A Novel Technique for English Font Recognition Using Support Vector Machines

Ramanathan, R.; Thaneshwaran, L.; Viknesh, V.; Arunkumar, T.; Yuvaraj, P.

doi:10.1109/artcom.2009.89

Cited by 32 publications

(16 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A 2D Gabor is a Gaussian modulated sinusoid in the spatial domain and a shifted as a shifted Gaussian in the frequency domain. It is represented by equations (1) and (2), x y x g (1) θ θ θ θ cos sin '…”

Section: Theory Of Gabor Filtersmentioning

confidence: 99%

“…In the problem of facial expression recognition, feature extraction can be done by using Fischer faces [3], Embedded Hidden Markov Model [4] etc. Gabor filters have been used extensively in many applications for feature extraction such as in facial expression recognition [2 & 6], Font recognition [1], and Neonatal pain recognition [5]. In the proposed method, a novel method of using Gabor filter for feature extraction is tested for OCR applications .The method is a very versatile method and has its use in many other applications.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Robust Feature Extraction Technique for Optical Character Recognition

Ramanathan

Nair

Thaneshwaran

et al. 2009

2009 International Conference on Advances in Computing, Control, and Telecommunication Technologies

Self Cite

View full text Add to dashboard Cite

Optical Character Recognition (OCR) is a classical research field and has become one of most thriving applications in the field of pattern recognition. Feature extraction is a key step in the process of OCR, which in fact is a deciding factor of the accuracy of the system. This paper proposes a novel and robust technique for feature extraction using Gabor Filters, to be employed in the OCR. The use of 2D Gabor filters is investigated and features are extracted using these filters. The technique generally extracts fifty features based on global texture analysis and can be further extended to increase the number of features if necessary. The algorithm is well explained and is found that the proposed method demonstrated better performance in efficiency. In addition, experimental results show that the method gains high recognition rate and cost reasonable average running time.

show abstract

Section: Theory Of Gabor Filtersmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Robust Feature Extraction Technique for Optical Character Recognition

Ramanathan

Nair

Thaneshwaran

et al. 2009

2009 International Conference on Advances in Computing, Control, and Telecommunication Technologies

Self Cite

View full text Add to dashboard Cite

show abstract

“…They were able to use second dimension wavelet decomposition for Italic font recognition. Ramanathan et al proposed a new method which used Gabor filter for English font recognition [3]. Zhu et al also utilized Gabor filter for recognition of Chinese fonts [4].…”

Section: Introductionmentioning

confidence: 99%

A New Method for Discriminating Printers Based on Contours Qualities of Printed Characters Using Wavelet Decomposition

Furukawa

2013

2013 12th International Conference on Document Analysis and Recognition

View full text Add to dashboard Cite

This article described a new method for discriminating models of laser printers by means of their printed characters, in particular details in contours of characters. A method we proposed was based on evaluating qualities in contours of printed characters using wavelet decomposition. Recently most of characters printed by laser printers were originally stored in printers or computers as vector outline such as Bezier or Spline. Raster Image Processor (RIP) implemented as hardware or software in printers or computers rasterized the outline into a pattern which was composed of subtle vast dots. There was a variety of types in contours of printed characters which were printed by each printer model. In Japan, stalkers typed their threatening letters using common used fonts such as MS Mincho in Japanese and Times or Century in English. Even though the kinds of fonts were known, there was not evidence since these fonts were equipped in almost all computers in Japan. Therefore a new method to discriminate models of printers was desired. Even though same too common fonts were used in threatening letters, subtle differences among contours of printed characters were observed since there was a variety of methods which each maker adopted in rendering and screening which converted from outline to pattern of dots. In order to detect the subtle differences among contours of printed documents, the article utilized wavelet decomposition and a high resolution i.e., 5400dpi flat bed image scanner. The article also used a simple method to analyze results of wavelet decomposition which was counting numbers of zerocrossing points at each scale of decomposition. The results of the experiment showed that the method we proposed was able to detect differences among models of laser printers even though using same too common fonts both Japanese and English.

show abstract

“…2 Remarkably, the computer vision research community has largely neglected the VFR problem. The few previous approaches [8,24,11,17,15,16,2] are mostly from the document analysis standpoint, focusing on small number of font classes on scanned document images that typically have high quality, in terms of resolution and contrast, and low geometric distortion. Consequently, tasks such as binarization, character segmentation, connected component analysis, geometric alignment, and OCR can be robustly applied.…”

Section: Introductionmentioning

confidence: 99%

“…Consequently, tasks such as binarization, character segmentation, connected component analysis, geometric alignment, and OCR can be robustly applied. Building on these image processing steps, global texture analysis [24,15], high-order statistical features [2] and typographical features [17,8] have been exploited for font recognition. However, such scanned image-based techniques are less applicable to our VFR problem, which needs to be effective even on very short strings or single words from noisy web images or photos taken with mobile devices.…”

Section: Introductionmentioning

confidence: 99%

Large-Scale Visual Font Recognition

Chen

Yang

Jin

et al. 2014

2014 IEEE Conference on Computer Vision and Pattern Recognition

View full text Add to dashboard Cite

This paper addresses the large-scale visual font recognition (VFR) problem, which aims at automatic identification of the typeface, weight, and slope of the text in an image or photo without any knowledge of content. Although visual font recognition has many practical applications, it has largely been neglected by the vision community. To address the VFR problem, we construct a large-scale dataset containing 2,420 font classes, which easily exceeds the scale of most image categorization datasets in computer vision. As font recognition is inherently dynamic and open-ended, i.e., new classes and data for existing categories are constantly added to the database over time, we propose a scalable solution based on the nearest class mean classifier (NCM). The core algorithm is built on local feature embedding, local feature metric learning and max-margin template selection, which is naturally amenable to NCM and thus to such open-ended classification problems. The new algorithm can generalize to new classes and new data at little added cost. Extensive experiments demonstrate that our approach is very effective on our synthetic test images, and achieves promising results on real world test images.

show abstract

A Novel Technique for English Font Recognition Using Support Vector Machines

Cited by 32 publications

References 6 publications

Robust Feature Extraction Technique for Optical Character Recognition

Robust Feature Extraction Technique for Optical Character Recognition

A New Method for Discriminating Printers Based on Contours Qualities of Printed Characters Using Wavelet Decomposition

Large-Scale Visual Font Recognition

Contact Info

Product

Resources

About