Binarization of MultiSpectral Document Images

Hollaus, Fabian; Sablatnig, Robert

doi:10.1007/978-3-319-23117-4_10

Cited by 8 publications

(10 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While these target detectors apply a linear transform to the data, the ACE applies a non-linear transformation based on second-order statistics. We are using ACE for target detection, since it proved to be superior to the afore mentioned methods [14], [1]. It is defined as…”

Section: B Target Detectionmentioning

confidence: 99%

“…The signature wavelength of the iron gall-based ink is varying between different images [1]. That is why a rough localization of potential ink pixels is performed prior to signature computation.…”

Section: A Pre-processingmentioning

confidence: 99%

“…The ACE reduces MS pixel vectors to a scalar which is close to one if a pixel vector is similar to the target signature. While we introduced this strategy in [1], the method is further improved in this paper by adding a spatial segmentation which forms the final binary image. In this processing stage the GrabCut [9], a semi-automated Graph-Cuts based segmentation, is fed with the results of the foreground estimation and the target detection rather than manual input.…”

Section: Introductionmentioning

confidence: 99%

“…While the majority of document image binarization techniques is designed for conventional gray-scale or RGB images, binarization techniques which use MS input data have become more mature recently [1]. Mitianodis and Papamarkos [6] use image fusion techniques such as the Independent Component Analysis (ICA) combined with a k-harmonic means classifier in order to binarize document images.…”

Section: Introductionmentioning

confidence: 99%

“…In addition, objects such as documents are imaged outside the visual range which renders faded-out or scraped-off ink legible again [1]. Document image binarization is a pre-processing step in document analysis systems which makes subsequent processing steps such as HandWriting Recognition (HWR) applicable or more efficient [2].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

MSIO: MultiSpectral Document Image BinarizatIOn

Hollaus

Sablatnig

2016

2016 12th IAPR Workshop on Document Analysis Systems (DAS)

Self Cite

View full text Add to dashboard Cite

MultiSpectral (MS) imaging enriches document digitization by increasing the spectral resolution. We present a methodology which detects a target ink in document images by taking into account this additional information. The proposed method performs a rough foreground estimation to localize possible ink regions. Then, the Adaptive Coherence Estimator (ACE), a target detection algorithm, transforms the MS input space into a single gray-scale image where values close to one indicate ink. A spatial segmentation using GrabCut on the target detection's output is computed to create the final binary image. To find a baseline performance, the method is evaluated on the three most recent Document Image Binarization COntests (DIBCO) despite the fact that they only provide RGB images. In addition, an evaluation on three publicly available MS datasets is carried out. The presented methodology achieved the highest performance at the MultiSpectral Text Extraction (MS-TEx) contest 2015.

show abstract

Section: B Target Detectionmentioning

confidence: 99%

Section: A Pre-processingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations