Abstract
Latent fingerprint identification is such a difficult task to law enforcement agencies and border security in identifying suspects. It is a too complicate due to poor quality images with non-linear distortion and complex background noise. Hence, the image quality is required for matching those latent fingerprints. The current researchers have been working based on minutiae points for fingerprint matching because of their accuracy are acceptable. In an effort to extend technology for fingerprint matching, our model is to propose the enhancementand matching for latent fingerprints using Scale Invariant Feature Transformation (SIFT). It has involved in two phases (i) Latent fingerprint contrast enhancement using intuitionistic type-2 fuzzy set (ii) Extract the SIFTfeature points from the latent fingerprints. Then thematching algorithm is performedwith n- number of images and scoresare calculated by Euclidean distance. We tested our algorithm for matching, usinga public domain fingerprint database such as FVC-2004 and IIIT-latent fingerprint. The experimental consequences indicatethe matching result is obtained satisfactory compare than minutiae points.
Similar content being viewed by others
Explore related subjects
Discover the latest articles, news and stories from top researchers in related subjects.Avoid common mistakes on your manuscript.
1 Introduction
To distinguish the human biometric security is known by two ways such as physical attributes and behavior. The biometric traits, for example: fingerprint, palm veins, iris recognition, retina, face recognition, DNA, palm print, hand geometry, odor, typing rhythm, gait, and voice which are exceptionally suitable for human acknowledgment because of their singularity, integrality and consistency [18]. Fingerprints have the elevated amount of dependability and widely utilized by criminological specialists as a part of criminal examinations. Fingerprints are undeniable at around 7 months of hatchling progression and finger ridge design does not change for the term of the life of a human except for in view of the setback, for instance, wounds and cuts on the finger. Despite for twins, it will never hint at change [3].
Fingerprint patterns are shaped in the epidermis on the fingertip. The fingerprint features such as incipient ridges, minutiae, pores, singular points, ridge orientation map, singular points, dots and frequency map are used for recognition to identify the person. The minutiae are called in the form of ridge ending and bifurcation. Fingerprint image has contained three types of patterns like Arch, Loop and Whorl and have nine- types of classifications were explained by Jain et al. [10]. Fingerprints have been generally utilized for dependable human ID as a part of crime scene investigation and law authorization applications for past decades. Rolled and plain fingerprint formats are used to collect the ten-print records of each and every caught criminal in law enforcement. It has captured by rolling a finger from nail to nail and pushing down a finger on the level of surface respectively [32,33,34,35,36].
These formats have an enough ridge details for individualization and it has termed as a model fingerprints [37,38,39,40]. To guarantee great nature of fingerprint images, reference fingerprints are gained under the supervision of a human administrator; along these lines, fingerprints can be recovered on account of low-quality impressions [25, 26]. Another sort of essential fingerprint recognizable proof includes seeking latent’s fingerprints against reference fingerprint databases. Latent fingerprints used to capture the fingerprint at crime scenes and are viewed as a critical origin of confirmation in crime scene examinations to recognize suspects.
Latent fingerprint images are less in quality because can present uncertainties, as the ridge and valley structure. So, it is required to enhance the original image to obtain a better quality (Fig. 1). The major idea of image enhancement is to create another image in such a way that, it uncovers data for investigation more than the first image.
However, feature extraction and matching has been extensively difficult task for latent fingerprints because of poor quality with complex background noise. The SIFT was introduced by Lowe [16]. It has been used for extracting thedistinctive invariant features from images and reliable matching between distinct images of the same object. The SIFT algorithm has been playing a major role in many active areas, namely, computer vision, pattern recognition, fuzzy and intuitionistic fuzzy, machine learning, etc [41,42,43,44].
In this paper, a new method is presented using intuitionistic type-2 fuzzy set theory to increase the contrast enhancement of latent fingerprint images. It is compared with different method to show the efficient of the proposed method and the results are found to be quit better. Next,our implementation is focus to develop the SIFT points in latent fingerprint images (Template and query image) for better matching results.
