This talk will demonstrate how Haleon uses computer vision and cloud technology to manage its digital assets.
Haleon is a world-leading consumer health company with a portfolio of product across the areas of oral health, pain relief, and digestion. As a market leader it holds a wealth of campaign material in its effort to bring these products to the consumers who need them. This comprises over a petabyte of digital assets such as packshot imagery, promo videos, web-banners, etc. One of the challenges in managing these assets is ensuring the metadata on the assets is present and complete. In this talk, we describe the efforts of the Data Science team at Haleon in collaboration with the Marketing and MarTech teams to enhance the metadata and enable accessibility, structure, and searchability of these assets to ensure their reuseability for adaptations. To achieve this, we devised a cloud-based solution that extracts key metadata from these assets, such as brand, category, campaign and market information, as well as mining creation dates.
We used Docker and Terraform for cloud management, and components like Event Grid, Kubernetes, and Cognitive Services in a pipeline for efficient processing. Throughout this presentation, we will take you through our journey of leveraging computer vision and cloud technology to improve end-to-end asset management at Haleon.
Technical Level: Technical practitioner