Methods and systems for machine learning-based article checking process

A non-transitory computer-readable memory comprising computer-executable instructions stored thereon that, when executed on a processor, cause the processor to perform steps comprising:

• receiving, from a handheld device comprising a camera, at least one image of a received consignment comprising one or more articles;
• obtaining an article identifier for the received consignment based on the at least one image;
• determining, with a first machine-learned model processing the at least one image, a class of each of the one or more articles forming a set of determined classes and a quantity for each class in the set of determined classes;
• making a first determination whether the article identifier matches the set of determined classes and the quantity of each class in the set of determined classes;
• generating, from the at least one image having been processed with the first machine-learned model, an isolated article image for each of the one or more articles;
• determining, with a second machine-learned model processing a first isolated image corresponding to a first article comprised by the one or more articles, whether the first article is damaged based on the determined class of the first article;
• and
• generating an alert based on a determination that the first article is damaged.

A system, comprising:

• a warehouse that receives a consignment comprising one or more articles;
• a handheld device comprising a camera;
• and
• a computer comprising one or more computer processors, the computer configured to:
• receive, from the handheld device comprising, at least one image of the received consignment;
• obtain an article identifier for the received consignment based on the at least one image;
• determine, with a first machine-learned model processing the at least one image, a class of each of the one or more articles forming a set of determined classes and a quantity for each class in the set of determined classes;
• make a first determination whether the article identifier matches the set of determined classes and the quantity of each class in the set of determined classes;
• generate, from the at least one image having been processed with the first machine-learned model, an isolated article image for each of the one or more articles;
• determine, with a second machine-learned model processing a first isolated image corresponding to a first article comprised by the one or more articles, whether the first article is damaged based on the determined class of the first article;
• and
• generate an alert based on a determination that the first article is damaged.