System and method for identifying a character-of-interest

A character-recognition system comprising;

• a vector-generating module configured to obtain image data relating to an acquired image of an object, the object having a character-of-interest on a surface thereof, the vector-generating module generating a feature vector based on the character-of-interest, the feature vector having a vector address;
• and
• an output-generating module configured to generate an output vector using a sparse distributed memory (SDM) module, the SDM module including an address space having hard locations, the hard locations having a stored vector location address and stored content counters, wherein the output-generating module is configured to:
• identify activated locations from the hard locations by determining relative distances between the vector address and the stored vector location addresses, the stored content counters of the activated locations including first and second stored sub-sets of counters;
• combine the counters of the first stored sub-sets of the activated locations using a first summation thread to provide a first combined sub-set of values;
• combine the counters of the second stored sub-sets of the activated locations using a second summation thread to provide a second combined sub-set of values, wherein the first and second summation threads are run in parallel;
• and
• generate an output vector that includes a set of coordinates comprising first and second output sub-sets, the first and second output sub-sets being based on the first and second combined sub-sets, respectively.

A method of generating an output vector to identify a character-of-interest using a processor and a sparse distributed memory (SDM) module, the SDM module including an address space having hard locations, the hard locations including a stored vector location address and stored content counters, the method comprising:

• obtaining a feature vector having a vector address, the feature vector being based on a character-of-interest in an acquired image;
• identifying activated locations from the hard locations by determining relative distances between the vector address and the stored vector location addresses, the stored content counters of the activated locations including first and second stored sub-sets of counters;
• combining the counters of the first stored sub-sets of the activated locations using a first summation thread to provide a first combined sub-set of values;
• combining the counters of the second stored sub-sets of the activated locations using a second summation thread to provide a second combined sub-set of values, wherein the first and second summation threads are run in parallel;
• and
• generating an output vector that includes a set of coordinates comprising first and second output sub-sets, the first and second output sub-sets being based on the first and second combined sub-sets, respectively.