PPT On A Transform (MRT) Naturally Suited to Directional Pattern Analysis
Download
A Transform (MRT) Naturally Suited to Directional Pattern Analysis Presentation Transcript:
1.A Transform (MRT) Naturally Suited to Directional Pattern Analysis
2.Introduction
Digital Signal Processing
Process sequence of numbers
Signal Transforms
Fourier transform, wavelet transform
Discrete Fourier Transform (DFT)
3.Global Patterns used in fingerprint analysis
4.Experimental Result
The black strips indicate the dominant direction of each 16x16 block estimated by using the proposed method.
5.Experimental Result
The black strips indicate the dominant direction of each 16x16 block estimated by using the proposed method.
6.Conclusion
MRT is a simple, alternate method of signal representation. A few MRT coefficients represent highly directional patterns and these can be used to identify directional features in images.
In the Fourier domain, the power spectrum of a directional pattern clusters along a line through the origin, and the orientation of the pattern is perpendicular to this line. The orientation information is thus spread over different frequency indices. In contrast, the MRT method associates each frequency index directly with a specific orientation.
A highly-directional subset of the MRT coefficients of each block of a fingerprint image was used to obtain the orientation field estimate of fingerprint images. Experimental results show that this procedure provides a good estimate of the orientation field.
The directional feature of MRT and its pattern structure has potential applications in areas which require a simple tool to describe directionality of images.
Download
A Transform (MRT) Naturally Suited to Directional Pattern Analysis Presentation Transcript:
1.A Transform (MRT) Naturally Suited to Directional Pattern Analysis
2.Introduction
Digital Signal Processing
Process sequence of numbers
Signal Transforms
Fourier transform, wavelet transform
Discrete Fourier Transform (DFT)
3.Global Patterns used in fingerprint analysis
4.Experimental Result
The black strips indicate the dominant direction of each 16x16 block estimated by using the proposed method.
5.Experimental Result
The black strips indicate the dominant direction of each 16x16 block estimated by using the proposed method.
6.Conclusion
MRT is a simple, alternate method of signal representation. A few MRT coefficients represent highly directional patterns and these can be used to identify directional features in images.
In the Fourier domain, the power spectrum of a directional pattern clusters along a line through the origin, and the orientation of the pattern is perpendicular to this line. The orientation information is thus spread over different frequency indices. In contrast, the MRT method associates each frequency index directly with a specific orientation.
A highly-directional subset of the MRT coefficients of each block of a fingerprint image was used to obtain the orientation field estimate of fingerprint images. Experimental results show that this procedure provides a good estimate of the orientation field.
The directional feature of MRT and its pattern structure has potential applications in areas which require a simple tool to describe directionality of images.
No comments:
Post a Comment