Device and method for hierarchical object recognition

A method comprising:

• at a device comprising:
• an augmented reality display device including a camera, a display device, and an eye-tracking device;
• an input device;
• a memory storing a hierarchical object recognition library arranged in a plurality of levels;
• and a controller, receiving, using the input device, an indication of a selected level of the hierarchical object recognition library;
• determining, using the eye-tracking device, an eye-gaze direction;
• recognizing, using the controller, at least one object in an image from the camera in the eye-gaze direction by comparing at least a region of the image in the eye-gaze direction with the selected level of the hierarchical object recognition library;
• controlling, using the controller, the display device to indicate a recognized object in the eye-gaze direction;
• and
• when an object to be recognized is partially out of a field-of-view of the camera, transmitting, using a communication interface of the device, a request to a server to acquire a respective image of the object to be recognized using one or more other cameras.

A device comprising:

• a display device;
• an input device;
• a memory storing a hierarchical object recognition library arranged in a plurality of levels;
• and
• a controller communicatively coupled to the display device, the input device, and the memory, the controller configured to:
• receive an image including a plurality of objects;
• receive, using the input device, an indication of a selected level of the hierarchical object recognition library;
• determine an eye-gaze direction;
• recognize at least one of the objects in the image by comparing at least a region of the image in the eye-gaze direction with the selected level of the hierarchical object recognition library;
• control the display device to indicate a recognized object in the eye-gaze direction;
• and
• when an object to be recognized is partially out of a field-of-view of a camera that acquired the image, transmit a request to a server to acquire a respective image of the object to be recognized using one or more other cameras.