The remaining work has composed as given by: Section 1 demonstrate a short narrative of background study of this work. Section 1.2 displays the notion of intuitionistic type-2 fuzzy set theory. Section 1.3 describes proposed methodology include contrast enhancement and SIFT feature based matching for latent fingerprints. Section 1.8 displays the experimental results and for finishing the conclusion in Section 2.
1.1 Related works
The fingerprint image enhancement is carried by Mao et al. [19] with the help of Gabor filter. A robust against gradient deviation technique provides us to estimate the orientation and frequencies of fingerprint in a local region. It allows effective Gabor filtering for fingerprint ridge and valley pattern enhancement [12].Type-2 fuzzy set was used by Chaira [7] for medical image enhancement.
The integration of Anisotropic Filter (AF) and Directional Median Filter (DMF) techniques are used for fingerprint image enhancement which is discussed by Wu et al. [27]. Bansal et al. [4] has described the fingerprints image enhancement using type-2 fuzzy set. The non stationary directional Fourier domain filtering enhancement technique is introduced by Sherlock et al. [23]. Greenberg et al. [8] developed the different types of filters for fingerprint image enhancement such as local histogram equalization, Wiener filtering, image binarization and a unique anisotropic filter. Selvi and George [22] explained the fuzzy based fingerprints enhancement technique, which was used to reduce the noise free image from the original image for further processing. Jayaram et al. [11] proposed a fuzzy inference system based contrast enhancement of gray level images. Data embedding in digital images using critical functions were introduced by Liao et al. [14, 15].
Yoon et al. [29] proposed an algorithm for latent fingerprint image enhancement which needs minimum margin in ROI and singular points to promote the automatic matching accuracy [33, 45, 46]. Later, they have planned to use manually marked ROI and singular points for latent fingerprint enhancement. Further in Yoon et al. [30] they have introduced the algorithm for Fingerprint Image Quality (LFIQ), which is used to latent fingerprint exemplar and crime investigation. Cao et al. [6] have approached a dictionary-based method for automatic latent segmentation and enhancement in the direction of the objective of accomplishing “lights-out” latent ID frameworks. Liao et al. [14, 15] have obtained Medical JPEG image steganography based on preserving inter-block dependencies [47,48,49].
Arora et al. [1] derived the algorithm for rolled and plain latent fingerprints. It performed to refine feature extraction in latent fingerprints with the eventual goal of improving the latent matching accuracy. Paulino et al. [21] used the Descriptor-Based Hough Transform (DBHT) for latent fingerprints matching whichwas based on minutiae points. Jain and Feng [9] have proposed the plain or rolled fingerprints matching using minutiae points and also extended feature including singularity, ridge quality map, ridge flow map, ridge wavelength map, and skeleton.
Partial fingerprint matching was proposed by Malathi and Meena [17] on the basis of score level fusion with the help of pore and SIFT features. SIFT feature was used by Park et al. [20] for fingerprint verification. On the basis of scale space and texture information, SIFT feature was extracted and matching was performed using the SIFT operator. To develop the matching accuracy and reduce the redundant error matching points were implemented by Yang et al. [28] for fingerprint matching algorithm using SIFT feature. Skrypnyk and Lowe [24] explores the effectiveness of the SIFT for image matching which is a generally used interest point detection method often applied to image matching the Harris corner detector. Distinct invariant feature was extracted by Skrypnyk and Lowe, 2004 for performing the reliable matching between different views of an object or scene [41, 42, 50, 51].
1.2 Preliminaries of fuzzy set
In 1965, Zadeh [31] introduced the notion of fuzzy set theory. It deals with degree of membership function. Later in 1985, Atanassov [2] defined the perception of intuitionistic fuzzy set theory which deals with degree of membership and non-membership function. He discussed the degree of non-membership function is equal to 1- degree of membership function, which is not always true, but there may be some hesitation degree that occur in the membership function. Likewise, it begins a major role in image processing, for example, image enhancement, morphological image, edge detection and so on.
