Multimodal Algorithm Research Articles

T58. Automatic seizure detection based on multimodal signal quantification

Introduction Quantitative analysis and automated seizure detection is able to increase efficiency of EEG review. However, acceptance of software assisted review is often low because results are inaccurate in real world patient cohorts and the reason for false detections cannot be deduced. The graphical software tool encevis visualizes detections of fast rhythmic activity and patterns defined by the ACNS critical care EEG terminology. Based on these detections as well as quantitative information of EEG, ECG, and EMG a multimodal seizure detection algorithm was developed. Simple rule based classification is utilized that facilitates easy interpretability. Aim of this work was to assess detection performance of different modalities and patient groups. Methods Our computer algorithm automatically detects seizures including rhythmic EEG patterns that show an increased amplitude compared to baseline. EMG signal is extracted by bandpass filtering EEG (30–60 Hz) to measure line length (LL) for detection of sustained and excessive ictal EMG activity of generalized tonic-clonic seizures (GTCS). High absolute values of LL and an increase of 500% to baseline trigger detections. Heart rate is calculated from ECG signals to detect ictal tachycardia (ITC) with more than 100 beats per minute and an increase of over 30% compared to baseline. To assess sensitivity and false detection rate a retrospective study was conducted including EEG/ECG recordings of 92 patients from two epilepsy monitoring units. Inclusion criteria were an age above 18 and at least one recorded epileptic seizure. EEGs were used without modification or manual editing of any kind. Automatic seizure detection was calculated for all 11,978 h of EEG (min = 23 h, max = 547 h). In total 410 manual seizure annotations were compared to automatic detections to define sensitivity (SE) and false detection in 24 h (FD/24 h). Results Combination of all three seizure detection methods (EMG + ECG + EEG) resulted in SE = 88% with 10.5 FD/24 h on average. By using only EMG based detections 100% of GTCS (n = 49) were found with an average false detection rate of 3.39 FD/24 h. Seizure detection solely based on ECG yielded SE = 31% with 1.35 FD/24 h. Analysis of the temporal lobe epilepsy patients showed SE = 93.3% and 6.75 FD/24 h, the extra temporal lobe patient group resulted in SE = 80% at 15.3 FD/24 h. Conclusion We showed that automatic seizure detection based on multimodal signal quantification can reach high sensitivity. The low false detection rate results in an average of 20 false detections per week than can be validated quickly by using EEG and time synchronized quantitative screens in parallel. By visualizing quantitative information that is the source of automatic seizure detection the interpretability of results is improved. Our proposed approach to automatic EEG analysis will raise efficiency of post hoc analysis compared to the current state of the art.

Reliability of IOTA score and ADNEX model in the screening of ovarian malignancy in postmenopausal women

BackgroundThe IOTA (International Ovarian Tumor Analysis) group has developed the ADNEX (Assessment of Different NEoplasias in the adneXa) model to predict the risk that an ovarian mass is benign, borderline or malignant. This study aimed to test reliability of these risks prediction models to improve the performance of pelvic ultrasound and discriminate between benign and malignant cysts. Material and methodsPostmenopausal women with an adnexal mass (including ovarian, para-ovarian and tubal) and who underwent a standardized ultrasound examination before surgery were included. Prospectively and retrospectively collected data and ultrasound appearances of the tumors were described using the terms and definitions of the IOTA group and tested in accordance with the ADNEX model and were compared to the final histological diagnosis. ResultsOf the 107 menopausal patients recruited between 2011 and 2016, 14 were excluded (incomplete inclusion criteria). Thus, 93 patients constituted a cohort in whom 89 had benign cysts (83 ovarian and 6 tubal or para-ovarian cysts), 1 had border line tumor and 3 had invasive ovarian cancers (1 at first stage, 1 at advanced stage and 1 metastatic tumor in the ovary). The overall prevalence of malignancy was 4.3%. Every benign ovarian cyst was classified as probably benign by IOTA score which showed also a high specificity with the totality of probably malignant lesion proved malignant by histological exam. The limit of this score was the important rate of not classified or undetermined cysts. However, the malignancy risks calculated by ADNEX model allowed identifying the totality of malignancy. Thus, the combination of the two methods of analysis showed a sensitivity and specificity rates of respectively 100% and 98%. Evaluation of malignancy risks by these 2 tests highlighted a negative predictive value of 100% (there was no case of false negative) and a positive predictive value of 80%. Discussion and conclusionOn the basis of our findings, the IOTA classification and the ADNEX multimodal algorithm used as risks prediction models can improve the performance of pelvic ultrasound and discriminate between benign and malignant cysts in postmenopausal women, especially for undetermined lesions.

