A vision task that involves both finding where objects are (localization with bounding boxes) and identifying what they are (classification).
Beyond simple image classification.
Used in self-driving cars and security systems.