INTRODUCTION situations concerning security and surveillance like

INTRODUCTION The advancementof Technology in the 21st Century makes a common intent of everyaction to be based on real time analysis and impeccable accuracy. In case ofImage retrieval or video-text detection and recognition, it poses a matter ofeven deeper insight while dealing with situations concerning security andsurveillance like directing Blind people on road or retrieving thealpha-numeric from number plate of a Car. The Quality of the videos are significantly degraded owing to factorslike motion blur, non- uniform illumination, complex background and textmovement.

Other than that the variance of lightings and perspective distortionshas a major negative impact on texts as well. So it remains a question of anattainable accuracy among the researchers. Text being one of the prime channelof communication, the retrieval of textual information from scene images andvideos frames has gained the attention of the researchers from the verybeginning. A robust text detection framework is required to detect text informationfrom a scene images and video frames.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now

Many theories and applications proposed,the accuracy of text detection from a video frame still remains a challenge asthe inputs are unrestricted to colors, fonts and even orientation.The methodologyof text detection can be resolved into namely three advances: connectedcomponent–bases, edge based and texture based. The first method uses theprinciple of color quantization and splitting to connect or group the adjacentpixel of similar color into connected components. But the factors like colorbleeding and the low contrast of text lines defy in manufacturing the completeimage and hence is not applicable for video images. In order to overcome thisimpending problem, edge based techniques are proposed. In this methodology boththe horizontal as well as the vertical profile of the edge map are analyzed.

Thoughexpedited, this process of text detection in video frames has the probabilityof producing false positive at large in instances of complex background. Inorder fix such problems, texture based approach is applied which considers thetext regions as texture or couture. Feature extractions in this texture based approachis done using Fast Fourier Transform, discrete cosine transform, waveletdecomposition, and Gabor filters. And such method involves classifiers like SVMand neural networks.

 But theseclassifiers needs extensive training data of text and non-text for attaininghigher accuracy.                Detectionof text with high precision on multi-oriented dimensions without restriction onthe background, alignment, and contrast and re-call still remains a difficulttask. Existing models focused on horizontal text–orientation fails when appliedon multi-oriented text frames.

Successful research on this field is muchlimited due to above mentioned restrictions. Accordingly, in this paper, wehave deduced a methodical model which will handle linear text in multiple orientationas well as curved texts. In addition to which we have proposed a HMM basedverification to attain higher accuracy.                Theproposed text detection framework which involves both linear text and curvetexts in any frames is developed from the very base to an advanced infrastructure.

For better filtering, Laplacian of Gaussian filter is used in the proposedmethod. One of the novelties of this method belongs to the domain of skeletal featuringof the texts which may reside in nonlinear dimensions and HMM basedverification of the text resulting in attaining a optimum level accuracy.    


I'm Gerard!

Would you like to get a custom essay? How about receiving a customized one?

Check it out