The appropriate action is to edit the product tags so that the incorrectly detected object is associated with its correct product classification. Product identification depends on the relationship between the detected item and the product label assigned to it. When two packages have highly similar visual characteristics, a detected object can require correction to ensure subsequent shelf-analysis results reflect the correct SKU or competitor product.

Share-of-shelf measurements depend on accurate identification of the products occupying the shelf. If an item belonging to a competitor is classified as the company ' s own product, the calculated shelf presence becomes overstated and the resulting compliance data is unreliable. Correcting the product tag addresses the classification problem at its source rather than manipulating the physical shelf or abandoning the assessment.

This approach also aligns with the underlying object-detection model. Detection systems generate identified objects and associate those detections with product classifications. Accurate labels and representative examples are fundamental to reliable image-based object recognition. Where product appearance changes substantially, maintaining accurate classification information is essential to preserving measurement quality.

Creating a different planogram would change the expected merchandising arrangement but would not correct an incorrectly classified detected product.

Study Guide Topics:Platform - > Einstein Object Detection - > Product Recognition - > Product Tags - > Share-of-Shelf Accuracy

===============