Videos
Search Videos
-
Data Curation and Feature Scaling
Run Time: 19 Minutes
In this IoT Central MicroSession with Edge Impulse, learn how to perform feature normalization and standardization prior to uploading data to Edge Impulse. This process involves calculating the mean and standard deviation of each channel and standardizing the data along a single axis to have a mean of 0 and a standard deviation of 1. The means and standard deviations will need to be recorded for inference during deployment. We use Google Colab to perform this analysis and transformation prior to uploading the data to Edge Impulse.
Instructor: Shawn Hymel Senior Developer Relations Engineer, Edge Impulse
Optional Hardware: None
Preparation: Free sign-up at https://studio.edgeimpulse.com/signupData Collection with a Custom Sensor
Run Time: 25 minutes
In this IoT Central MicroSession with Edge Impulse, learn how to collect data from custom sensors and store the readings in CSV files for analysis and uploading to Edge Impulse. For seamless efficiency, the Edge Impulse Data Forwarder tool can be used to automatically send data to the Edge Impulse Studio using an internet connection with your embedded system tethered to your computer via USB. By storing data to CSV files, you can then collect samples in the field without an internet connection.
Instructor: Shawn Hymel, Senior Developer Relations Engineer, Edge Impulse
Required Hardware: Arduino Nano 33 BLE Sense
Preparation: Free sign-up at https://studio.edgeimpulse.com/signupCreate Real-Time Object Detection on Low-Power Microcontrollers
Run Time: 19 Minutes
In this IoT Central MicroSession with Edge Impulse, learn how to create a real-time object detection system (FOMO) using low-power microcontrollers. You will learn how to collect a high quality object detection dataset to train and deploy a FOMO model to a microcontroller like the Himax WE-I Plus.
Instructor: Jenny Plunkett, Senior Developer Relations Engineer, Edge Impulse
Preparation: Free sign-up at https://studio.edgeimpulse.com/signupClassify Non-Human Voice Audio for Automation and Embedded Machine Learning
Run Time: 23 Minutes
In this IoT Central MicroSession, learn how to create a non-human voice audio classification system to identify when a faucet is leaking. This demo can also be adapted for industrial use cases, such as identifying machine failures with grinding sounds, or even for wildlife conservation purposes, to recognize the vocalizations of various endangered species in the wild.
Instructor: Jenny Plunkett, Senior Developer Relations Engineer, Edge Impulse
Optional Hardware: Mobile Phone
Preparation: Free sign-up at https://studio.edgeimpulse.com/signupAuto ML with EON Tuner
Run Time: 18 Minutes
In this IoT Central MicroSession with Edge Impulse, learn how to speed-up your machine learning pipeline design using the EON Tuner, Edge Impulse's tool for AutoML and much more. The EON Tuner analyzes your input data, potential signal processing blocks, and neural network architectures, providing an overview of possible model architectures that will fit your chosen device's latency and memory requirements. Louis will walk you through this process and show how to leverage the tool in just 18 minutes.
Instructor: Louis Moreau, Senior Developer Relations Engineer, Edge Impulse
Required Hardware: none
Preparation: Free sign-up at https://studio.edgeimpulse.com/signupAdvanced Sensor Anomaly Detection
In this IoT Central MicroSession with Edge Impulse, learn how you can easily create a machine learning classifier to detect anomalous sensor readings using only nominal training data for embedded machine learning models directly in the Edge Impulse Studio.
Instructor: Jenny Plunkett, Senior Developer Relations Engineer, Edge Impulse
Optional Hardware: mobile phone or a microcontroller with accelerometer sensor
Preparation: Free sign-up at https://studio.edgeimpulse.com/signupHow to Build Continuous Audio Classification Applications
Run Time: 16 Minutes
In this IoT Central MicroSession with Edge Impulse, quickly learn how to build continuous audio classification applications using Edge Impulse. First see a walkthrough of the theoretical understanding of the continuous classification methods before diving into the demo. Then follow a demo of the data collection and exploration, impulse design, and finally perform live inference on an Arduino Nano 33 BLE Sense, smartphone or anyone one of the boards listed below.
Instructor: Clinton Odour, Senior Developer Relations Engineer, Edge Impulse
Optional Hardware: Any smartphone and most boards with a microphone or an accelerometer. Examples include: ST B-L475E-IOT01A, Arduino Nano 33 BLE Sense, ESP32 FireBeetle with LIS3DHTR accelerometer, Himax WE-I Plus, Nordic Semiconductor nRF52840 DK, nRF5340 DK or nRF9160 DK, Silicon Labs Thunderboard Sense 2, Sony's Spresense, TI CC1352P LaunchPad, Raspberry Pi 4, Nvidia Jetson and more.
Preparation: Free sign-up at https://studio.edgeimpulse.com/signupHow to Classify Motions Using Embedded ML & Accelerometers
Run Time: 15 Minutes
In this IoT Central On-Demand MicroSession, learn how to classify three-axis motion signatures with machine learning using Edge Impulse. You will navigate through the full MLOps pipeline from the data collection to the live inferencing passing by the digital signal processing features extraction and the neural network training.
Instructor: Louis Moreau, Senior DevRel Engineer, Edge Impulse
Optional Hardware: Sensor or smartphone with 3-axis accelerometer
Preparation: Free sign-up at https://studio.edgeimpulse.com/signupHow to Create a Keyword Spotting System in 20 minutes
Run Time: 27 Minutes
In this IoT Central MicroSession, learn how to create a keyword spotting system to recognize spoken words. Follow along or use an Arduino board (details below) and classify words using Edge Impulse.
