Youtube Thumbnail (8).png

Urban Pothole Detection Use Case

Leveraging the current public transport system in Canada paired with ML/AI to pinpoint the location of detected potholes and reporting them directly based on their priority.

 

Product Information

DeepViewRT Benchmark Data

The DeepViewRT run time inference engine provides developers with the freedom to quickly deploy ML models to a broad selection of embedded devices and compute architectures without sacrificing flexibility or performance.

Product Brief
DeepView ModelPack for NPU, MPU

DeepView ModelPack is a bundle of state-of-the-art detection models pretrained and has been fully tested & optimized for NXP RT Crossover MCU’s and i.MX8 Application Processors.

Au-Zone Evaluation Trials and Demo Applications

A list of all available DeepView AI Middlware evaluation trials and application demos available on the Maivin AI Vision Starter Kit and Micro AI Vision Starter Kit. 

Product Brief
Maivin AI Vision Starter Kit

The Maivin AI Vision Starter Kit is a modular AI smart camera platform built on NXP’s i.MX8MPlus Applications Processor and production grade hardware and software components to enable rapid prototyping and field deployment of custom Vision Solutions. The Maivin targets applications where compute performance is the priority.

Product Brief
DeepView VisionPack for NPU, GPU, CPU

DeepView VisionPack provides developers with an end to end, hardware accelerated video pipeline for optimized AI based Vision applications. Fully integrated with DeepViewRT inference engine for high performance / low overhead AI vision solution on Applications Processors.

Product Brief
Micro AI Vision Starter Kit

The Micro Vision Starter Kit is a modular AI smart camera platform built on NXP’s i.MXRT1064 Crossover MCU and production grade hardware and software components to enable rapid prototyping and field deployment of custom Vision Solutions. The Micro targets applications where lower cost, design complexity and size are prioritized over computer performance.