Detecting overlapping communities in LBSNs by fuzzy subtractive clustering

With the increasing popularity of location-based social networks (LBSNs), community detection has emerged as an important and practical issue. One of the main shortcomings of the previous methods is that cluster’s centers have been selected randomly in clustering the communities; therefore, different results are obtained in each execution. This paper proposes an intelligent method to detect overlapping communities in LBSNs with respect to user check-ins and the attributes of venues and users. In the proposed approach, clustering is performed on the two-segment user-venue edges. The advantage of this approach is the use of subtractive clustering method, which determines cluster centers with a probability, based on the potential of each data point. Furthermore, the specified fuzzy rules have a significant role in determining the appropriate number of cluster centers. The paper concentrates on the data obtained from the Foursquare LBSN. The empirical results of the proposed approach are compared with the previous approach to clustering edges in multi-attribute and multi-mode (M2-clustering) algorithm, and a significant improvement is observed in the cost function of the community detection.

Phantom Preparation and Optical Property Determination

Tissue-like optical phantoms are important in testing new imaging algorithms. Homogeneous optical phantoms with determined optical properties are the first step of making a proper heterogeneous phantom for multi-modality imaging. Typical recipes for such phantoms consist of epoxy resin, hardener, India ink and titanium oxide. By altering the concentration of India ink and titanium oxide, we are able to get multiple homogeneous phantoms with different absorption and scattering coefficients by carefully mixing all the ingredients. After fabricating the phantoms, we need to find their individual optical properties including the absorption and scattering coefficients. This is achieved by solving diffusion equation of each phantom as a homogeneous slab under canonical illumination. We solve the diffusion equation of homogeneous slab in frequency domain and get the formula for theoretical measurements. Under our steady-state diffused optical tomography (DOT) imaging system, we are able to obtain the real distribution of the incident light produced by a laser. With this source distribution we got and the formula we derived, numerical experiments show how measurements change while varying the value of absorption and scattering coefficients. Then we notice that the measurements alone will not be enough for us to get unique optical properties for steady-state DOT problem. Thus in order to determine the optical properties of a homogeneous slab we want to fix one of the coefficients first and use optimization methods to find another one. Then by assemble multiple homogeneous slab phantoms with different optical properties, we are able to obtain a heterogeneous phantom suitable for testing multi-modality imaging algorithms. In this paper, we describe how to make phantoms, derive a formula to solve the diffusion equation, demonstrate the non-uniqueness of steady-state DOT problem by analysing some numerical results of our formula, and finally propose a possible way to determine optical properties for homogeneous slab for our future work.

MMSI: A Multi-Mode Service Invocation Algorithm To Improve The Connectivity In Accessing Cloud Services In Heterogeneous Mobile Cloud

Modern research in cloud environment is focused on where mobile users can access data through cloud services through arrangement of regional cloudlets when the connectivity with the cloud service provider is less or lost. When a mobile user is connected to a cloud service and the service level lowers on random mobility, this algorithm is used for choosing the cloudlet or an ad hoc cloud to provide identical service without any interruption. The client invokes the services through cellular networks in normal condition and every time interval, it computes the signal strength, energy level and delay factors in accessing the cloud service. When the estimated parameters are less than the threshold, it is connected with the local access point. The multi-mode algorithm computes the service invocation weight and selects the connectivity mode in continuing the service invocation; this helps the moving user to access cloud services without any interruption.

