Scene text segmentation using low variation extremal regions and sorting based character grouping

Šarić, Matko

doi:10.1016/j.neucom.2017.05.021

Cited by 12 publications

(8 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finally, heuristic rules are used for misclassified character filtering. In [20], the character candidates are extracted from low-variation ERs and classified using a support vector machine (SVM) and geometrical features. e obtained characters are grouped into text lines using heuristic rules, and a final restoration stage is considered if adjacent regions satisfy a set of predefined conditions.…”

Section: Related Workmentioning

confidence: 99%

“…In the last years, the MSER and SWT techniques have become the most used techniques for text detection process due to their invariance to scale and rotation transformations. Besides, not only the MSER but also all extremal regions (ERs) are used for text segmentation [16][17][18][19][20]. However, ERbased methods need to process multiple repeated regions to obtain correct character segmentation, generating classification errors and a high computational cost.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Natural Scene Text Detection and Segmentation Using Phase-Based Regions and Character Retrieval

Diaz-Escobar

Kober

2020

Mathematical Problems in Engineering

View full text Add to dashboard Cite

Multioriented text detection and recognition in natural scene images are still challenges in the document analysis and computer vision communities. In particular, character segmentation plays an important role in the complete end-to-end recognition system performance. In this work, a robust multioriented text detection and segmentation method based on a biological visual system model is proposed. The proposed method exploits the local energy model instead of a common approach based on variations of local image pixel intensities. Features such as lines and edges are obtained by searching for the maximum local energy utilizing the scale-space monogenic signal framework. The candidate text components are extracted from maximally stable extremal regions of the local phase information of the image. The candidate regions are filtered by their phase congruency and classified as text and nontext components by the AdaBoost classifier. Finally, misclassified characters are restored, and all final characters are grouped into words. Experimental results show that the proposed text detection and segmentation method is invariant to scale and rotation changes and robust to perspective distortions, blurring, low resolution, and illumination variations (low contrast, high brightness, shadows, and nonuniform illumination). Besides, the proposed method achieves often a better performance compared with state-of-the-art methods on typical natural scene datasets.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Natural Scene Text Detection and Segmentation Using Phase-Based Regions and Character Retrieval

Diaz-Escobar

Kober

2020

Mathematical Problems in Engineering

View full text Add to dashboard Cite

show abstract

“…It is done to effectively enhance the regions with different luminance and to mitigate the impact of over enhancement before conveying the enhanced image to text detection stage. The enhanced image is used for post text detection, segmentation and recognition [12][13][14][15][16][17]. The extracted text is then pronounced by using a Microsoft TTS [18].…”

Section: Contributionmentioning

confidence: 99%

“…The enhanced image is processed by the text candidates regions detection. In literature, different approaches are utilized to identify the text candidates such as the Stroke Feature Transform (SFT), the Stroke Width Transform (SWT), the Extremal Regions (ERs) and Maximally Stable Extremal Regions (MSERs) [12][13][14]. The MSER is well known for its text detection robustness [21] and is utilized in the proposed solution.…”

Section: Text Detectionmentioning

confidence: 99%

See 1 more Smart Citation

A Combination of DWT CLAHE and Wiener Filter for Effective Scene to Text Conversion and Pronunciation

Qaisar

Hammad

Khan

2020

J. Electr. Eng. Technol.

View full text Add to dashboard Cite

An effective scene to text conversion and its pronunciation is realized. An intelligent combination of Discrete Wavelet Transform (DWT), Contrast Limited Adaptive Histogram Equalization (CLAHE), Wiener filter and adaptive weighted average is utilized for the image enhancement. Subsequently, the Maximally Stable Extremal Region (MSER) is used to detect the text regions. Afterward, the geometrical and contour based approaches filter out the non-text MSERs. The connected component concept is used to group the text candidates. In next step the Optical Character Recognition (OCR) recognizes the text. The Microsoft speech to text synthesizer pronounces the extracted text. The system applicability is tested by using the standard robust reading competition dataset. The designed method secures 93% precision in text segmentation and 89.9% precision in end-to-end recognition.

show abstract

Scene Text Segmentation Based on Local Image Phase Information and MSER Method

Diaz-Escobar

Kober

2018

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Scene text segmentation using low variation extremal regions and sorting based character grouping

Cited by 12 publications

References 12 publications

Natural Scene Text Detection and Segmentation Using Phase-Based Regions and Character Retrieval

Natural Scene Text Detection and Segmentation Using Phase-Based Regions and Character Retrieval

A Combination of DWT CLAHE and Wiener Filter for Effective Scene to Text Conversion and Pronunciation

Scene Text Segmentation Based on Local Image Phase Information and MSER Method

Contact Info

Product

Resources

About