Object detection
Object detection uses computer vision to spot items such as phones, books or extra screens in the camera view during an exam.
Definition
Object detection is a computer vision technique that locates and labels items in an image. In proctoring it is used to spot objects that are not allowed in the exam, such as mobile phones, books, notes, headphones or additional screens, in the webcam or second camera view.
Detection quality depends on the camera angle, lighting and resolution. A front-facing webcam sees mostly the candidate's face, so objects on the desk are often out of view; a second camera placed to the side shows the desk and hands and gives object detection much more to work with.
Like other AI signals, object detection produces flags rather than verdicts. A phone detected on the desk may be switched off, and a book may be an allowed reference in an open-book exam. QProctor's AI detects phones, books and second screens, including on the second camera's view of the desk, and records the frame for review.