Posts by Jeff Prosise

My previous post introduced two popular algorithms for detecting faces in photographs: Viola-Jones, which relies on machine learning, and MTCNNs, which rely on deep learning. Face detection is a special case of object detection, in which computers detect and identify objects in images. Identifying an object is an image-classification problem, something at which CNNs excel.…

My previous post demonstrated how to use transfer learning with a CNN trained on millions of facial images to build a facial-recognition model that is remarkably adept at identifying faces. The images used to train the model were carefully cropped so that faces were consistently sized and positioned in the frame. Suppose you wanted to…

Not long ago, I boarded a flight to Europe and was surprised that I didn’t have to show my passport. I passed in front of a camera and was promptly welcomed aboard the flight. It was part of an early pilot for Delta Air Lines’ effort to push forward with facial recognition and offer a…

You are the leader of a group of climate scientists concerned about the planet’s dwindling rainforests. The world loses up to 10 million acres of old-growth rainforests each year, much of it due to illegal logging. Your team plans to convert thousands of discarded smart phones into solar-powered listening devices and position them throughout the…

My previous post demonstrated how to use transfer learning to build a model that with just 300 training images can classify photos of three different types of Arctic wildlife with 95% accuracy. One of the benefits of transfer learning is that it can do a lot with relatively few images. This feature, however, can also…

My post introducing convolutional neural networks (CNNs) used a dataset with photos of Arctic foxes, polar bears, and walruses to train a CNN to recognize Artic wildlife. Trained with 300 images – 100 for each of the three classes – the CNN achieved an accuracy around 60%. That’s not sufficient for most purposes. Imagine you’re…

Given a set of images with a relatively high degree of separation between classes, it is perfectly feasible to train a CNN to classify those images on a typical laptop or PC. A great example is the famous MNIST dataset, which contains 60,000 training images of scanned, handwritten digits, each measuring 28 x 28 pixels,…

Computer vision is a branch of AI in which computers extract perceptual information from images. Real-world uses include identifying objects in photos, removing inappropriate images from social-media sites, flagging defective parts coming off an assembly line, and facial recognition. Computer-vision models can even be combined with natural language processing (NLP) models to caption photos. I…

It’s not difficult to use Scikit-learn to build machine-learning models that analyze text for sentiment, identify spam e-mails, and classify textual data in other ways. But state-of-the-art text classification is most often performed with neural networks. You already know how to build neural networks that do binary classification and multiclass classification using rows of numbers…