Object tracking using object attributes

A computer-implemented method, comprising:

• receiving, by a computing system, data specifying visual attributes of a particular object detected in an image;
• identifying, by the computing system, a tracked object profile for a tracked object, the tracked object profile comprising data specifying visual attributes of the tracked object;
• determining a level of similarity between the particular object detected in the image and the tracked object based on comparison of visual attributes of the particular object with corresponding visual attributes of the tracked object, including normalizing similarities for the compared visual attributes, wherein a first normalization function used to normalize a level of similarity between first corresponding visual attributes of the particular object and the tracked object is different from a second normalization function used to normalize a level of similarity between second corresponding visual attributes of the particular object and the tracked object;
• determining, based on the level of similarity between the particular object detected in the image and the tracked object, that the particular object matches the tracked object;
• providing an indication that the particular object matches the tracked object;
• and
• updating the tracked object profile for the tracked object in response to determining that the particular object matches the tracked object, the updating including updating the visual attributes of the tracked object based on the visual attributes of the particular object.

A computer program product, encoded on one or more non-transitory computer storage media, comprising instructions that when executed by one or more computers cause the one or more computers to perform operations comprising:

• receiving data specifying visual attributes of a particular object detected in an image;
• identifying a tracked object profile for a tracked object, the tracked object profile comprising data specifying visual attributes of the tracked object;
• determining a level of similarity between the particular object detected in the image and the tracked object based on comparison of visual attributes of the particular object with corresponding visual attributes of the tracked object, including normalizing similarities for the compared visual attributes, wherein a first normalization function used to normalize a level of similarity between first corresponding visual attributes of the particular object and the tracked object is different from a second normalization function used to normalize a level of similarity between second corresponding visual attributes of the particular object and the tracked object;
• determining, based on the level of similarity between the particular object detected in the image and the tracked object, that the particular object matches the tracked object;
• providing an indication that the particular object matches the tracked object;
• and
• updating the tracked object profile for the tracked object in response to determining that the particular object matches the tracked object, the updating including updating the visual attributes of the tracked object based on the visual attributes of the particular object.