Amazon SageMaker’s Built-in Algorithm Webinar Series: Bonus Topic: Optical Character Recognition

Опубликовано 19 ноября 2018, 21:30

Learn more about Amazon SageMaker at - amzn.to/2zdoY7u.
This webinar replay presents a prototype for optical character recognition for handwritten text. An internal Amazon team has produced a prototype for optical character recognition for handwritten text, a use-case that is often requested by customers in verticals such as healthcare and finance. The prototype uses a combination of computer vision algorithms and natural language processing to break apart images that include handwritten notes and extract the textual content. In this webinar, we will show you how we use MXNet Gluon, outlining a pipeline for Optical Character Recognition that uses a CNN for page segmentation, SSD (Single Shot Multibox Detection) for line detection and CNN-BiLSTM with CTC Loss for handwriting recognition. We also investigate extensions to improve performance with Lexicon Search and Language Modelling.

Learn more - docs.aws.amazon.com/sagemaker/...