DewarpNet: Single-Image Document Unwarping With Stacked 3D and 2D Regression Networks

ICCV 2019 Sagnik Das Ke Ma Zhixin Shu Dimitris Samaras Roy Shilkrot

Capturing document images with hand-held devices in unstructured environments is a common practice nowadays. However, "casual" photos of documents are usually unsuitable for automatic information extraction, mainly due to physical distortion of the document paper, as well as various camera positions and illumination conditions... (read more)

PDF Abstract

Results from the Paper


 Ranked #1 on MS-SSIM on DocUNet (using extra training data)

     Get a GitHub badge
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK USES EXTRA
TRAINING DATA
BENCHMARK
MS-SSIM DocUNet DewarpNet MS-SSIM 0.47 # 1
SSIM DocUNet DewarpNet SSIM 0.493146 # 2
Local Distortion DocUNet DewarpNet LD 8.98 # 1

Methods used in the Paper


METHOD TYPE
🤖 No Methods Found Help the community by adding them if they're not listed; e.g. Deep Residual Learning for Image Recognition uses ResNet