The following definitions are explained below
-
Definition Typically, the fuzzy set theory indicates the membership functions which contain the closed interval range [0, 1]. The degree of membership function denoted as a fuzzy set A by μA. It can be represented as follows A = {(x, μA(x), νA(x))| x ∈ X} where μA(x) : X → [0, 1].
-
Definition Intuitionistic fuzzy set theory was introduced by Atanassov which includes two functions such as degree of membership and non-membership functions respectively. The degree of non-membership function denoted as an intuitionistic fuzzy set A by νA. It can be mathematically represented as follows A = {(x, μA(x), νA(x))| x ∈ X}; where μA(x), νA(x) : X → [0, 1]. πA(x) = 1 − μA(x) − νA(x) is the representation of the third parameter which is called the hesitation degree.
-
Definition Type-2 fuzzy set theory is utilized to find the accurate membership values where the membership values are hard. It contains more uncertainty account of membership [13]. It can be written as follows,
\( {A}_{Type-2}=\left\{x,{\overset{\frown }{\mu}}_A(x)|x\in X\right\} \); Where \( {\overset{\frown }{\mu}}_A(x) \) type-2 membership function. It includes the upper and lower of membership functions.
\( {\mu}^{lower}={\left[\mu (x)\right]}^{\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$\alpha $}\right.} \); Where α ∈ [0, 1].
The membership function on an element x ∈ X with the necessary condition
1.3 Proposed methodology
In this segment, we examine the latent fingerprint image pre-processing, enhancement, SIFT point’s extraction and matching.
1.4 Pre-processing
The Binarization and Thinning are the two procedure of pre-processing. It is used to upgrade a few elements in image for further handling. Binarization technique is used to convert a grayscale image into binary image which means every pixel contain single bit, i.e. 0 or 1 (Fig. 2a). Ridge Thinning is utilized to decrease the darkness of ridge lines (Fig. 2a). This process may suitable for recognition because enhancement is essential. Thinning is the transformation of a digital image into a simplified image but does not change over the original (x, y) area. It is utilized to decimate the additional pixel of ridges until just one pixel broad.
1.5 Contrast enhancement using intuitionistic type −2 fuzzy set
The feature of the image must be adequately high to guarantee high fingerprint recognition precision. Latent fingerprint ridge flow enhancement is an exceptionally vital and essential procedure before feature extraction.
This enhancement method is based on window scheme. The original image is split into six windows and each window maintained the same size for further processing. Enhancement is obtained from each window.
The following formula is used to fuzzified the N × N image
Where, g - gray level of the image ranges from 0 to L-1. gmax and gmin - max – min values of the image.
The membership and non-membership function is calculated for every window using Takagi-Sugeno-Kang (TSK) type intuitionistic fuzzy set (IFS) generator.
The TSK type IFS generator is followed as [5]:
Where K(1) = 0, K(0) = 1.
The non-membership function is derived by using Sugeno type intuitionistic fuzzy generator
The TSK-IFS become as:
For all windows, the hesitation degree can be mathematically written as follows,
In this experiment λ = 1 is utilized to maintain the quality of the image because on raising the λ value, the enhanced image will be worsening. At the same time, as λ increases, the Sugeno generator will diminish in this manner the non-membership quality will diminish and the hesitation degree will increment.
The average value of the enhanced features is considered in each window. Due to hesitation degree occur in interval range of membership function, the modified membership function can represent as follows:
Here, we introduce the type-2 fuzzy membership function with based on definition 3.3 in section 3. Based on the image quality the value of α is obtained by using trial and error method.
The upper-lower functions are considered based on Hamacher T-co norm [13]
μupper(gij) and μlower(gij) are the upper - lower membership function of the type-2 fuzzy set respectively which are processed utilizing definition 3.3
In this case of λ is taken as λ = 10 × im_avg, where im_avg is the average value of the pixels in the image. Form this experiment results, the value ‘10’ would be the better and results are achieved by separating the images into several windows.
1.6 SIFT feature point’s extraction
Generally object recognition was developed on the basis of SIFT feature. Translation, rotation, scale and other imaging parameters are the local invariant feature coordinates which are transformed from the image content (Fig. 3). Scale-space extrema detection, key point localization, orientation assignment and key point descriptor are important stages of computation used for generating the SIFT features. These stages are performed as follows: pick the key locations at local maximum and minimum of a Difference of Gaussian (DOG) function applied in scale space, which is constructed by successively down sampling the input image. Maximum and minimum of this scale space function are determined by comparing each pixel to its neighbors.
1.7 Matching
In the past decades, Automated Fingerprint Identification System (AFIS) has not been achieved successfully in latent fingerprints. The latent fingerprint matchingis a more challenging for all researchers by reason of it is not a quality image. In this algorithm achieved enhancement and matching using SIFT points to find the exact matching (Fig. 4).
For example figure:
The SIFT feature point matching is achieved by comparing each local extrema based on the associated descriptors. We consider I1 and I2 are two images for matching. Given a feature point p11 in I1, its closest point p21, second closest point p22, and their distances d1 and d2 are calculated from feature points in I2. When the ratio d1/d2 is sufficiently small, p11 is considered to match with p21. The number of matching points and their geometric configuration is decided the matching score between two images.
The algorithm is executed as taken after:
-
Step 1: Latent fingerprint images is an introduced as a input image
-
Step 2: Acquire the latent fingerprint image.
-
Step 3: Pre-processing the input image.
-
Step 4: Latent fingerprint image contrast enhancement using intuitionistic type-2 fuzzy set
-
Step 5: Extract the SIFT points from the image
-
Step 6: Extract the feature (SIFT) for query image which is taken from database
-
Step 7: Compute the Euclidean distance between query and database image.
-
Step 8: Sort the output image and discover which one is flawless match to the given query image.
Mathematically, the Euclidean distance is the ‘ordinary’ distance between two points in Euclidean space.
After the completion of matching process, the matched image will be displayed on the screen.
1.8 Experimental results and discussions
Fingerprint Verification Competition-2004 (FVC-2004) and Indraprastha Institute of Information Technology (IIIT) latent fingerprint database have been used for experiments. It has been public available fingerprint databases.
FVC-2004 consists of four databases and each database image has different size and resolution (http://bias.csr.unibo.it/fvc2004/download.asp). In particularly FVC-2004, we are used FVC-2004 database 1 and FVC-2004 database 2. IIIT latent fingerprint database contains 15 subjects (http://www.iab-rubric.org/resources.html). Each subject has all 10 fingerprints. Each image is of 4752 × 3168 pixel in size and has been scanned at 500 pixels per inch as a gray - scale image. Due to large pixel size of latent fingerprint images, while computation which gives the unacceptable computation error. Therefore, the ROI is manually marked and it will be changed as a size of 508 × 661 for better results (Fig. 5). Therefore, the results are obtained based on manually marked ROI image. It is only applicable for latent fingerprint images but not in fingerprint images. The enhancement and matching are mainly focused on latent fingerprint images. In addition, we are tested above mentioned ordinary fingerprint databases, so which are clear image compare than latent fingerprints.
1.9 Latent fingerprint image contrast enhancement
The latent fingerprint image quality evaluation is very difficult because it contains a lot of uncertainty and poorly illuminated. The evaluation uses the linear index of fuzziness / fuzzy entropy for calculating the fuzziness in enhanced image. The linear index of fuzziness is less than original image and this is the feature of enhancement. Proposed method is tested with linear index of fuzziness and it is better than the existing method (Figs. 6, 7, 8, and Table 1).
The linear index of fuzziness as follows:
The graph shows the proposed method of fuzziness is low compared than previous method such as crisp set, fuzzy set, intuitionistic fuzzy set and type-2 fuzzy set (Fig. 9).
1.10 SIFT feature points matching
This algorithm is used to extract the SIFT points from the image. Next we find the distance between two pixel (database image and query image) by using Euclidean formula. Then the matching scores are calculated with the help of proposed method and it has tested with 10 images from each database but it’s applicable for n-number of images. Furthermore, if a query image matches with any one of the template images, which match score is equal to 0 (exact match). Otherwise similarity match score will be shown.
