Seeing AI

Seeing AI is a free application from Microsoft, designed primarily for the blind and low-vision community. Developed by Saqib Shaikh—a Microsoft engineer who is himself visually impaired—the app leverages computer vision and artificial intelligence to narrate the world in real-time. The app functions by turning the smartphone camera into an intelligent sensor that can interpret a variety of visual cues. It is divided into "channels" to handle different tasks. The Short Text channel speaks text as soon as it appears in front of the camera, making it ideal for reading mail or signs. The Document channel provides audio guidance to capture a printed page and recognizes the text with its original formatting. The Product channel scans barcodes, using audio beeps to guide the user, and then identifies the item and its details. Furthermore, the app features a People channel that can recognize friends and describe their facial expressions, including an estimate of their age and mood. The Scene channel provides an overall description of the captured environment. More specialized channels include Currency (identifying banknotes), Color (identifying the color of clothes or objects), and Light (generating an audible tone based on the intensity of light in the room). Recent updates have integrated generative AI to allow users to ask follow-up questions about scanned documents, such as "What is the price of the burger on this menu?" or "Summarize this article."