A novel ring radius transform for video character reconstruction

Shivakumara, Palaiahnakote; Phan, Trung Quy; Bhowmick, Souvik; Tan, Chew Lim; Pal, Umapada

doi:10.1016/j.patcog.2012.07.008

Cited by 18 publications

(3 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Obstacles with large area or background objects passing through characters would frustrate most edge-based methods. Recently proposed Ring Radius Transform (RRT) [15] could be a remedy approach. One of our further works is to solve this common problem by extending LHA with the cooperation of RRT.…”

Section: Methodsmentioning

confidence: 99%

Scene Character Detection by an Edge-Ray Filter

Shivakumara

Uchida

2013

2013 12th International Conference on Document Analysis and Recognition

Self Cite

View full text Add to dashboard Cite

Abstract-Edge is a type of valuable clues for scene character detection task. Generally, the existing edge-based methods rely on the assumption of straight text line to prune away the noncharacter candidates. This paper proposes a new edge-based method, called edge-ray filter, to detect the scene character. The main contribution of the proposed method lies in filtering out complex backgrounds by fully utilizing the essential spatial layout of edges instead of the assumption of straight text line. Edges are extracted by a combination of Canny and Edge Preserving Smoothing Filter (EPSF). To effectively boost the filtering strength of the designed edge-ray filter, we employ a new Edge Quasi-Connectivity Analysis (EQCA) to unify complex edges as well as contour of broken character. Label Histogram Analysis (LHA) then filters out non-character edges and redundant rays through setting proper thresholds. Finally, two frequentlyused heuristic rules, namely aspect ratio and occupation, are exploited to wipe off distinct false alarms. In addition to have the ability to handle special scenarios, the proposed method can accommodate dark-on-bright and bright-on-dark characters simultaneously, and provides accurate character segmentation masks. We perform experiments on the benchmark ICDAR 2011 Robust Reading Competition dataset as well as scene images with special scenarios. The experimental results demonstrate the validity of our proposal.

show abstract

Section: Methodsmentioning

confidence: 99%

Scene Character Detection by an Edge-Ray Filter

Shivakumara

Uchida

2013

2013 12th International Conference on Document Analysis and Recognition

Self Cite

View full text Add to dashboard Cite

show abstract

“…The data is transformed into a more effectively processed format. Shivakumara et al [4,6] introduced a reconstruction technique in video images for the improvement of character recognition rate by using novel Ring Radius Transform (RRT) and the concept of medial pixels on characters with broken contours in the edge domain. The inner and outer contour symmetry information is exploited to reconstruct the gaps of the broken character that outperforms the state-of-the-art methods in terms of relative error and character recognition rate.…”

Section: Related Workmentioning

confidence: 99%

An Approach for Offline Handwritten Character Shape Reconstruction Using Active Contour and Morphological Techniques

Garg

Kaur

Parashar

2021

Advanced Computing and Intelligent Technologies

View full text Add to dashboard Cite

Optical Character Recognition (OCR) models have increased their popularity in the past decade. The accuracy of OCR models has been increased with the new research models, which have improved the recognition rate of the OCR algorithms. In this work, the improved model for character shape restoration practices of handwritten documents is proposed. The proposed model involves the active contour selection (ACS) model to detect object locations, where the shape restoration method can be applied. The contour selection model utilizes the snake model for marking and localization of the characters in all possible shapes using polygonal point marking for the extraction of the object boundary. Then the distantia factor is applied over the detected contour regions to analyze the distance between two contour regions, which decides upon the selection of the target regions, where the opening needs to be closed. The combination of adaptive dilation and adaptive erosion has been utilized for filling or closing of the target region, which verifies the angle of the contour objects and thickness for removal of the smudge regions. The proposed model has undergone various experiments to assess the proposed model quality against the existing models. The proposed model has been found efficient and improved in terms of obtained performance parameters of peak signal-to-noise ratio (PSNR), mean squared error (MSE) and relative error. The proposed model has been recorded with 0.0006 percent relative error compared to 0.03, 0.26 and 0.03 in the case of the ring radius method, restoration through the medial axis and iterative midpoint method, respectively, which shows the robustness of the proposed model.

show abstract

“…Extremal Regions are used in [17] to extract candidate text regions and non-text regions are removed by a classifier trained with text specific features. A novel ring radius transform [20] is proposed to restore broken contours in the edge domain to reconstruct a binary character. By combining with medial pixels and the symmetric properties of character stroke, it could restore the broken parts on the inner and outer contour of the character.…”

Section: Segmentation Based Approachmentioning

confidence: 99%