Search results for: Fast Face Detection

Authors: Pinaki Pratim Acharjya, Esha Dutta

Abstract:

In image segmentation contour detection is one of the important pre-processing steps in recent days. Contours characterize boundaries and contour detection is one of the most difficult tasks in image processing. Hence it is a problem of fundamental importance in image processing. Contour detection of an image decreases the volume of data considerably and useless information is removed, but the structural properties of the image remain same. In this research, a robust and effective contour detection technique has been proposed using mathematical morphology. Three different contour detection results are obtained by using morphological dilation and erosion. The comparative analyses of three different results also have been done.

Keywords: Image segmentation, contour detection, mathematical morphology.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1395

Authors: Omid Gervei, Ahmad Ayatollahi, Navid Gervei

Abstract:

In this paper we present an approach for 3D face recognition based on extracting principal components of range images by utilizing modified PCA methods namely 2DPCA and bidirectional 2DPCA also known as (2D) 2 PCA.A preprocessing stage was implemented on the images to smooth them using median and Gaussian filtering. In the normalization stage we locate the nose tip to lay it at the center of images then crop each image to a standard size of 100*100. In the face recognition stage we extract the principal component of each image using both 2DPCA and (2D) 2 PCA. Finally, we use Euclidean distance to measure the minimum distance between a given test image to the training images in the database. We also compare the result of using both methods. The best result achieved by experiments on a public face database shows that 83.3 percent is the rate of face recognition for a random facial expression.

Keywords: 3D face recognition, 2DPCA, (2D) 2 PCA, Rangeimage

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 3048

Authors: Fituri H Belgassem, Abdulbasit Nigrat, Seddeeq Ghrari

Abstract:

In this paper, the construction of fast algorithms for the computation of Periodic Walsh Piecewise-Linear PWL transform and the Periodic Haar Piecewise-Linear PHL transform will be presented. Algorithms for the computation of the inverse transforms are also proposed. The matrix equation of the PWL and PHL transforms are introduced. Comparison of the computational requirements for the periodic piecewise-linear transforms and other orthogonal transforms shows that the periodic piecewise-linear transforms require less number of operations than some orthogonal transforms such as the Fourier, Walsh and the Discrete Cosine transforms.

Keywords: Piece wise linear transforms, Fast transforms, Fast algorithms.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1644

Authors: Sandip Sadhukhan, Arpita Sarkar, Debprasad Sinha, Goutam Kumar Ghorai, Gautam Sarkar, Ashis K. Dhara

Abstract:

Accurate segmentation of the optic disc is very important for computer-aided diagnosis of several ocular diseases such as glaucoma, diabetic retinopathy, and hypertensive retinopathy. The paper presents an accurate and fast optic disc detection and segmentation method using an attention based fully convolutional network. The network is trained from scratch using the fundus images of extended MESSIDOR database and the trained model is used for segmentation of optic disc. The false positives are removed based on morphological operation and shape features. The result is evaluated using three-fold cross-validation on six public fundus image databases such as DIARETDB0, DIARETDB1, DRIVE, AV-INSPIRE, CHASE DB1 and MESSIDOR. The attention based fully convolutional network is robust and effective for detection and segmentation of optic disc in the images affected by diabetic retinopathy and it outperforms existing techniques.

Keywords: Ocular diseases, retinal fundus image, optic disc detection and segmentation, fully convolutional network, overlap measure.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 753

Authors: Samar Krimi, Kaïs Ouni, Noureddine Ellouze

Abstract:

The method described in this paper deals with the problems of T-wave detection in an ECG. Determining the position of a T-wave is complicated due to the low amplitude, the ambiguous and changing form of the complex. A wavelet transform approach handles these complications therefore a method based on this concept was developed. In this way we developed a detection method that is able to detect T-waves with a sensitivity of 93% and a correct-detection ratio of 93% even with a serious amount of baseline drift and noise.

