search

fh2019ustc / DocTr

The official code for “DocTr: Document Image Transformer for Geometric Unwarping and Illumination Correction”, ACM MM, Oral Paper, 2021.
Other
345 stars 48 forks source link

About training #7

Closed CVUsers closed 1 year ago

CVUsers commented 2 years ago

Hello, 1: when are you going to open source training? 2: How fast is your plan? 3: What is used in the document detection of your scheme and how is the effect

fh2019ustc commented 2 years ago

Hi, thanks for your attention to our work. 1&2. We will open the source code when our another work DocScanner is accepted.

  1. I do not quite understand the third question.
CVUsers commented 2 years ago

How is your docscanner implemented? Can you give us a general idea? Semantic or instance segmentation or key point detection?

fh2019ustc commented 2 years ago

You can refer to the paper for details. https://arxiv.org/pdf/2110.14968.pdf