Typically, the small number of minutiae pointsare appearing in a fingerprint images compare than SIFT points but in this experiment results are showed that the matching scores is high compared than SIFT points (See, Table 2, 3, and 4). In general, the less number of feature points are enough to show the efficient of matching but we obtain too many minutiae match score in each database which is lead to significant computational complexity. It’s observed that the pre-processing including enhancement step is elevated to the effect of matching using SIFT points (Fig. 10).
2 Conclusions
In this paper, SIFT operator can be used for fingerprint feature extraction and matching with proper pre-processing step. It gives better performance on public domain databases. Minutiae based technique uses preprocessing to connect the broken ridges. Further it removes all the texture information while extracting the skeletons of the ridge pattern. But, texture information is required for the SIFT operator. Therefore, this algorithm overcomes the drawbacks of conventional approaches to latent fingerprints due to a proper preprocessing step is used which would requires reducing the noise. MATLAB is used to implement this technique. The fingerprint recognizable proof is one of the procedures utilized in measurable science to help criminal examinations in day by day life, giving access control in budgetary security, visa related administrations etc. Our future work, will concentrate to use the large scale fingerprint database for identification using fusion (Minutiae and singular points) technique.
Change history
13 September 2022
This article has been retracted. Please see the Retraction Notice for more detail: https://doi.org/10.1007/s11042-022-13882-0
References
Arora S, Liu E, Cao K, Jain AK (2014) Latent fingerprint matching: performance gain via feedback from exemplar prints. IEEE Trans Pattern Anal Mach Intell 36(12):2452–2465
Atanassov KT (1986) Intuitionistic fuzzy sets. Fuzzy Sets Syst 20(1):87–96
Babler WJ (1991) Embryologic development of epidermal ridges and their configurations. Birth Defects Orig Artic Ser 27:95–112
Bansal R, Arora P, Gaur M, Sehgal P, Bedi P (2009) Fingerprint image enhancement using type-2 fuzzy sets. Proceedings of the IEEE sixth international conference on fuzzy systems and knowledge discovery. 3:412–417. Tianjin
Bustince H, Kacprzyk J, Mohedano V (2000) Intuitionistic fuzzy generators application to intuitionistic fuzzy complementation. Fuzzy Sets Syst 114(3):485–504
Cao K, Liu E, Jain AK (2014) Segmentation and enhancement of latent fingerprints: a coarse to fine ridge structure dictionary. IEEE Trans Pattern Anal Mach Intell 36(9):1847–1859
Chaira T (2013) Contrast enhancement of medical images using type II fuzzy set. Proceedings of the IEEE national conference on communications. 1–5. India
Greenberg S, Aladjem M, Kogan D, Dimitrov I (2000) Fingerprint image enhancement using filtering techniques. Proceedings of the 15th international conference on pattern recognition. 3, 322–325. Barcelona
Jain AK, Feng J (2011) Latent fingerprint matching. IEEE Trans Pattern Anal Mach Intell 33(1):88–100
Jain AK, Flynn P, Ross AA (2007) Handbook of biometrics. Springer Science and Business Media, New York
Jayaram B, Narayana K, Vetrivel V (2011) Fuzzy inference system based contrast enhancement. Proceedings of the international conference on EUSFLAT-LFA. 311–318. France
Karimi AS, Kuo CC (2008) A robust technique for latent fingerprint image segmentation and enhancement. Proceeding of the IEEE international conference on image processing. 1492–1495. Canada
Kumar, P. M., Gandhi, U., Varatharajan, R., Manogaran, G., Jidhesh, R., & Vadivel, T. (2017). Intelligent face recognition and navigation system using neural learning for smart security in internet of things. Clust Comput 1–12. doi:https://doi.org/10.1007/s10586-017-1323-4
Lee KH (2006) First course on fuzzy theory and applications. Springer Science and Business Media, Germany