Keywords: ECG, Modulus Maxima Wavelet Transform, Performance, T-wave detection

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1837

Authors: Shobha Sharma, Amita Dev, Akanksha Kant

Abstract:

Digital signal processor, image signal processor and FIR filters have multipliers as an important part of their design. On the basis of Vedic mathematics, Vedic multipliers have come out to be very fast multipliers. One of the image processing applications is edge detection. This research presents a small area and high speed 8 bit Vedic multiplier system comprising of compressor based adders. This results in faster edge detection. This architecture is tested on Xilinx vertex 4 FPGA board and simulations were carried out using the Xilinx synthesis tool. Comparisons are made and this system is found to be smaller in area with high speed (the lesser propagation delay). This compressor based Vedic multiplier is 1.1 times speedier than a typical Vedic multiplier. Also, this Vedic Multiplier is 2 times speedier than a ‘simple’ multiplier.

Keywords: Detection of edges, Vedic multiplier, image processing, Urdhva Tiryakbhyam sutra.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1798

Authors: Yang Weiping, Zhang Zhilong, Li Jicheng, Chen Zengping, He Jun

Abstract:

The background estimation approach using a small window median filter is presented on the bases of analyzing IR point target, noise and clutter model. After simplifying the two-dimensional filter, a simple method of adopting one-dimensional median filter is illustrated to make estimations of background according to the characteristics of IR scanning system. The adaptive threshold is used to segment canceled image in the background. Experimental results show that the algorithm achieved good performance and satisfy the requirement of big size image-s real-time processing.

Keywords: Point target, background estimation, median filter, adaptive threshold, target detection.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1825

Authors: Masrour Dowlatabadi, Jalil Shirazi

Abstract:

In this paper, an improved edge detection algorithm based on fuzzy combination of mathematical morphology and wavelet transform is proposed. The combined method is proposed to overcome the limitation of wavelet based edge detection and mathematical morphology based edge detection in noisy images. Experimental results show superiority of the proposed method, as compared to the traditional Prewitt, wavelet based and morphology based edge detection methods. The proposed method is an effective edge detection method for noisy image and keeps clear and continuous edges.

Keywords: Edge detection, Wavelet transform, Mathematical morphology, Fuzzy logic.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2388

Authors: Raid R. O. Al-Nima, S. S. Dlay, W. L. Woo

Abstract:

A relationship between face and signature biometrics is established in this paper. A new approach is developed to predict faces from signatures by using artificial intelligence. A multilayer perceptron (MLP) neural network is used to generate face details from features extracted from signatures, here face is the physical biometric and signatures is the behavioural biometric. The new method establishes a relationship between the two biometrics and regenerates a visible face image from the signature features. Furthermore, the performance efficiencies of our new technique are demonstrated in terms of minimum error rates compared to published work.

Keywords: Behavioural biometric, Face biometric, Neural network, Physical biometric, Signature biometric.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1669

Authors: Deng Shiwei

Abstract:

The users are now expecting higher level of DSP(Digital Signal Processing) software quality than ever before. Prevention and detection of defect are critical elements of software quality assurance. In this paper, principles and rules for prevention and detection of defect are suggested, which are not universal guidelines, but are useful for both novice and experienced DSP software developers.

Keywords: defect detection, defect prevention, DSP-software, software development, software testing.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1786

Authors: C. P. Providakis, G. M. Angeli, M. J. Favvata, N. A. Papadopoulos, C. E. Chalioris, C. G. Karayannis

Abstract:

An effort for the detection of damages in the  reinforcement bars of reinforced concrete members using PZTs is  presented. The damage can be the result of excessive elongation of  the steel bar due to steel yielding or due to local steel corrosion. In  both cases the damage is simulated by considering reduced diameter  of the rebar along the damaged part of its length. An integration  approach based on both electromechanical admittance methodology  and guided wave propagation technique is used to evaluate the  artificial damage on the examined longitudinal steel bar. Two  actuator PZTs and a sensor PZT are considered to be bonded on the  examined steel bar. The admittance of the Sensor PZT is calculated  using COMSOL 3.4a. Fast Furrier Transformation for a better  evaluation of the results is employed. An effort for the quantification  of the damage detection using the root mean square deviation  (RMSD) between the healthy condition and damage state of the  sensor PZT is attempted. The numerical value of the RSMD yields a  level for the difference between the healthy and the damaged  admittance computation indicating this way the presence of damage  in the structure. Experimental measurements are also presented.