Abstract IA14: Optimizing head and neck cancer treatment outcomes with MRI

Abstract Changes in the epidemiology of head and neck cancers have resulted in an increasing number of younger and healthier patients being treated with definitive external beam radiotherapy (EBRT). The long-term consequences of radiotherapy in a patient population with good clinical outcomes and extended life expectancy are becoming increasingly relevant in the management of treatment-associated morbidity and mortality. Osteoradionecrosis (ORN) of the mandible is a challenging issue related to irradiation, occurring in up to 16% of patients with various types of head and neck cancers. Altered bone vascularity and opportunistic infections within the oral cavity contribute to the development of ORN, leading to an inexorable process of bone destruction that does not follow the normal sequence of healing events. Early-stage ORN is often managed using antibiotics, local wound care, and hyperbaric oxygen (HBO). Advanced ORN requires surgical resection and reconstruction with healthy nonirradiated tissue. Successful management of this disease process requires an enhanced ability to identify patients at risk for ORN, monitor the effectiveness of conservative management, and improve preoperative planning to ensure clear margins at the time of resection. However, a standardized, objective staging and monitoring system for ORN is not currently available. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is a clinically available quantitative imaging method that is increasingly employed to assess microvascular function in the study of solid tumors of the head and neck. At our institution, DCE-MRI is integrated into a multimodality clinical algorithm aimed at improving the diagnosis, staging, and oncologic surveillance of head and neck tumors. DCE-MRI can detect altered bone vascularity associated with bone healing, necrosis, and metastatic involvement, with excellent spatial resolution. We hypothesize that DCE-MRI can be used to detect alterations in bone vascularity following irradiation to monitor ORN clinical progression and response to treatment. We are evaluating the potential of DCE-MRI to identify patients at risk for mandibular ORN, monitor response to conservative management, and determine the extent of advanced mandibular ORN to assist in surgical planning. This work has the potential to revolutionize the diagnosis and management of mandibular ORN. For the first time, clinicians will be able to identify patients at risk for ORN and manage post-radiotherapy care appropriately. The effectiveness of currently employed conservative measures could be tested using an objective measure and improved preoperative planning could reduce the rate of surgical failure due to residual compromised bone. Citation Format: Vlad C. Sandulache, Abdallah S.R. Mohamed, Yunyun Chen, Katherine A. Hutcheson, Brian P. Hobbs, Mark S. Chambers, Jason M. Johnson, Michelle D. Williams, James A. Bankson, Clifton D. Fuller, Stephen Y. Lai. Optimizing head and neck cancer treatment outcomes with MRI [abstract]. In: Proceedings of the AACR-AHNS Head and Neck Cancer Conference: Optimizing Survival and Quality of Life through Basic, Clinical, and Translational Research; April 23-25, 2017; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2017;23(23_Suppl):Abstract nr IA14.

Development of a transportability evaluation system using swept path analysis and multi-body dynamic simulation

In this study, we develop a transportability evaluation system for the transportation process or design process in product development engineering. The system extracts the constraint condition and transportability using swept path analysis and multi-body dynamic simulation. This system consists of two main software modules: a feature curve extraction module and a multi-body dynamic simulation module. The feature curve extraction module extracts the feature curve list of the entire transportation path with swept path analysis to transmit the information to the simulation module. The multi-body dynamic simulation module extracts the constraint condition for the product design process that a designer uses as the transportation constraint in the design process or quick checks the transportability of the product files. The entire structure of the system is accessible by a web-based platform. When the user (designer) inputs the product files and the transportation information, the system gives the constraints and transportability to the user. The entire analysis is performed by a background process on the analysis server. We also propose a multimodal transportability evaluation algorithm that considers design and dynamic conditions.

Attention driven multi-modal similarity learning

To learn a function for measuring similarity or relevance between objects is an important machine learning task, referred to as similarity learning. Conventional methods are usually insufficient for processing complex patterns, while more sophisticated methods produce results supported by parameters and mathematical operations that are hard to interpret. To improve both model robustness and interpretability, we propose a novel attention driven multi-modal algorithm, which learns a distributed similarity score over different relation modalities and develops an interaction-oriented dynamic attention mechanism to selectively focus on salient patches of objects of interest. Neural networks are used to generate a set of high-level representation vectors for both the entire object and its segmented patches. Multi-view local neighboring structures between objects are encoded in the high-level object representation through an unsupervised pre-training procedure. By initializing the relation embeddings with object cluster centers, each relation modality can be reasonably interpreted as a semantic topic. A layer-wise training scheme based on a mixture of unsupervised and supervised training is proposed to improve generalization. The effectiveness of the proposed method and its superior performance compared against state-of-the-art algorithms are demonstrated through evaluations based on different image retrieval tasks.

