Machine Learning for Computer Vision: Foundations and Use Cases

A catalogue of Heartbeat posts exploring computer vision, including important background, use cases, frameworks, and more

With advances in camera quality, image fidelity, and neural network research focused on solving image- and video-based challenges, computer vision continues to capture the attention and imaginations of machine learning researchers and practitioners.

But computer vision is an incredibly broad umbrella term that encompasses an array of specific tasks and challenges, and the field continues to expand.

With that in mind, we’ve selected some of our favorite Heartbeat posts that cover a range of topics within computer vision and separated them into two categories:

  • Computer Vision Foundations
  • Use Cases

Computer Vision Foundations

The Ancient Secrets of Computer Vision (condensed, by Joseph Redmon)

Condensed summaries and discussion of Joseph Redmon’s sprawling examination of all things computer vision.

— by Angus Addlesee

The 5 Computer Vision Techniques That Will Change How You See The World

In-depth overviews of common computer vision techniques: Image classification, object detection, object tracking, semantic segmentation, and instance segmentation.

— by James Le

Explore this detailed look at the trends in computer vision that have defined the last couple of years and will continue to make waves moving forward.

— by James Le

Artificial Art: How GANs are making machines creative

Exploring how advances in neural networks have created a new wave of creative machines capable of producing art, literature, and music — not exclusively CV, but an interesting look at generative computer vision within.

— by Machine’s Creativity

Convolutional Neural Networks: An Intro Tutorial

An overview of convolutional neural networks—the foundational building block of deep learning-based computer vision—including how and why they work and an implementation with Keras and TensorFlow.

— by Derrick Mwiti

Use Cases: Image Classification

Basics of Image Classification with PyTorch

Learn how to build a complete image classification pipeline with PyTorch — from scratch.

— by John Olafenwa

Scannable Chess Scoresheets with a Convolutional Neural Network and OpenCV

In this tutorial, you’ll learn how to build a ready-to-ship machine learning model that can automatically scan chess scoresheets.

— by Rithwik Sudharsan

Training an Image Classification Convolutional Neural Net to Detect Plant Disease Using fast.ai

Learn how to combine the powers of image classification and fast.ai’s training library to detect plant disease with a neural network.

— by Steve Mutuvi

Building a Vision-Controlled Car Using Raspberry Pi — From Scratch

Build a car robot from scratch and control it using Raspberry Pi based on images captured and by a USB camera and then classified.

— by Ahmed Gad

Using TensorFlow.js to Train a “Rock-Paper-Scissors” Model

Train an image classification machine learning model in your browser in ~10 minutes.

— by Gant Laborde

Classification with TensorFlow and Dense Neural Networks

Learn how to perform classification using TensorFlow and its dense neural network classifier module.

Mohammed Rampurawala

Binary Classification using Keras in R

RStudio comes equipped with an interface for using Keras to build machine learning models in R. Learn how to use this interface to create a binary classification model.

— by Derrick Mwiti

Use Cases: Object Detection

Detecting objects in videos and camera feeds using Keras, OpenCV, and ImageAI

Learn how to detect objects in single video frames from camera feeds with Keras, OpenCV, and ImageAI.

— by Moses Olafenwa

Real-Time Object Detection on Raspberry Pi Using OpenCV DNN

Learn how to use OpenCV’s Deep Neural Network module (DNN) to detect objects in real time on a Raspberry Pi.

— by Saumya Shovan Roy (Deep)

Gentle guide on how YOLO Object Localization works with Keras

Learn how to build and implement a YOLO object localization model with Keras.

— by Chengwei Zhang

Real-Time Person Tracking on the Edge with a Raspberry Pi

A look at model architectures and optimizations for real-time detection and tracking of people on the edge.

— by Priya Dwivedi

Object detection in just 3 lines of R code using Tiny YOLO

Using R and a Tiny YOLO model, learn how you can detect objects in just 3 lines of code.

— by AbdulMajedRaja RS

Use Cases: Generative Adversarial Networks (GANs)

Introduction to Generative Adversarial Networks (GANs): Types, and Applications, and Implementation

This tutorial will introduce Generative Adversarial Networks (GANs), explore the different variations, their applications, and help you learn to build your own simple GAN using Keras.

— by Derrick Mwiti

Animated StyleGAN image transitions with RunwayML

Create an endless loop of synthetically-generated StyleGAN landscapes with smooth transitions using RunwayML and P5.js.

— by Mike Heavers

My MangaGAN: Building My First Generative Adversarial Network

Build a generative adversarial network to make your very own anime characters with Keras.

— by Nikita Sharma

Use Cases: Style Transfer

Neural Style Transfer with PyTorch

Neural style transfer is an exciting technology that generates images in the style of another image. Learn how this works, along with a simple implementation in PyTorch.

— by Derrick Mwiti

Art & Soul— A Style Transfer Website based on Tkinter and Django

Parts one and two of a complete walkthrough for building your own style transfer website using VGG16 and Django.

— by Nikita Sharma

Part 1: Building a style transfer model:

Part 2: Implementing the model in a website built with Tkinter and Django:

Use Cases: Pose Estimation

“Just Point It”: Machine Learning on iOS with Pose Estimation + OCR Using Core ML and ML Kit

A look a “Just Point It”, an iOS app that leverages machine learning to allow users to point at text on a paper document and look up the word that’s being pointed at.

— by Doyoung Gwak

One Last Cool Computer Vision Thing

Real-Time 2D/3D Feature Point Extraction from a Mobile Camera

Explore the intersection of machine learning and augmented reality, and learn how to extract 2D and 3D feature points from a smartphone camera.

— by Hart Woolery

Avatar photo

Fritz

Our team has been at the forefront of Artificial Intelligence and Machine Learning research for more than 15 years and we're using our collective intelligence to help others learn, understand and grow using these new technologies in ethical and sustainable ways.

Comments 0 Responses

Leave a Reply

Your email address will not be published. Required fields are marked *