A graph-based approach for segmenting touching lines in historical handwritten documents

Fernandez-Mota, David; Lladós, Josep; Fornés, Alícia

doi:10.1007/s10032-014-0220-0

Cited by 31 publications

(19 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Evaluation results on the ICDAR 2013 dataset and comparison to existing methods [3,10,17,20,25,28] are summarized in Table 3. As shown, the proposed method compares favorably with existing methods.…”

Section: Evaluation Resultsmentioning

confidence: 99%

“…Since the result map had negative values in the textlines and positive values in the background, minimal energy paths that pass between the textlines were extracted by using the dynamic programming. Fernández-Mota et al [3] also proposed a path-finding method based on the similar framework. Yin et al [31] exploited the hierarchical structure in document images and presented a bottom-up grouping method: CCs of an input image were grouped into textlines by using a learned metric.…”

Section: Methods On Textline Detection In Handwritten Documentsmentioning

confidence: 99%

See 1 more Smart Citation

Textline detection in degraded historical document images

Ahn

Ryu

Koo

et al. 2017

J Image Video Proc.

View full text Add to dashboard Cite

This paper presents a textline detection method for degraded historical documents. Our method follows a conventional two-step procedure that the binarization is first performed and then the textlines are extracted from the binary image. In order to address the challenges in historical documents such as document degradation, structure noise, and skews, we develop new methods for the binarization and textline extraction. First, we improve the performance of binarization by detecting the non-text regions and processing only text regions. We also improve the textline detection method by extracting main textblock and compensating the skew angle and writing style. Experimental results show that the proposed method yields the state-of-the-art performance for several datasets.

show abstract

Section: Evaluation Resultsmentioning

confidence: 99%

Section: Methods On Textline Detection In Handwritten Documentsmentioning

confidence: 99%

Textline detection in degraded historical document images

Ahn

Ryu

Koo

et al. 2017

J Image Video Proc.

View full text Add to dashboard Cite

show abstract

“…The two text lines shown in (a.1) and (b.1) include many singleand multiple-touching pairs, but the system can segment and recognize them. In the first text line, the components 10 and 11 are a single-touching pair, abbreviated as p (10,11), but they can be separated by the coarse segmentation. The multiple-touching pairs p (4,5), p (19,20) and p (20,21) are segmented by the fine segmentation.…”

Section: Discussionmentioning

confidence: 99%

“…The multiple-touching pairs p (4,5), p (19,20) and p (20,21) are segmented by the fine segmentation. In the second text line, the coarse segmentation with stroke width estimation is effective for single-touching pairs p (11,12), p (12,13) and p (13,14) while the fine segmentation works for a singletouching pair p (7,8) and a multiple-touching pair p (14,15). On the other hand, Fig.…”

Section: Discussionmentioning

confidence: 99%

“…For text-line segmentation, current methods are classified into five groups: methods making use of projection profiles [7], those based on the Hough transform [8], smearing methods [9], component grouping methods [10] and other methods [11], [12]. We employ a combination of the smearing method and the method using projection profiles [5], which we do not discuss in this paper.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Segmentation Method of Single- and Multiple-Touching Characters in Offline Handwritten Japanese Text Recognition

Nguyen

Nakagawa

2017

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

SUMMARYThis paper presents a method to segment single-and multiple-touching characters in offline handwritten Japanese text recognition with practical speed. Distortions due to handwriting and a mix of complex Chinese characters with simple phonetic and alphanumeric characters leave optical handwritten text recognition (OHTR) for Japanese still far from perfection. Segmentation of characters, which touch neighbors on multiple points, is a serious unsolved problem. Therefore, we propose a method to segment them which is made in two steps: coarse segmentation and fine segmentation. The coarse segmentation employs vertical projection, stroke-width estimation while the fine segmentation takes a graphbased approach for thinned text images, which employs a new bridge finding process and Voronoi diagrams with two improvements. Unlike previous methods, it locates character centers and seeks segmentation candidates between them. It draws vertical lines explicitly at estimated character centers in order to prevent vertically unconnected components from being left behind in the bridge finding. Multiple candidates of separation are produced by removing touching points combinatorially. SVM is applied to discard improbable segmentation boundaries. Then, ambiguities are finally solved by the text recognition employing linguistic context and geometric context to recognize segmented characters. The results of our experiments show that the proposed method can segment not only single-touching characters but also multiple-touching characters, and each component in our proposed method contributes to the improvement of segmentation and recognition rates.

show abstract