Liao X, Qin Z, Ding L (2017a) Dataembedding in digital images using critical functions. Signal Process Image Commun 58:146–156
Liao X, Yin J, Guo S, Li X, Sangaiah AK (2017b) Medical JPEG image steganography based onpreserving inter-block dependencies. Comput Electr Eng. https://doi.org/10.1016/j.compel-eceng.2017.08.020
Lopez, D. and Gunasekaran, M. (2015). Assessment of vaccination strategies using fuzzy multicriteriadecision making. In Proc. Proceedings of the Fifth International Conference on Fuzzy and NeuroComputing (FANCCO-2015). Springer International, Cham, pp. 195–208
Lopez, D., & Manogaran, G. (2016). Big data architecture for climate change and disease dynamics. Geetam S. Tomar (eds.) et al. The human element of big data: issues, analytics, and performance. CRC Press, Florida
Lopez, D., & Manogaran, G. (2017). Parametric model to predict H1N1 influenza in vellore district, Tamil Nadu, India. In handbook of statistics, vol. 37. Elsevier, Tamil Nadu, pp. 301–316
Lopez D, Sekaran G (2016) Climate change and disease dynamics - a big data perspective. Int J Infect Dis 45:23–24
Lopez, D., Gunasekaran, M., Murugan, B. S., Kaur, H., and Abbas, K. M. (2014, October). Spatial big data analytics of influenza epidemic in Vellore, India. In Proc. 2014 IEEE International Conference onBig Data. IEEE, pp. 19–24. doi: https://doi.org/10.1109/BigData.2014.7004422
Lopez D, Manogaran G, Jagan J (2017) Modelling the H1N1 influenza using mathematical and neural network approaches. Biomed Res 28(8):1–5
Lowe DG (1999) Object recognition from local scale-invariant features. Proceedings of the seventh IEEE international conference on computer vision. 2:1150–1157. Kerkyra
Malathi S, Meena C (2011) Improved partial fingerprint matching based on score level fusion using pore and sift features. Proceeding of the IEEE International conference on process automation control and computing. 1–4. Coimbatore
Maltoni D, Maio D, Prabhakar S, Jain AK (2009) Handbook of fingerprint recognition. Springer Science and Business Media, London
Manogaran G, Lopez D (2016) Health data analytics using scalable logistic regression with stochastic gradient descent. Int J Adv Intell Paradig 9:1–15
Manogaran G, Lopez D (2017) Disease surveillance system for big climate data processing and dengue transmission. Int J Ambient Comput Intell 8(2):1–25
Manogaran G, Lopez D (2017) Spatial cumulative sum algorithm with big data analytics for climate change detection. Comput Electr Eng. https://doi.org/10.1016/j.compeleceng.2017.04.006
Manogaran, G., & Lopez, D. (2017). A Gaussian process based big data processing framework in cluster computing environment. Clust Comput 1–16. doi:https://doi.org/10.1007/s10586-017-0982-5
Manogaran G, Lopez D (2017) A survey of big data architectures and machine learning algorithms in healthcare. Int J Biomed Eng Technol 25(2–4):182–211
Manogaran, G., Varatharajan, R., & Priyan, M. K. (2017). Hybrid recommendation system for heart disease diagnosis based on multiple kernel learning with adaptive neuro-fuzzy inference system. Multimed Tools Appl 1–21. doi:https://doi.org/10.1007/s11042-017-5515-y
Manogaran G, Varatharajan R, Lopez D, Kumar PM, Sundarasekar R, Thota C (2017) A new architecture of Internet of Things and big data ecosystem for secured smart healthcare monitoring and alerting. Futur Gener Comput Syst. https://doi.org/10.1016/j.future.2017.10.045
Manogaran, G., Lopez, D., Thota, C., Abbas, K. M., Pyne, S., & Sundarasekar, R. (2017a). Big data analytics in healthcare internet of things. In Innovative healthcare systems for the 21st century. Springer International Publishing, Berlin, pp. 263–284
Manogaran, G., Thota, C., Lopez, D., Vijayakumar, V., Abbas, K. M., & Sundarsekar, R. (2017b). Big data knowledge system in healthcare. In Internet of things and big data technologies for next generation healthcare. Springer International Publishing, Berlin, p. 133–157
