Systems and methods for identifying different product identifiers that correspond to the same product
In some embodiments, apparatuses and methods are provided herein useful to processing captured images. In some embodiments, there is provided a system for processing captured images of objects at a product storage facility including a trained machine learning model; and a control circuit. The control circuit may group a plurality of product identifiers into one or more clusters based on at least one of visual similarity of corresponding images, textual similarity of corresponding associated descriptions, and associated relationships between product identifiers of the plurality of product identifiers; determine clusters having common elements that are at least within a similarity threshold of each other; merge the clusters with the common elements; and generate a mapping dataset used to retrain the trained machine learning model to identify a plurality of objects. The mapping dataset may include a plurality of associations of associated product identifiers to a single object.
1 . A system for processing captured images of objects at a product storage facility, the system comprising:
one or more image capture devices configured to capture images of the objects at the product storage facility;
a trained machine learning model stored in a memory, wherein the trained machine learning model is trained on the images and metadata associated with a plurality of product identifiers, a plurality of product reference images, a plurality of planogram data, or combination thereof; and
a control circuit executing the trained machine learning model configured to:
convert the captured images into a plurality tensors;
group the plurality of product identifiers into one or more clusters based on at least one of visual similarity of corresponding images, textual similarity of corresponding associated descriptions, and associated relationships between product identifiers of the plurality of product identifiers, wherein the visual similarity of the corresponding images is determined based on a calculation of hamming distances pairwise among the plurality of tensors;
determine clusters having common elements that are at least within a similarity threshold of each other;
merge the clusters with the common elements; and
generate a mapping dataset used to retrain the trained machine learning model to identify a plurality of objects, wherein the mapping dataset comprises a plurality of associations of associated product identifiers to a single object.
2 . The system of claim 1 , further comprising
a database configured to store the images.
3 . The system of claim 2 , wherein at least one of the one or more image capture devices is coupled to a motorized robotic unit.
4 . The system of claim 1 , wherein the visual similarity of the corresponding images is determined based on a calculation of hamming distances between the corresponding images and a grouping of images having corresponding hamming distances that are less than a distance threshold.
5 . The system of claim 1 , wherein the textual similarity of the corresponding associated descriptions is determined based on a match of a predefined text associated with the corresponding associated descriptions.
6 . The system of claim 1 , wherein the product storage facility comprises at least one of: a product distribution center, a fulfillment center, and a retail store.
7 . The system of claim 1 , wherein the common elements comprise one or more product identifiers.
8 . The system of claim 1 , wherein the control circuit executing the trained machine learning model is further configured to generate a second mapping dataset used to retrain the trained machine learning model to identify the plurality of objects, wherein the second mapping dataset comprises a plurality of associations of each of the associated product identifiers to at least one corresponding product storage facility.
9 . The system of claim 1 , wherein the associated relationships comprise different product identifiers that are variants of same object.
10 . The system of claim 9 , wherein the variants of the same object comprise differences in at least one of: size, color, and pattern.
11 . A method for processing captured images of objects at a product storage facility, the method comprising:
capturing, by one or more image capture devices, images of the objects at the product storage facility;
training a machine learning model on the images and metadata associated with a plurality of product identifiers, a plurality of product reference images, a plurality of planogram data, or combination thereof;
converting, by a control circuit, the captures images into a plurality of tensors;
grouping, by the control circuit executing the trained machine learning model stored in a memory, the plurality of product identifiers into one or more clusters based on at least one of visual similarity of corresponding images, textual similarity of corresponding associated descriptions, and associated relationships between product identifiers of the plurality of product identifiers, wherein the visual similarity of the corresponding images is determined based on a calculation of hamming distances pairwise among the plurality of tensors;
determining, by the control circuit executing the trained machine learning model, clusters having common elements that are at least within a similarity threshold of each other;
merging, by the control circuit executing the trained machine learning model, the clusters with the common elements; and
generating, by the control circuit executing the trained machine learning model, a mapping dataset used to retrain the trained machine learning model to identify a plurality of objects, wherein the mapping dataset comprises a plurality of associations of associated product identifiers to a single object.
12 . The method of claim 11 , further comprising
storing, by a database, the images.
13 . The method of claim 12 , wherein at least one of the one or more image capture devices is coupled to a motorized robotic unit.
14 . The method of claim 12 , wherein the visual similarity of the corresponding images is determined based on a calculation of hamming distances between the corresponding images and a grouping of images having corresponding hamming distances that are less than a distance threshold.
15 . The method of claim 11 , wherein the textual similarity of the corresponding associated descriptions is determined based on a match of a predefined text associated with the corresponding associated descriptions.
16 . The method of claim 11 , wherein the product storage facility comprises at least one of: a product distribution center, a fulfillment center, and a retail store.
17 . The method of claim 11 , wherein the common elements comprise one or more product identifiers.
18 . The method of claim 11 , further comprising generating, by the control circuit executing the trained machine learning model, a second mapping dataset used to retrain the trained machine learning model to identify the plurality of objects, wherein the second mapping dataset comprises a plurality of associations of each of the associated product identifiers to at least one corresponding product storage facility.
19 . The method of claim 11 , wherein the associated relationships comprise different product identifiers that are variants of same object.
20 . The method of claim 19 , wherein the variants of the same object comprise differences in at least one of: size, color, and pattern.