Computerized system, method and processor-executable code to autonomously screen mailable items to identify candidate items for content inspection or testing

A non-transitory computer readable medium encoded with processor-executable code that, when executed by at least one processor, causes the at least one processor to autonomously screen in real-time each respective item of a stream of mailable or shippable items being conveyed through an item processing system, the executable code comprising:

• a routine to acquire data indicative of physical characteristics and digital information in connection with each respective item of the stream of mailable or shippable items being conveyed through the item processing system,
• a routine to process the data indicative of the physical characteristics and the digital information in connection with each respective item to form a respective virtual model of each respective item;
• a routine to generate a unique feature set for each respective item based on the respective virtual model of each respective item;
• a routine to comparatively analyze the unique feature set of each respective item with respect to reference feature sets;
• a routine to estimate a match or near match of at least some of elements of the unique feature set of a respective item of the stream of mailable or shippable items with corresponding elements of at least one of the reference feature sets;
• in response to the match or near match, a routine to identify the respective item of the stream of mailable or shippable items as a candidate for content inspection and/or testing;
• and
• a routine to command the item processing system to divert the respective identified item of the stream of mailable or shippable items for content inspection and/or testing.

A computerized method comprising:

• acquiring data indicative of physical characteristics and digital information in connection with each respective item of a stream of mailable or shippable items being conveyed through an item processing system,
• processing, by an artificial intelligence-based system, the data indicative of the physical characteristics and the digital information in connection with each respective item to form a respective virtual model of each respective item;
• generating, by the artificial intelligence-based system, a unique feature set for each respective item based on the respective virtual model of each respective item;
• comparatively analyzing, by the artificial intelligence-based system, the unique feature set of each respective item with respect to reference feature sets;
• in response to estimating a match or near match by the artificial intelligence-based system of the unique feature set of a respective item of the stream of mailable or shippable items with at least one of the reference feature sets, identifying the respective item of the stream of mailable or shippable items as a candidate for content inspection and/or testing;
• commanding the item processing system to divert the respective identified item of the stream of mailable or shippable items for content inspection and/or testing;
• and
• generating, by a machine-learning subsystem, the reference feature sets and respective match confidence thresholds for each reference feature set, wherein the generating by the machine-learning subsystem of at least some of the reference feature sets and respective match confidence thresholds of the at least some of the reference feature sets being in response to a training dataset configured to define constituent elements and respective relevance values of the defined constituent elements of the at least some of the reference feature sets, the defined constituent elements and the respective relevance values of each reference feature set being configured to predict items, which are candidates for content inspection and/or testing.