Description:

Images are one of the main elements on eCommerce sites to support the selection of the product and the sales. We have done two major improvements over these images:

1. Isolate the images which do not provide enough value to be shared (redundant, poor quality, or misleading) - Images with low resolution, image quality, sharpness, white balance, product centering, margin issues etc.

2. Remove duplicate images that pollute the presentation of our products

AI technologies used:

Image Classification Image Processing Deep Neural Networks

Data Input:

Digital Asset Management - Image library Product Information Data

Model Output:

Custom Image and metadata data model tailored to be uploaded to the DAM

Prerequisites:

Python

Image manipulation libraries - OpenCV and Pillow

Parallel Processing - At code level using GPUs and at Infrastructure level

Cloud Computing and Infrastructure as Code - AWS; Terraform;

Deep Learning basics - CNNs

Speaker Info:

Speaker: Kiran Y - Senior AI Engineering Manager - Data Science @ Schneider Electric, I have over 20 years of experience in software engineering and product development, with a PhD in computer science focused on image processing, NLP and predictive Time series forecast & modeling. I have experience of leading a team of data scientists and AI engineers to develop high quality E2E AI solutions and non-AI solutions for various domains, such as healthcare, telecom, supply chain, and retail.

Along with:

Pavan kamath - Lead, AI Product Manager

Lisa J - Lead AI Software Engineer

Nirmal J - AI Software Engineer

Kranthi Kumar Mutyalapalli - Senior Data Scientist

Speaker Links:

Linkedin

Simulation Study Publication

PeerJ Publication