Unsupervised deep learning for text line segmentation

Barakat, Berat Kurar; Droby, Ahmad; Alaasam, Reem; Madi, Boraq; Rabaev, Irina; Shammes, Raed; El-Sana, Jihad

doi:10.1109/icpr48806.2021.9413308

Cited by 14 publications

(11 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We compare our results with those of supervised learning methods, Mask-RCNN [14] and FCN+EM [14], and an unsupervised deep learning method, UTLS [15]. Mask-RCNN is an instance segmentation algorithm which is fully supervised using the pixel labels of the text lines.…”

Section: Results On the Vml-ahte Datasetmentioning

confidence: 99%

“…The advantage of the proposed method on the supervised methods is zero labelling effort. Also UTLS [15] has zero labelling effort, however it requires to adjust a heuristic formula. The proposed method eliminates this formula by assuming the neighbouring patches contain the same text line patterns.…”

Section: Results On the Vml-ahte Datasetmentioning

confidence: 99%

“…However, the main concern is to find an objective function that will use a representation to capture text lines, although they are not labelled. Kurar et al [15] formulated this concern as the answer to the question of whether a given document patch contains a text line or space line. The answer is based on a human adjusted score.…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Unsupervised Learning of Text Line Segmentation by Differentiating Coarse Patterns

Barakat

Droby

Saabni

et al. 2021

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Despite recent advances in the field of supervised deep learning for text line segmentation, unsupervised deep learning solutions are beginning to gain popularity. In this paper, we present an unsupervised deep learning method that embeds document image patches to a compact Euclidean space where distances correspond to a coarse text line pattern similarity. Once this space has been produced, text line segmentation can be easily implemented using standard techniques with the embedded feature vectors. To train the model, we extract random pairs of document image patches with the assumption that neighbour patches contain a similar coarse trend of text lines, whereas if one of them is rotated, they contain different coarse trends of text lines. Doing well on this task requires the model to learn to recognize the text lines and their salient parts. The benefit of our approach is zero manual labelling effort. We evaluate the method qualitatively and quantitatively on several variants of text line segmentation datasets to demonstrate its effectivity.

show abstract

Section: Results On the Vml-ahte Datasetmentioning

confidence: 99%

Section: Results On the Vml-ahte Datasetmentioning

confidence: 99%

See 1 more Smart Citation

Unsupervised Learning of Text Line Segmentation by Differentiating Coarse Patterns

Barakat

Droby

Saabni

et al. 2021

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

show abstract

“…Barakat et al [9] suggested an unsupervised method for extracting text lines, which was driven by the relative variation in text lines and space between text lines. The number of foreground pixels over text lines differs significantly from the number of foreground pixels over text line gaps.…”

Section: Related Workmentioning

confidence: 99%

An approach to analysis of arabic text documents into text lines, words, and characters

Abdo

Abdu

Manza

et al. 2022

IJEECS

View full text Add to dashboard Cite

Text line extraction from a text document image and segmenting it into isolate words and segmenting these words into individual characters are considered as one of the most critical processes in OCR systems development and turning the document into a searchable electronic representation, this paper presents a new approach to analyze the Arabic text documents, the proposed approach contains four steps, preprocessing, text line segmentation, word segmentation, character segmentation. The horizontal projection method are used to detect and extract the text line from preprocessed text documents image, in word segmentation step The space threshold are computed to determine the spaces among connected components in text line as within-word space or between-words space for segmenting the text line into isolate words, finally thinning method applied to find the skeleton of segmented word and analyses geometric characteristics of the characters to detect ligatures and characters. The proposed approach was tested and evaluated on a set of 115 text images, this set contains images from the KFUPM Handwritten Arabic TexT (KHATT) database and some images produced by the authors. The experiment results are extremely encouraging, with a success rate of 98.6% for lines segmentation, 96% for words segmentation, and 87.1% for characters segmentation.

show abstract

“…To the best of our knowledge, there is no existing work that assigns a corresponding order to text translations. Of course, there are some works dealing with document layout, such as text line extraction [1,2], text line segmentation [3,4], and scene text detection [5]. Tere are also several works on recognizing characters such as Gurmukhi [6][7][8][9][10][11] and Devanagari [12], as well as a work on machine translation [13].…”

Section: Introductionmentioning

confidence: 99%

Learning-Based Ordering Characters on Ancient Document

Lee

Baek

Choi

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

Digitalizing and translating a scanned document image entails detecting the characters using a detector and translating the characters in the order they were detected with a translator. However, it is impossible to translate these characters correctly because the detector often detects them in any order. As a result, since it is critical to organize the recognized characters for proper translation, we propose ordering characters from documents with multiple variations using the strength of the learning-based model that learns the necessary operations from the data. In this task, it is difficult to order the characters written on antique handwritten documents that have deviations such as a bent or split line, as opposed to official records that have lines placed uprightly one by one. Because dealing with these many variants using a human-designed algorithm is problematic, we arrange characters printed on papers with diverse variations by taking advantage of a training model that can learn the appropriate function from data. Our method outputs both line id and y-axis and combines them to assign the sequential index. It is difficult to train using simply local regions because sequential character indexes in a large range include long-range dependencies. To solve this problem, we use network architecture to expand the receptive field as wide as possible. The network must learn to give various indexes to characters in similar places for each document because the number and area of characters vary for each document. We offer the ground truth assign method based on the absolute position to assign similar indexes to characters in similar places. Furthermore, even if the network uses absolute ground truth, the network may assign the incorrect line if the center coordinates of characters are biased in one direction. As a result, we employed the Region of Interest (ROI) from the pretrained coordinate layer, which contains position and trend information. We used the modified edit distance to compare the similarity of character indexes from the ground truth and our technique. In addition, we computed the modified fisher criterion to assess the degree of the clustering line. Consequently, our edit distance is just 0.43 times that of the human-designed algorithm, and our fisher criterion is 1.46 times that of the human-designed algorithm, improving the performance of human-designed algorithm.

show abstract

Unsupervised deep learning for text line segmentation

Cited by 14 publications

References 35 publications

Unsupervised Learning of Text Line Segmentation by Differentiating Coarse Patterns

Unsupervised Learning of Text Line Segmentation by Differentiating Coarse Patterns

An approach to analysis of arabic text documents into text lines, words, and characters

Learning-Based Ordering Characters on Ancient Document

Contact Info

Product

Resources

About