Researchers have developed a technique that could allow deep learning algorithms to learn the visual features of images in a self-supervised fashion, without the need for annotations by human researchers. Tech Explore brought this information to us in their article, “A new machine learning strategy that could enhance computer vision.”
The Universitat Autonoma de Barcelona, Carnegie Mellon University and the International Institute of Information Technology in Hyderabad, India aim to give computers the capability to read and understand textual information in any type of image in the real-world. In their study, they designed computational models that join textual information about images with the visual information contained within them, using data from Wikipedia and other online platforms. They then used these models to train deep-learning algorithms on how to select good visual features that semantically describe images. The model does not require specific annotations for each image. Instead, the textual context where the image is found acts as the supervisory signal.
By training their algorithms using images from the internet, the researchers highlighted the value of content that is readily available online.
Melody K. Smith
Sponsored by Data Harmony, a unit of Access Innovations, the world leader in indexing and making content findable.