Proceeding to block 408, the images also may be associated with each other by the physical proximity of the spaces which they respectively represent. For example, images may be determined to be of proximate spaces based on a location of one image space being within a first distance of the location of another image space. The locations may be determined to be proximate based on the images sharing a similarity such as an image of another object. At block 410 the images may be identified and/or categorized by type of objects in the images using image recognition. The images may be stored locally on the smart glasses and/or uploaded to a web storage server.

The image recognition and classification above may be executed using a machine learning algorithm such as a neural network (NN). An input training set may be used to train a NN using pre-classified object images.

FIG. 5 illustrates subsequent logic. Commencing at block 500, suppose the wearer of the smart glasses has lost an object. He may speak the name of the object, which may be detected by a microphone on the smart glasses and processed using speech recognition. If desired, at block 502 indicates that input of approximate time of loss and/or location of loss may be received. Even if such information is not input, the logic may move to block 504 to access images with objects matching the input at block 500. Time/location, if input at block 502, may be used to winnow the images returned at block 506 by presenting the images on a display of the smart glasses.