Managing hemorrhaging hepatic tumors: the place of microwave ablation in a multimodality algorithm

Managing hemorrhaging hepatic tumors: the place of microwave ablation in a multimodality algorithm

Adaptive Multimodal Continuous Ant Colony Optimization

Seeking multiple optima simultaneously, which multimodal optimization aims at, has attracted increasing attention but remains challenging. Taking advantage of ant colony optimization (ACO) algorithms in preserving high diversity, this paper intends to extend ACO algorithms to deal with multimodal optimization. First, combined with current niching methods, an adaptive multimodal continuous ACO algorithm is introduced. In this algorithm, an adaptive parameter adjustment is developed, which takes the difference among niches into consideration. Second, to accelerate convergence, a differential evolution mutation operator is alternatively utilized to build base vectors for ants to construct new solutions. Then, to enhance the exploitation, a local search scheme based on Gaussian distribution is self-adaptively performed around the seeds of niches. Together, the proposed algorithm affords a good balance between exploration and exploitation. Extensive experiments on 20 widely used benchmark multimodal functions are conducted to investigate the influence of each algorithmic component and results are compared with several state-of-the-art multimodal algorithms and winners of competitions on multimodal optimization. These comparisons demonstrate the competitive efficiency and effectiveness of the proposed algorithm, especially in dealing with complex problems with high numbers of local optima.

Multimode Controller Design for DSTATCOM integrated with Battery Energy Storage for Smart Grid Applications

A Distribution Static Compensator (DSTATCOM) can provide reactive power support, power factor correction and voltage regulation to a distribution system feeder. In this paper, a multimode multifunctional control algorithm is proposed for the controller of the Voltage Source Converter (VSC) to provide DSTATCOM and load leveling functionalities for smart grid applications through a Battery Energy Storage System (BESS) connected to the grid through the VSC. The designed controller enables the VSC to function in both grid-connected and standalone modes of operation. In grid-connected mode, the VSC provides load leveling, dynamic reactive power exchange and unity power factor functionalities based on consumer's choice. When a fault occurs on the grid, the controller opens the circuit breaker and isolates the load from the grid and then switches to the voltage controller for the standalone mode of operation. The novelty of the proposed controller is, the inverter structure which provides load leveling and DSTATCOM functionalities during grid-connected mode will also serve the load during a grid failure, thereby providing uninterrupted power supply to the load. The simulation results show the ability of the proposed multimode controller to successfully operate the multifunctional DSTATCOM-BESS system in various modes.

Multimodal Estimation of Distribution Algorithms.

Taking the advantage of estimation of distribution algorithms (EDAs) in preserving high diversity, this paper proposes a multimodal EDA. Integrated with clustering strategies for crowding and speciation, two versions of this algorithm are developed, which operate at the niche level. Then these two algorithms are equipped with three distinctive techniques: 1) a dynamic cluster sizing strategy; 2) an alternative utilization of Gaussian and Cauchy distributions to generate offspring; and 3) an adaptive local search. The dynamic cluster sizing affords a potential balance between exploration and exploitation and reduces the sensitivity to the cluster size in the niching methods. Taking advantages of Gaussian and Cauchy distributions, we generate the offspring at the niche level through alternatively using these two distributions. Such utilization can also potentially offer a balance between exploration and exploitation. Further, solution accuracy is enhanced through a new local search scheme probabilistically conducted around seeds of niches with probabilities determined self-adaptively according to fitness values of these seeds. Extensive experiments conducted on 20 benchmark multimodal problems confirm that both algorithms can achieve competitive performance compared with several state-of-the-art multimodal algorithms, which is supported by nonparametric tests. Especially, the proposed algorithms are very promising for complex problems with many local optima.

Combined Raman and autofluorescence ex vivo diagnostics of skin cancer in near-infrared and visible regions.

The differentiation of skin melanomas and basal cell carcinomas (BCCs) was demonstrated based on combined analysis of Raman and autofluorescence spectra stimulated by visible and NIR lasers. It was ex vivo tested on 39 melanomas and 40 BCCs. Six spectroscopic criteria utilizing information about alteration of melanin, porphyrins, flavins, lipids, and collagen content in tumor with a comparison to healthy skin were proposed. The measured correlation between the proposed criteria makes it possible to define weakly correlated criteria groups for discriminant analysis and principal components analysis application. It was shown that the accuracy of cancerous tissues classification reaches 97.3% for a combined 6-criteria multimodal algorithm, while the accuracy determined separately for each modality does not exceed 79%. The combined 6-D method is a rapid and reliable tool for malignant skin detection and classification.

