Blog

When used in the right situation, neural networks can be an awesome solution to your learning problem. Neural networks allow you to feed in structured data (numerical and categorical data), wait for some magic to happen (note: not actual magic — it’s just maths), and out pops your answer — for example, maybe you’re trying to predict the result of an election from data from news sources. Convolutional neural networks allow you to do something similar but for images — for example, maybe you’re trying to predict whether an image is a hotdog or a not-hotdog.

Continue reading Building a mixed-data neural network in Keras to predict accident locations

Instagram is one of the leading social media platforms in the world with over a billion monthly users. It’s great for sharing pictures, connecting with friends and family, getting updates about our favorite celebrities, and posting stories using AR camera filters.

Continue reading Creating Instagram and Facebook AR filters with Spark AR Studio — A beginner’s guide (Part 1)

Google introduced a new product in the Firebase Suite earlier this year, Firebase’s Machine Learning Kit. ML Kit brings Google’s machine learning expertise to both Android and iOS apps in a powerful way. In this post I will dive into how we can make use of it in order to build a real-time face detector for an Android app.

Continue reading “Real-Time Face Detection on Android with ML Kit”

Welcome back to part 5 of our series exploring how to build a side-scroller mobile game with Unreal Engine 4. If you haven’t yet read part 4, I’d recommend reading it here:

In the last part, we integrated our winning and losing mechanisms. Here, we’ll work on creating a boss enemy to make our game a bit more challenging.

Continue reading Creating a side-scroller mobile game with Unreal Engine 4 (Part 5): Creating a Boss Enemy

TensorFlow.js is a very powerful library when it comes to using deep learning models directly in the browser. It includes support for a wide range of functions, covering basic machine learning, deep learning, and even model deployment.

Another important feature of Tensorflow.js is the ability to use existing pre-trained models for quickly building exciting and cool applications.

Continue reading “Constructing a 3D Face Mesh from Face Landmarks in Real-Time with TensorFlow.js and Plot.js”

When building an app, sometimes we find ourselves creating data just so we can see our UI start to come to life. Whether we’re adding a new feature, making a POC (proof of concept), or starting from scratch, it’s very easy to just put data together for the sake of it.

For instance, we’ve been building a Team Profiles app to learn SwiftUI. And, so far, we’ve managed to make a class for our Profiles, but even with just three instances, our UI code looks bulky. It takes up more space in our code than the actual UI code does in the ContentView!

Continue reading SwiftUI: Data Driven UI

When working with standard photo editors, users will generally encounter a common suite of tools — scaling, cropping, color, rotation, blurring, filtering, etc.

But what do users do when they want something a little more powerful and complex? Until recently, the answer has most likely been to transfer photos to powerful desktop editing platforms.

But heavyweight editing tools can be both difficult to learn and cost-prohibitive. And with recent advances in smartphone cameras, more and more high-quality photography is happening on mobile devices. Moving back and forth between a phone’s camera roll and desktop software can quickly become cumbersome, especially when working with multiple images.

Continue reading Superimpose X: Bringing professional photo editing tools to smartphones with mobile machine learning

Unnecessary background within an image, photograph, or video makes the image lose its focal point. By removing the background, we can draw the attention of the viewer back to the focal point, thereby making the image more informative and easy to view.

In this article, we will learn to create segmentation and the camera texture to create an effect that replaces unnecessary background in Facebook’s Spark AR Studio. We will make use of a segmentation mask texture to separate the foreground and background of the scene. Then, we will implement the camera texture to a material in the foreground to stream a video of the person using the effect back into the scene. You can then transform the environment behind them and replace the background.

Continue reading Replace the Background in Facebook’s Spark AR Studio

Often in machine learning, and specifically with classification problems, we encounter imbalanced datasets. This typically refers to an issue where the classes are not represented equally, which can cause huge problems for some algorithms.

In this article, we’ll explore a technique called resampling, which is used to reduce this effect on our machine learning algorithms.

Continue reading Resampling to Properly Handle Imbalanced Datasets in Machine Learning