Entity

Time filter

Source Type

Beijing, China

Pan P.,Fujitsu R and nter Co. | Schonfeld D.,University of Illinois at Chicago
IEEE Signal Processing Letters | Year: 2011

In this letter, we extend the first-order Markov chain model commonly used in visual tracking and present a novel framework of visual tracking using high-order Monte Carlo Markov chain. By using graphical models to obtain conditional independence properties, we derive a general expression for the posterior density function of an m th-order hidden Markov model. We subsequently use Sequential Importance Sampling (SIS) to estimate the posterior density and obtain the high-order particle filtering algorithm for visual object tracking. Experimental results demonstrate that the performance of our proposed algorithm is superior to traditional first-order particle filtering (i.e., particle filtering derived based on first-order Markov chain). © 2006 IEEE. Source


Yin X.-C.,University of Science and Technology Beijing | Hao H.-W.,University of Science and Technology Beijing | Sun J.,Fujitsu R and nter Co. | Naoi S.,Fujitsu R and nter Co.
Proceedings of the International Conference on Document Analysis and Recognition, ICDAR | Year: 2011

Document images captured by a mobile phone camera often have perspective distortions. In this paper, fast and robust vanishing point detection methods for such perspective documents are presented. Most of previous methods are either slow or unstable. Based on robust detection of text baselines and character tilt orientations, our proposed technology is fast and robust with the following features: (1) quick detection of vanishing point candidates by clustering and voting on the Gaussian sphere space, and (2) precise and efficient detection of the final vanishing points using a hybrid approach, which combines the results from clustering and projection analysis. The rectified image acceptance rate for Mobile Cam-based documents, signboards and posters is more than 98% with an average speed of about 100ms. © 2011 IEEE. Source


Zhu Y.,Tianjin Normal University | Sun J.,Fujitsu R and nter Co. | Naoi S.,Fujitsu R and nter Co.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2012

In this paper, a natural scene character recognition method using convolutional neural network(CNN) and bimodal image enhancement is proposed. CNN based grayscale character recognizer has strong tolerance to degradations in natural scene images. Since character image is bimodal pattern image in essence, bimodal image enhancement is adopted to improve the performance of CNN classifier. Firstly, a maximum separability based color-to-gray method is used to strengthen the discriminative power in grayscale image space. Secondly, grayscale distribution normalization based on histogram alignment is performed. Through increasing the data consistency among grayscale training and test samples, it leads to a better CNN classifier. Thirdly, a shape holding grayscale character image normalization is adopted. Based on these measures, a high performance natural scene character recognizer is constructed. The recognition rate of 85.96% on ICDAR 2003 robust OCR dataset is higher than existing works, which verified the effectiveness of the proposed method. © 2012 Springer-Verlag Berlin Heidelberg. Source


Zhu Y.,Tianjin Normal University | Sun J.,Fujitsu R and nter Co. | Naoi S.,Fujitsu R and nter Co.
Proceedings of the International Conference on Document Analysis and Recognition, ICDAR | Year: 2013

This paper proposed a sub-structure learning based method for handwritten Chinese text recognition. In conventional methods, a standard character recognizer is trained on character classes only. Unreliable recognition results on character segments will decrease final recognition precision. By discovering stable sub-structure patterns from real character segment samples automatically, both character and sub-structure patterns are trained in character recognizer. The judgment reliability of segments being characters is significantly improved. Furthermore, to deal with millions of training segment samples, a two-stage clustering method is proposed for sub-structure learning. Experiment results on HIT-MW database show that the sub-structure learning based method improves performance significantly. The F1-measure evaluation of handwritten Chinese text recognition is improved by 8.84%. © 2013 IEEE. Source


Zhu Y.,Tianjin Normal University | Sun J.,Fujitsu R and nter Co. | Naoi S.,Fujitsu R and nter Co.
Proceedings - International Conference on Pattern Recognition | Year: 2012

Comparing with conventional character normalization methods not taking the discriminative information into account, this paper proposes a novel normalization method - Discriminative Normalization. Saliency regions contain most of discriminative information among similar characters. According to different types, they are enlarged in character normalization to increase their influence in recognition. As a result, discrimination power among similar characters is enhanced which is benefit to separating similar characters. The experiment on CASIA dataset shows that error rate is reduced by 9.97%. Comparing with similar character recognition without discriminative normalization, 46.0% more errors are reduced. That verifies its effectiveness. © 2012 ICPR Org Committee. Source

Discover hidden collaborations