Multimodal ultrasonographic algorithm in the differentiation of submandibular masses

Conclusion: The multimodal diagnostic algorithm is a simple diagnostic tool in the pre-operative assessment of submandibular masses that enables one-stage surgical concepts.Background: The pre-operative assessment of submandibular masses is of major clinical impact. This study assesses the diagnostic utility of a newly introduced diagnostic algorithm in the differentiation of submandibular masses based on epidemiological and B-mode ultrasonographic data.Methods: One hundred and eighty-eight patients with submandibular triangle masses were included. Epidemiological and ultrasonographic data of 128 consecutively included patients were assessed and structured in a multimodal algorithm. The diagnostic algorithm was prospectively validated in a further 60 patients.Results: Single epidemiological and ultrasonographic data do not reliably predict the lesional entity. The multimodal diagnostic algorithm achieved a diagnostic sensitivity/specificity of 100%/100% in sialolithiasis, 44%/100% in benign submandibular gland diseases (BSD), 94%/96% in nodal lymphoma/unspecific lymphadenitis, and 91%/84% in carcinomas. Reduced sensitivity in BSD or specificity in carcinomas increased after intra-operative fresh frozen section in patients suspected for BSD.

Volterra series based blind equalization for nonlinear distortions in short reach optical CAP system

In this paper, we propose a blind Volterra series based nonlinear equalization (VNLE) with low complexity for the nonlinear distortion mitigation in short reach optical carrierless amplitude and phase (CAP) modulation system. The principle of the blind VNLE is presented and the performance of its blind adaptive algorithms including the modified cascaded multi-mode algorithm (MCMMA) and direct detection LMS (DD-LMS) are investigated experimentally. Compared to the conventional VNLE using training symbols before demodulation, it is performed after matched filtering and downsampling, so shorter memory length is required but similar performance improvement is observed. About 1dB improvement is observed at BER of 3.8×10−3 for 40Gb/s CAP32 signal over 40km standard single mode fiber.

Discriminating between intentional and unintentional gaze fixation using multimodal-based fuzzy logic algorithm for gaze tracking system with NIR camera sensor

Gaze tracking systems are widely used in human–computer interfaces, interfaces for the disabled, game interfaces, and for controlling home appliances. Most studies on gaze detection have focused on enhancing its accuracy, whereas few have considered the discrimination of intentional gaze fixation (looking at a target to activate or select it) from unintentional fixation while using gaze detection systems. Previous research methods based on the use of a keyboard or mouse button, eye blinking, and the dwell time of gaze position have various limitations. Therefore, we propose a method for discriminating between intentional and unintentional gaze fixation using a multimodal fuzzy logic algorithm applied to a gaze tracking system with a near-infrared camera sensor. Experimental results show that the proposed method outperforms the conventional method for determining gaze fixation.

Multimode Image Compression Algorithm Employing Multiple-Choice Knapsack Problem-Based Encoding Mode Selection

In this paper, we first formulate the encoding mode selection of a multimode image compression algorithm, which adaptively determines the encoding mode to encode a patch of a given image as a multiple-choice knapsack problem (MCKP). Then, we present a multimode image compression algorithm that takes advantage of the MCKP-formulated mode selection problem and a memory-efficient implementation of the MCKP-based mode selection algorithm. To emphasize the effect of the MCKP-based mode selection, the proposed multimode image compression algorithm adopts encoding modes from conventional multimode image compression algorithms. In experiments using the Kodak test image set, the proposed algorithm outperformed benchmark algorithms by 2.6–7.8 dB in the average peak signal-to-noise ratio when the target compression ratio is 1/6.

Automated multimodal segmentation of an abnormal breast mass in mammogram.

