This hands-on workshop teaches you the foundational elements of intelligent video analytics (IVA) by integrating deep learning models and algorithms into a ready-to-use NVIDIA DeepStream pipeline. First, you’ll gain an in-depth understanding of data normalization, annotation, and metadata formatting in IVA applications, followed by a fundamental understanding of object detection in video frames. You’ll proceed to train object tracking models and learn how to utilize temporal features of videos to achieve more efficient detection models. You’ll also learn to deploy and accelerate the development of your IVA application by plugging deep neural networks into an end-to-end stream processing pipelineusing the DeepStream framework. Throughout the workshop, you’ll get hands-on coding experience using a live GPU-accelerated environment to deploy and train your models. At the end, you’ll have access to additional resources to design and deploy IVA applications on your own.
NVIDIA-Deep Learning for Intelligent Video Analytics (DL-IVA)
Schedule
- No schedule events found for this course.
- PC
Private Class
Privately train a group of your employees at your facility, virtually, or any of our locations.
- PC
- LCLive Classroom
Live Classroom
Learn and interact with your instructor and peers in-person in our classrooms. - VCVirtual Classroom
Virtual Classroom
Attend any of our instructor-led classes virtually regardless of your physical location. - PCPrivate Class
Private Class
Privately train a group of your employees at your facility, virtually, or any of our locations. - GTRGuaranteed to Run
Guaranteed to Run
GTR classes are guaranteed to run as promised and delivered.
Course Summary
Show All
Description
Objectives
- Wrangle video data and perform raw data ingestion into underlying models
- Deploy deep learning models for accurate and effective object detection and tracking applications
- Accelerate the development of IVA applications by using the DeepStream frame
Prerequisites
Experience with deep networks (specifically variations of CNNs); intermediatelevel experience with C++ and Python
Outline
Object Detection for Intelligent Video Analytics (IVA) (120 minutes)
- Prepare metadata for object detection models and learn the fundamentals of Object Detection API
- Apply frame-by-frame detection models and compare their KPI results and utilize advanced techniques of qualitative analysis and performance measurement
Learn the fundamentals of object detection methods in IVA applications, as well as preliminaries of raw data processing and metadata formatting. Get hands-on experience with the Object Detection API and learn how to measure accuracy and performance of the models using IoU metrics.
Using Transfer Learning and Multiple Object Tracking Techniques in IVA (120 minutes)
- Generate TFRecords to optimize the training performance and use transfer learning techniques to fine tune the model
- Measure and visualize model performance
Get familiar with the nuances of fine tuning an IVA application and learn about the implications of modeling. Understand how object detectors can be bootstrapped into your IVA application.
Deploying the Application Using NVIDIA DeepStream (120 minutes)
- Create and deploy easily configurable, end-to-end video processing pipelines
- Learn how to easily plugin multiple inference models, and explore methods for visualizing the inference data.
Learn to deploy the video analytics models into a ready-to-use video processing pipeline using DeepStream and understand the fundamentals of creating robust smart city applications. Learn about various plugins to embed into your pipeline, as well as NVIDIA's dedicated video codec plugins.
Next Steps and Q&A (45 Mins)
- Discuss the next-steps and questions.
Use this time to discuss any questions about assessment/material.