Computer vision uses machine learning techniques to recognize objects in large amounts of images. A key development was the use of deep neural networks, which can recognize new images not in the training data as well as or better than humans. Graphics processing units (GPUs) enabled this breakthrough due to their ability to accelerate deep learning algorithms. Computer vision tasks involve both unsupervised learning, such as clustering visually similar images, and supervised learning, where algorithms are trained on labeled image data to learn visual classifications and recognize objects.