Computer Vision for Omnichannel Retail: Intelligent Analysis and Selection of Product Images at Scale
Computer Vision for Omnichannel Retail: Intelligent Analysis and Selection of Product Images at Scale


Content quality of product catalog plays a critical role in retail. In particular, visual content such as product images influences customers' engagement and purchase decisions more heavily than plain text. With the rapid growth of omnichannel retail experience and the advent of computer vision and AI, traditional content management systems are giving way to automated scalable systems. This talk presents such a machine learning driven system for evaluating, understanding and optimally selecting product images for customers. For a given product, the system aggregates images from various sources, analyzes and filters them for poor quality or non-compliant content, de-duplicates them, classifies them based on viewpoint and relevance to the product, and finally, selects a set of images with optimal count and quality. The system is empowered by an array of state-of-the-art technologies, ranging from deep image classification, deep regression, object detection to traditional computer vision. The talk will describe individual models and their performances as well as how they work together in production. More importantly, the talk will highlight the unique challenges of developing machine learning models in the real world such as noisy and scarce data, severe class imbalance, strict budget on annotation and review by humans and it will discuss the tricks and strategies such as artificial data generation, incremental generation of training set, transfer learning etc. that were adopted to address these challenges.


Abon leads a data science team at Walmart Labs, Sunnyvale. His team develops machine learning and computer vision based solutions for content enrichment and content quality improvement in the e-commerce domain. His team uses deep learning to classify and analyze product images and to infer item-to-item relationships. Earlier, Abon worked as a Research and Development Engineer in the Department of Computational Imaging at Intel Corporation. He developed scalable data analysis and visualization algorithms at Argonne National Laboratory and Oak Ridge National Laboratory. Abon graduated from the Ohio State University, Columbus, OH with a PhD in Computer Science and Engineering Department. His doctoral research and publications focused on exploratory analysis of large-scale scientific data.

Privacy Settings
We use cookies to enhance your experience while using our website. If you are using our Services via a browser you can restrict, block or remove cookies through your web browser settings. We also use content and scripts from third parties that may use tracking technologies. You can selectively provide your consent below to allow such third party embeds. For complete information about the cookies we use, data we collect and how we process them, please check our Privacy Policy
Consent to display content from - Youtube
Consent to display content from - Vimeo
Google Maps
Consent to display content from - Google