Extracting information from images requires an end-to-end deep learning pipeline consisting of dataset creation, data preprocessing and augmentation, model design using backbone architectures and transfer learning/fine-tuning, model training, evaluation, deployment, and explainability. In this talk, I will discuss how to employ such a pipeline to solve a variety of image problems like classification, object detection, autoencoders, image generation, counting, and captioning.

Local ODSC chapter in Los Angeles, USA

Instructor's Bio

Valliappa Lakshmanan

Director for Data Analytics and AI Solutions at Google

He is the Director for Data Analytics and AI Solutions on Google Cloud. His team builds software solutions for business problems using Google Cloud’s data analytics and machine learning products. His books on Machine Learning Design Patterns, BigQuery: The Definitive Guide and Data Science on Google Cloud Platform are available from Amazon. He designed and built much of the Warning Decision Support System Integrated Information (WDSS-II). The sample code that goes with his books (ml-design-patterns, bigquery-oreilly-book, data-science-on-gcp, and analysis-of-spatial-grids) is open-source and available from GitHub.


  • 1

    End-to-end machine learning on images

    • Ai+ Training

    • Webinar Link

    • Join ODSC West 2021 Training Conference