Historical Document Image Binarization: A Review

Tensmeyer, Chris; Martinez, Tony

doi:10.1007/s42979-020-00176-1

Cited by 45 publications

(14 citation statements)

References 162 publications

(314 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To measure the quality of the binarized content, we followed the same evaluation protocol used by Kota et al [32] where the metrics and tools from H-DIBCO [49] were used. These metrics try to approximate human perception of quality on binary image [51]. First, we consider the peak signal to noise ratio (PSNR) which measures how close the binary image is to the ground truth (higher is better).…”

Section: A Handwritten Content Extractionmentioning

confidence: 99%

FCN-LectureNet: Extractive Summarization of Whiteboard and Chalkboard Lecture Videos

Davila

Setlur

et al. 2021

IEEE Access

View full text Add to dashboard Cite

Recording and sharing of educational or lecture videos has increased in recent years. Within these recordings, we find a large number of math-oriented lectures and tutorials which attract students of all levels. Many of the topics covered by these recordings are better explained using handwritten content on whiteboards or chalkboards. Hence, we find large numbers of lecture videos that feature the instructor writing on a surface. In this work, we propose a novel method for extraction and summarization of the handwritten content found in such videos. Our method is based on a fully convolutional network, FCN-LectureNet, which can extract the handwritten content from the video as binary images. These are further analyzed to identify the unique and stable units of content to produce a spatial-temporal index of handwritten content. A signal which approximates content deletion events is then built using information from the spatial-temporal index. The peaks of this signal are used to create temporal segments of the lecture based on the notion that sub-topics change when large portions of content are deleted. Finally, we use these segments to create an extractive summary of the handwritten content based on key-frames. This will facilitate content-based search and retrieval of these lecture videos. In this work, we also extend the AccessMath dataset to create a novel dataset for benchmarking of lecture video summarization called LectureMath. Our experiments on both datasets show that our novel method can outperform existing methods especially on the larger and more challenging dataset. Our code and data will be made publicly available.INDEX TERMS fully convolutional networks, handwritten text detection, image binarization, lecture videos, video summarization.

show abstract

Section: A Handwritten Content Extractionmentioning

confidence: 99%

FCN-LectureNet: Extractive Summarization of Whiteboard and Chalkboard Lecture Videos

Davila

Setlur

et al. 2021

IEEE Access

View full text Add to dashboard Cite

show abstract

“…This has been shown to be effective in situations where a colour band shows a cluster of contrast level (Fujita and Aya, 2004). Image binarisation typically uses threshold algorithms to split histograms into two, sometimes using several sub-processes such as background detection and edge detection (Malepati, 2010) to ensure foreground textures stand out as much as possible against their backgrounds (Tensmeyer and Martinez, 2020). This can be especially useful when, despite all previous efforts, tracers are still not prevalent to their background and surroundings.…”

Section: Image Manipulationmentioning

confidence: 99%

Considerations When Applying Large-Scale PIV and PTV for Determining River Flow Velocity

et al. 2021

View full text Add to dashboard Cite

Large-scale image velocimetry is a novel approach for non-contact remote sensing of flow in rivers. Research within this topic has largely focussed on developing specific aspects of the image velocimetry work-flow, or alternatively, testing specific tools or software using case studies. This has resulted in the development of a multitude of techniques, with varying practice being employed between groups, and authorities. As such, for those new to image velocimetry, it may be hard to decipher which methods are suited for particular challenges. This research collates, synthesises, and presents current understanding related to the application of particle image velocimetry (PIV) and particle tracking velocimetry (PTV) approaches in a fluvial setting. The image velocimetry work-flow is compartmentalised into sub-systems of: capture optimisation, pre-processing, processing, and post-processing. The focus of each section is to provide examples from the wider literature for best practice, or where this is not possible, to provide an overview of the theoretical basis and provide examples to use as precedence and inform decision making. We present literature from a range of sources from across the hydrology and remote sensing literature to suggest circumstances in which specific approaches are best applied. For most sub-systems, there is clear research or precedence indicating how to best perform analysis. However, there are some stages in the process that are not conclusive with one set method and require user intuition or further research. For example, the role of external environmental conditions on the performance of image velocimetry being a key aspect that is currently lacking research. Further understanding in areas that are lacking, such as environmental challenges, is vital if image velocimetry is to be used as a method for the extraction of river flow information across the range of hydro-geomorphic conditions.

show abstract

“…In this review, we begin by presenting the two categories and subsequently discuss competitions in document binarization that have been held in conjunction with the ICFHR and ICDAR conferences. In addition, we refer readers to further details in two current and comprehensive research efforts [10], [11] in which a review of important works in the field of the document binarization has been explored and possible future research directions in this subject have been presented.…”

Section: Related Workmentioning

confidence: 99%

Binarization of Degraded Document Images Using Convolutional Neural Networks and Wavelet-Based Multichannel Images

2020

View full text Add to dashboard Cite

Convolutional neural networks (CNNs) have previously been broadly utilized to binarize document images. These methods have problems when faced with degraded historical documents. This paper proposes the utilization of CNNs to identify foreground pixels using novel input-generated multichannel images. To create the images, the original source image is decomposed into wavelet subbands. Then, the original image is approximated by each subband separately, and finally, the multichannel image is constituted by arranging the original source image (grayscale image) as the first channel and the approximated image by each subband as the remaining channels. To achieve the best results, two scenarios are considered, that is, two-channel and four-channel images, and then fed into two types of CNN architectures, namely, single and multiple streams. To investigate the effect of the multichannel images proposed as network inputs, the CNNs used in the architectures are three popular networks, namely, Unet, SegNet, and DeepLabv3+. The experimental results of the scenarios demonstrate that our method is more successful than the three CNNs when trained by the original source images and proves competitive performance in comparison with state-of-the-art results using the DIBCO database.

show abstract

Historical Document Image Binarization: A Review

Cited by 45 publications

References 162 publications

FCN-LectureNet: Extractive Summarization of Whiteboard and Chalkboard Lecture Videos

FCN-LectureNet: Extractive Summarization of Whiteboard and Chalkboard Lecture Videos

Considerations When Applying Large-Scale PIV and PTV for Determining River Flow Velocity

Binarization of Degraded Document Images Using Convolutional Neural Networks and Wavelet-Based Multichannel Images

Contact Info

Product

Resources

About