Look Who it Was Random video

Auditing in-store Point of Sale Material using machine learning on commodity hardware

443 views · Published 3 October 2018 · 47:06 · Indexed 20 September 2026

Channel: AIM Network · 2018 · Science & Technology

Watch on YouTube

At Cypher 2018
By Vivek V Krishnan Retail Consultant , Satyajit Nair Principal Engineer at Johnson Controls

More Details at www.analyticsindiasummit.com

Manufacturers spend a lot of time and money on deploying their marketing materials within the store to gain visibility and boost their sales. Often, the in-store marketing campaign suffers from poor execution and damage within the store which in-turn leads to poor ROI for manufacturers. Auditing the deployment of marketing materials is plagued by inaccuracy and lengthy turnaround times due to the extensive use of manual methods to audit these stores. Further, verification of the findings of these audits is a difficult and time-consuming process for manufacturers. Our proposed solution aims to increase the accuracy, while simultaneously reducing the turnaround times for these in-store audits by applying the latest in computer vision to make these audits objective and speedy. By using an advanced CNN architecture like Google’s InceptionNet, we train a neural network to detect brands and eventually different types of in-store displays. The output of the neural networks is used to automatically tag and score images from audits. Automating the process will also make the verification of these audits quick and easy for clients. This would save them critical time and arm them with timely information to react quickly to new developments in the market.

More from this channel