U-Net-bin: hacking the document image binarization contest
| Autor: | Dmitry P. Nikolaev, Dmitrii Alexeevich Ilin, Pavel Bezmaternykh |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2019 |
| Předmět: |
Computer science
ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION 02 engineering and technology CONTEST 01 natural sciences Bin Image (mathematics) 010309 optics 0103 physical sciences 0202 electrical engineering electronic engineering information engineering lcsh:Information theory lcsh:QC350-467 Electrical and Electronic Engineering Hacker dibco Information retrieval business.industry historical document processing Deep learning deep learning Document analysis lcsh:Q350-390 Atomic and Molecular Physics and Optics training dataset augmentation Computer Science Applications document analysis ComputingMethodologies_DOCUMENTANDTEXTPROCESSING 020201 artificial intelligence & image processing u-net architecture Artificial intelligence binarization business lcsh:Optics. Light |
| Zdroj: | Компьютерная оптика, Vol 43, Iss 5, Pp 825-832 (2019) |
| ISSN: | 2412-6179 0134-2452 |
| Popis: | Image binarization is still a challenging task in a variety of applications. In particular, Document Image Binarization Contest (DIBCO) is organized regularly to track the state-of-the-art techniques for the historical document binarization. In this work we present a binarization method that was ranked first in the DIBCO`17 contest. It is a convolutional neural network (CNN) based method which uses U-Net architecture, originally designed for biomedical image segmentation. We describe our approach to training data preparation and contest ground truth examination and provide multiple insights on its construction (so called hacking). It led to more accurate historical document binarization problem statement with respect to the challenges one could face in the open access datasets. A docker container with the final network along with all the supplementary data we used in the training process has been published on Github. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|