Jetson is Great for Computer Vision.
For On-Edge Computer Vision Projects, the Jetson Nano provides the following benefits:
Compatibly with TensorFlow, PyTorch etc.
But How Do I Get Started with Jetson for AI-CV Applications?
Jetson Computer Vision Course
Be at the Cutting-Edge of Computer Vision.
- Setting up any Jetson for AI.
- DeepStream SDK
- 6 x Real-World Edge-CV Jetson Applications.
OS Image + Code
8 Hours+ Content
"If you want to develop your own Low-Powered & Portable Machine Vision models on-Edge, then this course is for you!"
Introduction to Jetson.
- Overview of Course
- Jetson Nano Features
- Applications of Jetson
- Jetson Variants Comparison
Getting Started with Jetson.
- How Jetson is better than RPi.
- Which SD-Card to buy.
- How to download image & flash SD-Card
- Running Image on Jetson for the first time.
Installing Libraries & Setting up AI Computer
- Installing python and other supportive libraries.
- Explanation on various libraries and their usage e.g opencv, pytorch etc.
- Installing OpenCV from scratch with CUDA support.
- Installing other supportive libraries.
- Installing Pytorch and torchvision.
Computer Vision Basics on Jetson + PyTorch
- How to perform some basic image operations using OpenCV
- PyTorch and torchvision Basics & How to use it.
- Combining OpenCV & Torch to perform basic image operations.
- Object detection introduction and how it is performed.
- Brief about different YOLO versions.
- Demo of object detection using YOLOX
YOLOX object detection on Custom Dataset
- Number Plate Dataset & annotation for object detection.
- Train the model on some existing dataset.
- Perform object detection using pre-trained model for ANPR.
TensorRT Intro & Setup
- What is TensorRT (TRT)?
- TensorRT benefits.
- Installing dependencies for TRT.
- Setting up the environment for TRT.
Optimising YOLOX Model using TensorRT.
- Converting the YOLOX model to TensorRT
- Testing TensorRT model
- Comparing Results
What is DeepStream?
- Introduction to DeepStream.
- How it works?
- DeepStream Demo
- DeepStream Applications
- Setting up an environment for DeepStream SDK
- Testing the DeepStream SDK on Jetson.
Running DeepStream SDK with Multiple IP cameras
- Setting IP cameras with DeepStream.
- Performing multiple camera synchronisation using DeepStream SDK.
- Perform object detection on multiple cameras simultaneously.
App 1 - Vehicle Counting + Tracking
- Vehicle Detection
- Object Tracking for Vehicles
- Counting Vehicles
App 2 - Automatic Number Plate Recognition [ANPR]
- Introduction to Training with the Jetson.
- How to annotate data in YOLO format?
- Brief about Google colab and setting environment for training data.
- Training the model on custom dataset of number plates
- Extract number plate using YOLO model
- Apply Paddle-OCR to extract alphabets & digits on license plates.
App 3 - Pose Estimation on Jetson
- How pose estimation is performed?
- Installing PoseNet library.
- Implementing PoseNet estimation on Jetson.
App 4 - Fake Face Classification
- What is DeepFake?
- Perform training on videos of DeepFake.
- Fake/real video classification on pre-trained model.
App 5 - Face Recognition and Attendance - Clock in, clock out.
- Introduction to Face Recognition.
- Implementing face recognition.
- Face recognition for Attendance System.
Once you complete the course, you will graduate with an Official Certificate embedded with a unique ID that you can share on LinkedIn.
99k Subscribers - YouTube
70k Students - Augmented Startups
M(Eng.) Electronic - University of Joburg
Course Development Timeline
The Jetson Course Development commences. During this time lectures will be created and rolled out to you as they are completed.
Mar. - May 2022.
The beta version of the course will be completed. Feedback on lessons are collected and updates are rolled out based on your feedback.
31st June. 2022
Course is finalized with updates to the course.
Special Ends in: