Back to results
US 12,162,041 B2Granted
Systems and methods for waste item detection and recognition
OwnerHassan Murad, Vivek Hiteshbhai Vyasinventors
Life of this patent
Estimating the term…
Abstract
Embodiments described herein relate to hardware and software for waste item detection and recognition, along with an education or feedback system. Embodiments described herein use artificial intelligence, which embodies machine learning and computer vision, to detect waste items and generate feedback to nudge the user to dispose the waste items into appropriate receptacles while generating smart operational insights of a designated premise.
Claims27
A system for waste item detection comprising:
- sensors for detecting an approaching object to trigger a camera to capture image data and one or more processors to process the image data;
- the camera for capturing image data of the approaching object;
- non-transitory computer readable storage medium with executable instructions for causing the one or more processors to:
- process the image data using a neural network to detect a waste item within the image data and determine a category for the waste item, wherein the neural network defines classes for different waste items and maps the detected waste item to the classes for the different waste items to compute a pairing of the detected waste item and a class, the class being associated with the category for the waste item;
- and
- generate feedback data indicating the category for the waste item;
- wherein the camera or the sensors capture additional data indicating disposal of the waste item, wherein the one or more processors use the additional data to determine whether the waste item was disposed in the appropriate receptacle, and generates additional feedback data based on the determination.
4 more elements
A system for waste item detection comprising:
- sensors for detecting an approaching object to trigger image processing for detection and recognition;
- a camera for capturing image data of the approaching object;
- non-transitory computer readable storage medium with executable instructions for causing one or more processors to:
- process the image data using a neural network to detect a waste item within the image data and determine a category for the waste item;
- determine an appropriate receptacle to dispose of the detected waste item;
- and
- generate feedback data indicating the category for the waste item;
- wherein the camera or the sensors capture additional data indicating disposal of the waste item, wherein the one or more processors use the additional data to determine whether the waste item was disposed in the appropriate receptacle, and generates additional feedback data based on the determination;
- wherein the one or more processors are configured to use the neural network or an additional neural network to detect brand data within the image data and tag the image data with metadata indicating the brand data.
6 more elements
Dependencies are inferred from the claim text.