[Adapt] [Seminar] Fault-tolerant OCR from Semi-structured Medical Images

[Jinyi Lu]

Hi Adapters,

Our seminar will take place tomorrow at 5pm in room SEIEE3-528. I will give
a talk about my project "Fault-tolerant Optical Character Recognition from
Semi-structured Medical Images".

In the talk, I will first describe the problem we aim to solve, that is,
extraction of text fields from semi-structured medical images. Then I will
show you a spatial, fault-tolerant description language that specifies
rough formats of the images, and how we automatically generate the most
reliable parsing of the images. I will also talk about our incremental
correction framework that further improves our performance.

Hope you can enjoy this talk.

Thanks,
Jinyi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://cs.sjtu.edu.cn/pipermail/adapt/attachments/20150324/581cb9f6/attachment.html>