Instructor: Shawn Hymel, Senior DevRel Engineer, Edge Impulse
Optional Hardware: Arduino Nano 33 BLE Sense
Preperation: Free sign-up at https://studio.edgeimpulse.com/signupInference with Custom Sensors
Run Time: 20 Minutes
In this IoT Central MicroSession, learn how to create a keyword spotting system to recognize spoken words. Follow along or use an Arduino board (details below) and classify words using Edge Impulse.
Instructor: Shawn Hymel, Senior DevRel Engineer, Edge Impulse
Optional Hardware: Arduino Nano 33 BLE Sense
Preparation: Free sign-up at https://studio.edgeimpulse.com/signupLearn How to Integrate an Object Detection Model in a Custom Application
Run Time: 12 Minutes
In this IoT Central MicroSession with Edge Impulse, learn how to integrate object detection models with the Linux Python SDK to build fully-customizable applications. In this example, you will see how an object detection model running locally can output a live video stream with the bounding boxes in a web application.
Instructor: Louis Moreau, Senior Developer Relations Engineer, Edge Impulse
Optional Hardware: Linux x86 or MacOS
Preparation: Free sign-up at https://studio.edgeimpulse.com/signupLearn to Build Object Detection Using Transfer Learning
Run Time: 15 Minutes
In this IoT Central MicroSession with Edge Impulse, quickly learn how to build an object detection model using Transfer Learning. You will also learn how to run the live inference using Edge Impulse Linux CLI or MacOS-based hardware. This is the first ion a two part series (coming soon).
Instructor: Louis Moreau, Senior Developer Relations Engineer, Edge Impulse
Optional Hardware: Linux-based or MacOS-based hardware
Preparation: Free sign-up at https://studio.edgeimpulse.com/signupUsing Machine Learning for Sensor Fusion
Run Time: 22 Minutes
In this IoT Central MicroSession, learn about use cases for sensor fusion and how it can be accomplished using neural networks. You will walk through how a machine learning model performing sensor fusion can be trained in Edge Impulse and tested on a microcontroller.
Instructor: Shawn Hymel, Senior Developer Relations Engineer, Edge Impulse
Optional Hardware: Arduino Nano 33 BLE Sense
Preparation: Free sign-up at https://studio.edgeimpulse.com/signupUsing Image Data Augmentation to Train a Better ML Model
Run Time: 18 Minutes
In this IoT Central MicroSession with Edge Impulse, learn the process of performing data augmentation on an image dataset, which includes flipping, translating, zooming, rotating, and adding noise. This process generates modified copies of the original data. See how to train a convolutional neural network (CNN) on Edge Impulse with the augmented dataset and demonstrate how it is more accurate than a model trained on the original data.
Instructor: Shawn Hymel, Senior Developer Relations Engineer, Edge Impulse
Required Hardware: None
Preparation:
Free sign-up at https://studio.edgeimpulse.com/signup
Code and dataset: https://github.com/ShawnHymel/computer-vision-with-embedded-machine-learning
Additional Resources:
A non-augmented version of the project used in the video: studio.edgeimpulse.com/public/36514/latest
An example of the project with augmented data: studio.edgeimpulse.com/public/36800/latestUsing the C++ SDK Library to Perform Inference on Any Device
Run Time: 29 Minutes
In this IoT Central MicroSession with Edge Impulse, see how to use the Edge Impulse C++ SDK library to perform inference on any platform. You will learn how to build a C++ application that includes the necessary library functions to predict class probabilities from a static set of raw features, as well as construct a Makefile to link the libraries together.
Instructor: Shawn Hymel, Senior Developer Relations Engineer, Edge Impulse
Optional Hardware: Lunix computer, (i.e Raspberry Pi)
Preparation: Free sign-up at https://studio.edgeimpulse.com/signupData Collection with a Custom Sensor
Collect data from custom sensors and store the readings in CSV files for analysis and uploading to Edge Impulse.
Using Image Data Augmentation to Train a Better ML Model
Learn the process of performing data augmentation on an image dataset, which includes flipping, translating, zooming, rotating, and adding noise.
Advanced Sensor Anomaly Detection
How to easily create a machine learning classifier to detect anomalous sensor readings using only nominal training data for embedded machine learning models directly in the Edge Impulse Studio.
Learn How to Integrate an Object Detection Model in a Custom Application
Learn how to integrate object detection models with the Linux Python SDK to build fully-customizable applications.
Data Curation and Feature Scaling
Learn how to perform feature normalization and standardization prior to uploading data to Edge Impulse. This process involves calculating the mean and standard deviation of each channel and standardizing the data along a single axis to have a mean of 0 and a standard deviation of 1. The means and standard deviations will need to be recorded for inference during deployment. We use Google Colab to perform this analysis and transformation prior to uploading the data to Edge Impulse.
Instructor: Shawn Hymel Senior Developer Relations Engineer, Edge Impulse
Optional Hardware: None
Preparation: Free sign-up at https://studio.edgeimpulse.com/signupInference with Custom Sensors
Learn how to create a keyword spotting system to recognize spoken words. Follow along or use an Arduino board (details below) and classify words using Edge Impulse.
Auto ML with EON Tuner
Learn how to speed-up your machine learning pipeline design using the EON Tuner, Edge Impulse's tool for AutoML and much more.
Automate your ML Pipeline
In this IoT Central MicroSession with Edge Impulse, learn how to create complete automated data pipelines so you can work on your active learning strategies.