Takenobu Matsuura

On-line writer recognition for Thai based on velocity of barycenter of pen-point movement

We propose an on-line writer recognition method for Thai based on velocity of barycenter of pen-point movement. The barycenter is determined from the center point of script and two adjacent pen-point positions with respect to time in handwriting process. In this paper, the Fourier coefficients or the velocity and trajectory of the barycenter are considered as the input and output of FIR (finite impulse response) system, respectively. In this case, the impulse response of the FIR system is interpreted as the feature of handwriting. The K-L expansion of the impulse response is used for writer recognition. Writer recognition experiments are performed on database consisting of 6,642 Thai scripts written by 81 writers. As the result. Type I (false rejection) and Type II (false acceptance) error rates were 1.50% and 0.05%. respectively.

A simple technique to determine calibration parameters for coplanar camera calibration

Coplanar camera calibration is the process of determining the extrinsic and intrinsic parameters of camera, where a set of image and two-dimensional (2D) control points are known. This paper presents a simple and efficient technique to compute the calibration parameters for coplanar calibration. Computing of the calibration parameters is directly performed on computer image array in the frame buffer. Thereby, the image center can simultaneously computed together with other parameters. However, the parameters can be solved using available three orthogonality constraints, where the scale factor is assumed as one and the lens distortion is not considered. The proposed algorithm is tested with noise-free synthetic image and synthetic image corrupted with known noise. The results show that the proposed technique is accurate and robust to noise.

FIR signature verification system characterizing dynamics of handwriting features

This paper proposes an online signature verification method based on the finite impulse response (FIR) system characterizing time-frequency characteristics of dynamic handwriting features. First, the barycenter determined from both the center point of signature and two adjacent pen-point positions in the signing process, instead of one pen-point position, is used to reduce the fluctuation of handwriting motion. In this paper, among the available dynamic handwriting features, motion pressure and area pressure are employed to investigate handwriting behavior. Thus, the stable dynamic handwriting features can be described by the relation of the time-frequency characteristics of the dynamic handwriting features. In this study, the aforesaid relation can be represented by the FIR system with the wavelet coefficients of the dynamic handwriting features as both input and output of the system. The impulse response of the FIR system is used as the individual feature for a particular signature. In short, the signature can be verified by evaluating the difference between the impulse responses of the FIR systems for a reference signature and the signature to be verified. The signature verification experiments in this paper were conducted using the SUBCORPUS MCYT-100 signature database consisting of 5,000 signatures from 100 signers. The proposed method yielded equal error rate (EER) of 3.21% on skilled forgeries.

On-line signature verification based on FIR system characterizing velocity and direction change of barycenter trajectory

We propose an on-line signature verification method based on finite impulse response(FIR) system characterizing velocity and direction change of barycenter trajectory. First, the discrete cosine transforms (DCTs) of the characteristics are used to reduce fluctuation and extract the feature of handwriting in signing process. Then the signature verification system is realized by the three FIR subsystems. The obtained impulse responses of the three FIR subsystems are used as the individual feature for signature verification. Signature can be verified by evaluating the difference between the impulse responses of the FIR subsystems for a reference signature and the signature to be verified. The signature verification experiments were performed on the SUBCORPUS-100 MCYT signature database[6] consisting of 5,000 signatures from 100 signers.

Seal imprint verification with rotation invariance

This work discusses on a seal imprint verification method with rotation invariance. The rotation invariant feature is represented here by the coefficients of 2D Fourier series expansion of the log-polar image. Seal imprint can be verified by the eigen vector corresponding to the largest eigen value of a matrix defined in terms of the above Fourier coefficients. The seal imprint verification experiments are given to show the effectiveness of the proposed method.

Firearm identification based on rotation invariant feature of cartridge case

This paper, proposes firearm identification method based on rotation invariant feature of cartridge case image. The rotation invariance feature is represented by the absolute value of Fourier coefficients of polar image of cartridge case on circles with different radii. The absolute value of Fourier coefficients is considered as the individual feature of the particular gun. Finally, the firearm can be identified by the distance between the absolute value of Fourier coefficients obtained from the reference cartridge case and the cartridge case to be identified. The firearm identification experiments are given to show the effectiveness of the proposed method.

On-line writer identification method based on FIR system characterizing pen-tip movement

In this paper on-line writer identification problem is reduced to FIR system identification problem. Then person can be identified by evaluating the difference between the impulse responses of the FIR system for the reference and the person to be identified. The FIR system is realized by considering the DCTs of the horizontal and vertical components of the pen-tip movement with reduced fluctuation as the input and output, respectively. Finally, writer identification experiments are given to show the effectiveness of the proposed method.

An Algorithm for Coplanar Camera Calibration

This paper proposes an algorithm to determine the camera parameters for coplanar camera calibration. The proposed method requires only one image to compute the camera parameters. The intermediate parameters defined in terms of the camera parameters are determined by a linear method. Then, the final intermediate parameters and the lens distortion coefficients are computed iteratively using a nonlinear optimization method in which the parameters obtained from the linear method are used as initial parameters. Finally, the camera parameters are determined from the final intermediate parameters. The proposed method is tested and compared with the conventional method using the synthetic images. The experimental results show that the camera parameters determined by the proposed method give accurate and stable.

Online Unconstrained Handwritten Thai Character Recognition Using Multiple Representations

This paper proposes an unconstrained handwritten Thai character recognition method using multiple representations. The proposed method can recognize different handwritten Thai characters having small curve segments with clockwise and counter clockwise directions, although the conventional method such as the elastic matching method was difficult to recognize them. As the experimental results, the recognition rate was 97.50% and the recognition time (average) using the PC with 1.4 GHz Pentium 4 was 0.01 second.

Determine Calibration Parameters with Satisfied Constraints for Coplanar Camera Calibration

Coplanar camera calibration is the process of determining a set of calibration parameters that describe the mapping between 2D control points and 2D image coordinates. However, the optimal solution for coplanar calibration can be obtained by solving a set of constrained optimization problem. This paper presents a technique to determine the calibration parameters for coplanar calibration, where a set of orthogonality constraints is fully satisfied. The solving of camera calibration consists of two steps. The first step is an estimating of calibration parameters using a closed form solution based on distortion free. In the second step, the parameters estimated in first step are iteratively improved by nonlinear optimization method, where the objective function is enforced to satisfy the constraints during optimization process. The computing of parameters is directly performed on computer image array by assuming that scale factor is equal to 1. Consequently, the image center can be simultaneously computed with the other parameters. The proposed technique is tested with a synthetic image that known distortion. The result shows that the proposed technique gives sufficient accuracy.