Dynamic Product Marketing Through Computer Vision

1 . A method comprising:

collecting image data;

identifying a set of product instances within the environment

for each product instance of the set of product instances, analyzing image data of the product instance and thereby determining a product presentation variation associated with the product instance;

detecting user-item interactions associated with the set of product instances; and

analyzing user-item interactions associated with the set of product instances according to the product presentation variations.

2 . The method of claim 1 , wherein identifying the set of product instances within the environment comprises, for each product instance of the set of product instances, identifying the product instance with a product identifier.

3 . The method of claim 2 , wherein analyzing user-item interactions associated with the set of product instances according to the product presentation variations comprises analyzing product instances with a shared product identifier and determining a select product presentation variation for the product identifier.

4 . The method of claim 3 , further comprising updating a marketing system with the select product presentation variation.

5 . The method of claim 3 , further comprising altering settings in a product operations computing system based on the select product presentation variation.

6 . The method of claim 1 , further comprising, for each product instance, segmenting the product instance in the image data to yield the image data of the product instance that used in determining the product presentation variation.

7 . The method of claim 6 , wherein analyzing image data of the product instance further comprises classifying product packaging of the product instance as a packaging variation classification using the image data of the product instance, wherein the product presentation variation determined for the product instance is based, at least in part, on the packaging variation classification.

8 . The method of claim 7 , wherein classifying the product packaging as a packaging variation classification comprises applying a product packaging classification model to the image data of the product instance.

9 . The method of claim 7 , wherein classifying the product packaging as a packaging variation classification comprises performing product packaging comparison of the image data of the product instance and determining a product presentation variation that includes at least a packaging variation classification.

10 . The method of claim 6 , wherein analyzing image data of the product instance further comprises extracting text from the image data of the product instance by performing optical character recognition on the image data of the product instance; and the product presentation variation is based at least in part on the text of the product instance.

11 . The method of claim 7 , receiving specification of a set of pre-configured product presentation variations; and wherein classifying the product packaging comprises classifying the product packaging from a set of product presentation variations that includes at least the set of pre-configured product presentation variations.

12 . The method of claim 7 , wherein classifying the product packaging comprises classifying the image data as a product presentation variation as one of a pre-existing product presentation variation classifications upon detecting a matching condition with a pre-existing product presentation variation and, upon failing to detect a matching condition, associating the product instance with a new product presentation variation classification.

13 . The method of claim 1 , wherein, for each product instance of at least a subset of the set of product instances, analyzing image data of the product instance comprises detecting a marketing item instance in the image data, analyzing image data of the marketing item instance, and thereby determining the product presentation variation based on analysis of the marketing item instance.

14 . The method of claim 1 , wherein, for each product instance of at least a subset of the set of product instances, analyzing image data of the product instance comprises detecting a marketing item instance in the image data, analyzing image data of the marketing item instance, and thereby determining the product presentation variation based on analysis of the marketing item instance.

15 . The method of claim 1 , further comprising receiving a product map data model, wherein identifying the set of product instances comprises identifying the set of product instances in the image data using the product map.

16 . The method of claim 1 , wherein detecting user-item interactions associated with the set of product instances comprises detecting instances of detecting user attention, detecting item pickup, and detecting a product purchase.

17 . A non-transitory computer-readable medium storing instructions that, when executed by one or more computer processors of a communication platform, cause the communication platform to:

collecting image data;

identifying a set of product instances within the environment

for each product instance of the set of product instances, analyzing image data of the product instance and thereby determining a product presentation variation associated with the product instance;

detecting user-item interactions associated with the set of product instances; and

analyzing user-item interactions associated with the set of product instances according to the product presentation variations.

18 . A system comprising of:

one or more computer-readable mediums storing instructions that, when executed by the one or more computer processors, cause a computing platform to perform operations comprising:

collecting image data;

identifying a set of product instances within the environment

for each product instance of the set of product instances, analyzing image data of the product instance and thereby determining a product presentation variation associated with the product instance;

detecting user-item interactions associated with the set of product instances; and

analyzing user-item interactions associated with the set of product instances according to the product presentation variations.