Manogaran, G., Thota, C., & Lopez, D. (2018). Human-computer interaction with big data analytics. In HCI challenges and privacy preservation in big data security. IGI Global, India, pp. 1–22
Manogaran, G., Vijayakumar, V., Varatharajan, R., Kumar, P. M., Sundarasekar, R., & Hsu, C. H. Machine learning based big data processing framework for cancer diagnosis using hidden markov model and GM clustering. Wirel Pers Commun, 1–18. doi:https://doi.org/10.1007/s11277-017-5044-z
Mao K, Zhu Z, Jiang H (2010) A fast fingerprint image enhancement method. Proceedings of the IEEE third international joint conference on computational science and optimization. 1, 222–226. China
Park U, Pankanti S, Jain AK (2008) Fingerprint verification using SIFT features. Proceedings of the international society for optics and photonics in SPIE defense and security symposium. 69440K–69440K. Orlando
Paulino AA, Feng J, Jain AK (2013) Latent fingerprint matching using descriptor-based houghtransform. IEEE Trans Inf Forensics Secur 8(1):31–45
Selvi M, George A (2013) FBFET: fuzzy based fingerprint enhancement technique based on adaptive thresholding. Proceedings of the IEEE fourth international conference on computing. Communications and networking technologies. 1–5. Tiruchengode
Sherlock BG, Monro DM, Millard K (1994) Fingerprint enhancement by directional Fourier filtering. IET proceedings vision. Image Signal Process 141(2):87–94
Skrypnyk I, Lowe DG (2004) Scene modeling, recognition and tracking with invariant image features. Proceeding of the third IEEE and ACM international symposium on mixed and augmented reality. 110–119. USA
Thota, C., Sundarasekar, R., Manogaran, G., Varatharajan, R., & Priyan, M. K. (2018). Centralized fog computing security platform for IoT and cloud in healthcare system. In Exploring the convergence of big data and the internet of things. IGI Global, Hershey, pp. 141–154
Varatharajan, R., Manogaran, G., & Priyan, M. K. (2017). A big data classification approach using LDA with an enhanced SVM method for ECG signals in cloud computing. Multimed Tools Appl 1–21. doi:https://doi.org/10.1007/s11042-017-5318-1
Varatharajan R, Manogaran G, Priyan MK, Balaş VE, Barna C (2017a) Visual analysis of geospatial habitat suitability model based on inverse distance weighting with paired comparison analysis. Multimed Tools Appl 1–21. doi: https://doi.org/10.1007/s11042-017-4768-9
Varatharajan R, Vasanth K, Gunasekaran M, Priyan M, Gao XZ (2017b) An adaptive decision based kriging interpolation algorithm for the removal of high density salt and pepper noise in images. Comput Electr Eng. https://doi.org/10.1016/j.compeleceng.2017.05.035
Wu C, Shi Z, Govindaraju V (2004) Fingerprint image enhancement method using directional median filter. Proceedings of the International society for optics and photonics in SPIE Defense and security symposium. 66–75. Orlando
Yang Y, Liu W, Zhang L (2010) Study on improved scale invariant feature transform matching algorithm. Proceeding of the second pacific-asia conference on circuits. communications and system. 1:398–401. China
Yoon S, Feng J, Jain AK (2011) Latent fingerprint enhancement via robust orientation field estimation. Proceeding of the IEEE international joint conference on biometrics. 1–8. Washington
Yoon S, Cao K, Liu E, Jain AK (2013) LFIQ: latent fingerprint image quality. Proceeding of the IEEE sixth international conference on theory. Applications and systems. 1–8. Arlington
Zadeh LA (1965) Fuzzy sets. Inf Control 8(3):338–353
Author information
Authors and Affiliations
Corresponding author
About this article
Cite this article
Manickam, A., Devarasan, E., Manogaran, G. et al. RETRACTED ARTICLE: Score level based latent fingerprint enhancement and matching using SIFT feature. Multimed Tools Appl 78, 3065–3085 (2019). https://doi.org/10.1007/s11042-018-5633-1
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11042-018-5633-1