New Wavelet and Color Features for Text Detection in Video

Palaiahnakote, Shivakumara; Phan, Trung Quy; Tan, Chew Lim

doi:10.1109/icpr.2010.972

Cited by 32 publications

(24 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For each patch, retrieve the distinct properties that can separate text regions from background ones. Such properties include distribution of wavelet coefficients [12,13], DCT coefficients [14], edge feature [15], spatial variance [16]. Then use some classifier, such as support vector machine (SVM) [17], neural network [18], and Adaboost [19], to classify text patches and non-text ones.…”

Section: Relation To Prior Workmentioning

confidence: 99%

“…What is more, they are unable to detect slanted texts effectively. Region-based methods first group the pixels which share the same properties, such as the edges' geometrical arrangement [20,21] and color uniformity [12,22]. Then for each group of pixels, some geometrical constraints and other text features are used to remove non-text groups.…”

Section: Relation To Prior Workmentioning

confidence: 99%

See 1 more Smart Citation

Transform invariant text extraction

et al. 2013

View full text Add to dashboard Cite

Automatically extracting texts from natural images is very useful for many applications such as augmented reality. Most of the existing text detection systems require that the texts to be detected (and recognized) in an image are taken from a nearly frontal viewpoint. However, texts in most images taken naturally by a camera or a mobile phone can have a significant affine or perspective deformation, making the existing text detection and the subsequent OCR engines prone to failures. In this paper, based on stroke width transform and texture invariant low-rank transform, we propose a framework that can detect and rectify texts in arbitrary orientations in the image against complex backgrounds, so that the texts can be correctly recognized by common OCR engines. Extensive experiments show the advantage of our method when compared to the state of art text detection systems.

show abstract

Section: Relation To Prior Workmentioning

confidence: 99%

Section: Relation To Prior Workmentioning

confidence: 99%

Transform invariant text extraction

et al. 2013

View full text Add to dashboard Cite

show abstract

“…For instance, the examples for text detection before and after text candidate region detection are respectively shown in Fig. 1, where one can see the text detection method which works based on wavelet and color features [8] fails to detect all the text lines in the image and gives more false positives as shown in Fig. 1(a) when text candidate region detection is not applied, while the same text detection method detects all the text lines with less false positives when applied after text candidate region detection as shown in Fig.…”

Section: Introductionmentioning

confidence: 99%

A New Multi-modal Technique for Bib Number/Text Detection in Natural Images

Roy

Shivakumara

Mondal

et al. 2015

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. The detection and recognition of racing bib number/text, which is printed on paper, cardboard tag, or t-shirt in natural images in marathon, race and sports, is challenging due to person movement, non-rigid surface, distortion by non-illumination, severe occlusions, orientation variations etc. In this paper, we present a multi-modal technique that combines both biometric and textual features to achieve good results for bib number/text detection. We explore face and skin features in a new way for identifying text candidate regions from input natural images. For each text candidate region, we propose to use text detection and recognition methods for detecting and recognizing bib numbers/texts, respectively. To validate the usefulness of the proposed multi-modal technique, we conduct text detection and recognition experiments before text candidate region detection and after text candidate region detection in terms of recall, precision and f-measure. Experimental results show that the proposed multi-modal technique outperforms the existing bib number detection method.

show abstract

“…In the experiment, the performance was evaluated objectively by measures of precision, recall, f-measure, and misdetection rates (Shivakumara, Phan, & Tan, 2010). The scenes were classified into the following categories by our detection method.…”

Section: Learning Activitymentioning

confidence: 99%

A rapid auto-indexing technology for designing readable e-learning content

Yu¹,

Liao²,

Su³

et al. 2012

IRRODL

View full text Add to dashboard Cite

A rapid scene indexing method is proposed to improve retrieval performance for students accessing instructional videos. This indexing method is applied to anchor suitable indices to the instructional video so that students can obtain several small lesson units to gain learning mastery. The method also regulates online course progress. These anchored points not only provide students with fast access to specific material but also can link to certain quizzes or problems to show the interactive e-learning content that course developers deposited in the learning management system, which enhances the learning process. This allows students to click on the anchored point to repeat their lesson, or work through the quizzes or problems until they reach formative assessment. Hence, their learning can be guided by the formative assessment results.In order to quickly find the scene to index, some specific description of it was needed. Actually, most of the instructional videos were recorded by teachers and were part of their PowerPoint presentations. Based on the features of the PowerPoint slides in the instructional videos, such as the title or page number, the specified scene can be found. Since we used specific scene descriptions, it was easy to employ the rapid scene detection method using an image filter and Sobel mask. Finally, we applied an experimental design to check the precision of the scene detection and evaluate user satisfaction. The results showed that rapid scene indexing can definitely assist learners in their online learning; that is, it gives them better learning mastery and provides regulation for the online learning environment.

show abstract

New Wavelet and Color Features for Text Detection in Video

Cited by 32 publications

References 11 publications

Transform invariant text extraction

Transform invariant text extraction

A New Multi-modal Technique for Bib Number/Text Detection in Natural Images

A rapid auto-indexing technology for designing readable e-learning content

Contact Info

Product

Resources

About