Invariant-based dimensional reduction of object recognition features, systems and methods

摘要

A sensor data processing system and method is described. Contemplated systems and methods derive a first recognition trait of an object from a first data set that represents the object in a first environmental state. A second recognition trait of the object is then derived from a second data set that represents the object in a second environmental state. The sensor data processing systems and methods then identifies a mapping of elements of the first and second recognition traits in a new representation space. The mapping of elements satisfies a variance criterion for corresponding elements, which allows the mapping to be used for object recognition. The sensor data processing systems and methods described herein provide new object recognition techniques that are computationally efficient and can be performed in real-time by the mobile phone technology that is currently available.

主权项

1. A sensor data processing system comprising:
a controlled environment having environmental parameters having corresponding environmental attributes; and an image processing engine programmed to:
obtain a first training data set representative of at least one object under a defined environmental state within the controlled environment;derive a first recognition trait from the first training data set according to a trait extraction algorithm, the first recognition trait comprising a first plurality of elements, wherein deriving includes generating a trait vocabulary as a function of descriptor clusters derived from at least the first training data set;configure the controlled environment to create a second environmental state by adjusting an environmental attribute through adjusting a corresponding environment parameter;obtain a second training data set representative of the at least one object under the second environmental state within the controlled environment with respect to the adjusted environmental attribute;derive a second recognition trait from the second training data set according to the trait extraction algorithm, the second recognition trait comprising a second plurality of elements, and wherein the second recognition trait has a correspondence to the first recognition trait;identify a mapping that maps a plurality of elements of the first recognition trait and the second recognition trait in a new representation space, wherein the mapping of the plurality of elements in the new representation space satisfies trait element variance criteria among corresponding elements in traits across the first training data set and the second training data set; andstore the mapping in a memory.