Solving computer vision and machine learning problems for real-time shelf visibility
In this talk, we provide an exposition into Zebra’s Next Generation Shelf Sensing solution. It is a solution which captures data in various customer environments, performing analysis of the captured data to provide greater visibility into retailers’ shelf and store environments.
At present, retailers visually asses the shelf status by scanning the "Shelf Label" barcodes to collect relevant information regarding the product which is supposed to be on the shelf from the retailer's SKU database. This is extremely labor intensive and error prone. The data generated by the human associates is often not in a machine readable form, requiring manual processing before the problem can be resolved. This leads to delays in the identifying the out of stock or non compliance to standard store operating procedure such as ensuring the right product is placed at the right shelf location at the right price. These errors or non compliance results in lost sales and dissatisfied customers.
The Next Generation Shelf Sensing solution provides reliable data in a machine readable form, providing increased inventory visibility, improved quality of merchandise displays and higher customer satisfaction. The core technology within the Next Generation Shelf Sensing Solution is the use of computer vision and machine learning algorithms providing analysis on the data captured through various sensors. We thus provide a high level overview into the kinds of problems with a computer vision and machine learning perspective that the Next Generation Shelf Sensing Solution encounters and demonstrate how they would be solved.