Thus, fingerprint recognition is one of the security enforcements and essentially had to be developed. As in Chapter 1, this paper discusses the background to GLCM and DWT and the reason this project was undertaken. Apart from that, this paper also discusses the problem that had been faced earlier to recognize fingerprints optimally.
This paper also discusses the objectives and limitations of this project in this chapter. In the next chapter, the story about GLCM as well as DWT was discussed at length that made it. Definition of term, equation and equation related to GLCM and DWT was also explained.
In the third chapter, this paper reviews the method that will be approached for the project for the entire eight-month time frame. As for the last chapter, several initial conclusions were drawn about the techniques for fingerprint recognition.
INTRODUCTION
- Background of Study
- History
- Problem Statement
- Objectives
- Scope of Study
- Relevancy of Study
- Feasibility of Study
Nowadays, the technology has developed a new approach that corresponds to the identity management as well as access control regarding the fingerprint or thumbprint identification or rather, recognition. During the early 20th century, several conventional scientists such as Henry Faulds, Francis Galton as well as Edward Henry started to develop the fingerprint recognition approach for the knowledge development intention. Among the early development are murders, crimes and perpetrators identification establishment through the use of the fingerprint recognition [5].
Nevertheless, by the end of the 20th century, the largest fingerprint recognition system had emerged, developed by the Integrated Automated Fingerprint Identification System (IAFIS). The information collected is included in the demographic statistics and is supplemented with an index of 10 fingerprints [5]. Gray-level co-occurrence matrices (GLCM) have been around for almost forty years and are still widely used.
In this work, this paper will investigate a fingerprint recognition system that merged two feature extraction techniques, namely Gray-Level Co-occurrence Matrices (GLCM) as well as Discrete Wavelet Transform (DWT). The final step is to evaluate the system performance in terms of correct detection.
LITERATURE REVIEW
Definition
GLCM
In addition to that, GLCM has also been used as discrete Fourier transform normalization to convert rotation-dependent features into rotation-invariant ones and tested on four different datasets of natural and synthetic images. The goal can be achieved by considering all pixels located approximately at a given distance from it, extracting rotation-dependent features for each direction defined by the neighborhood and converting the rotation-dependent features to rotation-independent ones" (Francesco Bianconi. Many quantitative measures of texture exist and use 3D co-occurrence matrices in CBIR applications.
The aim of the related papers is to generalize the concept of adjacency matrices to dimensional Euclidean spaces and to extract more features from the matrix.
DWT
Since the perfect reconstruction is a sum of inverse low-pass and inverse high-pass filters, the output of the inverse high-pass filter can be calculated. The first phase involves understanding the computation involved in a multi-expansion wavelet transform and determining the best layout for the SPROC chip, a digital signal processing chip that uses parallel processing and pipelining for efficiency. The SPROC chip is essentially a RISC processor with an instruction set geared towards DSP applications [15].
Although it seemed fairly certain that the final version of the wavelet transformer would be a grating filter, matrix methods were studied to obtain a basic understanding of wavelets, the results of which are presented in the discussion of Chapter 4.
METHODOLOGY
Project Activities
The total value and the division of the GLCM numerical values are performed after each value of the normalized GLCM in the cell content is multiplied by a factor [6]. The equation will be used primarily to calculate the pixel weight in the captured image. There is no contrast created in the diagonal GLCM table, but the contrast will increase as the value moves away from the diagonal, which is also affected by increasing the factor.
The chi-square dissimilarity equation (5) will then be used to track the dissimilarity between the original database fingerprint as well as the captured fingerprint image. This will dilate a digital signal and each pixel of the image will be dilated by a decimator [15]. This data is passed through two convolution functions, each of which creates an output stream that is half the length of the original input.
While in the case of a grating filter, the low- and high-pass outputs are usually called odd and even outputs, respectively [16]. The event or high-pass output contains the difference between the true input and the value of the reconstructed input if it were reconstructed only from the information given in the odd output. However, each output has half the frequency band of the input, so the frequency resolution has been doubled.
For the first block, it indicates that the entire fingerprint will be stored in a database in a. Next block, it indicates that the entire fingerprint will be converted to gray in color so that the size of the fingerprint database will be much smaller. This process also ensures that the line of the finger or finger ridge pattern can be more easily traced and scanned.
Thirdly, Contrast, Correlation, Energy and Homogeneity of the fingerprint will be calculated and compared with each other. And finally, the entire data set will be compared with each other to check the dissimilarity of the fingerprint. Then the image of the fingerprint will be filtered as high pass and low pass to calculate for the next process.
Project Timeline
Project Key-Milestone
SUMMARY OF PROJECT PROGRESS AND FUTURE WORK
Expected Result
Compared to the GLCM technique, DWT is more sensitive and can detect larger differences between the dataset. The higher the contrast value, the more the conversion process to gray color needs to be done. But the value of inequality in experiment 3 and experiment 2 is almost the same and must be rounded up to 4 decimal places.
Future Works
CONCLUSION AND RECOMMENDATIONS
Normalization equation
This contrast group is specified to measure the related weight or factor contrast against the distance from the GLCM diagonal. The contrast becomes zero value if the integer channel is set with either 8-bit channel or 16-bit channel, so it must be introduced with only real numbers.
Contrast equation
Dissimilarity equation
Homogeneity equation
Dissimilarity Chi-squared equation
Low Pass Filter Equation
High Pass Filter Equation
Then, the received fingerprint will be compared with the set threshold and later it will be decided whether it can be accepted or not.