A lot of computer-aided diagnosis systems have been attempted to segment automatically breast mass from a mammogram and to classify the mass as benign and malignant quantitatively. This study aimed to develop an automated computer-aided diagnosis system to evaluate the disease with high accuracy using the proposed multimodal segmentation algorithm when compared to an abnormal breast mass region outlined in mammogram by radiologists of American College of Radiology as "standard." In this study, a total number of 150 mammograms were downloaded from the DDSM database for screening mammography. Based on the available diagnostic report, the studied data were classified as follows: (1) Group I: normal (n = 50, mean ± SD age = 55 ± 8 years), (2) Group II: benign breast cancer (n = 50, mean ± SD age = 58 ± 11 years), and (3) Group III: malignant breast cancer (n = 50, mean ± SD age = 58±9 years). It was found that the proposed multimodal segmentation algorithm processed all the mammograms of different mass types, density, shapes, size, margin, calcification type, and distortion successfully, and it segmented the mass automatically with high accuracy. In this study, a computer-aided diagnosis system was developed to segment the breast mass automatically in a mammogram with high accuracy of 96%. The sensitivity and specificity of the system were found to be 94% and 97%, respectively, when compared to abnormal region outlined in mammogram by radiologists of American College of Radiology as standard.

Ptychographic iterative engine with partially coherent illumination for weakly scattering samples

As an expansion of coherent diffraction imaging, ptychographic iterative engine (PIE) not only inherits advantages such as ultra-high resolution and compact optical system, but also expands the field of view in quantitative imaging, thus PIE is widely used in short wavelength imaging such as X-ray and electron beam imaging, and then extended to visible light field. However, PIE requires coherent illumination for both phase and amplitude retrieval, while traditional X-ray or electron beam sources often cannot satisfy this strict coherent condition, which leads to poor-quality information retrieval with low signal-to-noise ratio. Though several proposed methods such as multiple wavelength and multi-mode algorithms can eliminate incoherency influence to some extent, various details such as quantitative spectrum of illuminating source should be obtained before information retrieval, which complicates computing procedures. In addition, it is hard to acquire the spectrum of the illuminating source in most cases. In order to acquire high-quality information based on PIE with partially coherent illumination, a newly designed enhanced phase retrieval method for weakly scattering samples in PIE with partially coherent illumination is presented in this paper, in which only the bright field of the diffraction patterns is used in the iterative procedures mimicking the coherent cases especially for weakly scattering samples without any prior illuminating details. The bright field area can be regarded as purely coherent diffraction patterns composed of a strengthened zeroth order beam and a weakened diffracted beam. While the dark-field area generated by interference of diffracted beams cannot satisfy the requirement for coherence, therefore, dark-field diffraction patterns should be excluded in sample information extraction and only the bright field is used for phase retrieval via iterative process. Compared with the proposed multiple wavelength and multi-mode algorithms, the proposed method can simplify sample reconstructing procedures due to needing no prior knowledge. Moreover, in order to enhance the information of weakly scattering samples in retrieval, high order iteration method is also introduced in the paper. Since the bright field can be regarded as purely coherent diffraction patterns composed of a strengthened zeroth order beam and a weakened diffracted beam. For weakly scattering sample, the weakened diffracted beam is much lower than zeroth order beam, thus it is difficult to acquire high-contrast information with classical PIE algorithms. Introducing high order iterative tactic, the contrast of weakly scattering sample is obviously improved and the details of weakly scattering sample can be retrieved clearly. Both theoretical analysis and numerical simulations are illustrated in detail, proving the robustness and availability of the designed method: high-contrast phase information can be obtained with the proposed method, while traditional phase retrieval algorithm almost loses all details of the sample. In order to mimic the real experimental situation, a 30 dB white noise is added into the simulation, the details of weakly scattering sample phase information can also be retrieved clearly by using the bright field of the diffraction patterns with high order iteration method. With the newly designed enhanced phase retrieval method for weakly scattering samples with partially coherent illumination, sample retrieval via PIE can not only use ordinary X-ray source or electron beam as illumination source, thereby avoiding the dependence on complete coherent source, but also obviously improve the retrieval quality of the sample characteristics, which widely expands the application fields of the PIE.

Neural Network Based Normalized Fusion Approaches for Optimized Multimodal Biometric Authentication Algorithm

A multimodal biometric system is applied to recognize individuals for authentication using neural networks. In this paper multimodal biometric algorithm is designed by integrating iris, finger vein, palm print and face biometric traits. Normalized score level fusion approach is applied and optimized, encoded for matching decision. It is a multilevel wavelet, phase based fusion algorithm. This robust multimodal biometric algorithm increases the security level, accuracy, reduces memory size and equal error rate and eliminates unimodal biometric algorithm vulnerabilities.