 

Keywords: Concrete reinforcement, damage detection, electromechanical admittance, experimental measurements, finite element method, guided waves, PZT.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2638

Authors: Zhang Xiaodong, Tang Zixin, Liu Xin

Abstract:

In order to avoid self-collision of space manipulators during operation process, a real-time detection method is proposed in this paper. The manipulator is fitted into a cylinder-enveloping surface, and then, a kind of detection algorithm of collision between cylinders is analyzed. The collision model of space manipulator self-links can be detected by using this algorithm in real-time detection during the operation process. To ensure security of the operation, a safety threshold is designed. The simulation and experiment results verify the effectiveness of the proposed algorithm for a 7-DOF space manipulator.

Keywords: Space manipulator, Collision detection, Self-collision, the real-time collision detection.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2018

Authors: Medjram Sofiane, Babahenini Mohamed Chaouki, Mohamed Benali Yamina

Abstract:

Skin detection is an important task for computer vision systems. A good method of skin detection means a good and successful result of the system. The colour is a good descriptor for image segmentation and classification; it allows detecting skin colour in the images. The lighting changes and the objects that have a colour similar than skin colour make the operation of skin detection difficult. In this paper, we proposed a method using the YCbCr colour space for skin detection and lighting effects elimination, then we use the information of texture to eliminate the false regions detected by the YCbCr skin model.

Keywords: Skin detection, YCbCr, GLCM, Texture, Human skin.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2429

Authors: Israa Al-Neami, Ali J. Al-Askery, Martin Johnston, Charalampos Tsimenidis

Abstract:

In this paper, we focus on the design of a multi-line copper wire (MLCW) communication system. First, we construct our proposed MLCW channel and verify its characteristics based on the Kolmogorov-Smirnov test. In addition, we apply Middleton class A impulsive noise (IN) to the copper channel for further investigation. Second, the MIMO G.fast system is adopted utilizing the proposed MLCW channel model and is compared to a single line G-fast system. Second, the performance of the coded system is obtained utilizing concatenated interleaved Reed-Solomon (RS) code with four-dimensional trellis-coded modulation (4D TCM), and compared to the single line G-fast system. Simulations are obtained for high quadrature amplitude modulation (QAM) constellations that are commonly used with G-fast communications, the results demonstrate that the bit error rate (BER) performance of the coded MLCW system shows an improvement compared to the single line G-fast systems.

Keywords: G.fast, Middleton Class A impulsive noise, mitigation techniques, copper channel Model.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 682

Authors: H. Krouma, M. Barkat, K. Kemih, M. Benslama, Y. Yacine

Abstract:

In this paper, we consider the analysis of the acquisition process for a hybrid double-dwell system with antenna diversity for DS-CDMA (direct sequence-code division multiple access) using an adaptive threshold. Acquisition systems with a fixed threshold value are unable to adapt to fast varying mobile communications environments and may result in a high false alarm rate, and/or low detection probability. Therefore, we propose an adaptively varying threshold scheme through the use of a cellaveraging constant false alarm rate (CA-CFAR) algorithm, which is well known in the field of radar detection. We derive exact expressions for the probabilities of detection and false alarm in Rayleigh fading channels. The mean acquisition time of the system under consideration is also derived. The performance of the system is analyzed and compared to that of a hybrid single dwell system.

Keywords: Adaptive threshold, hybrid double-dwell system, CA-CFAR algorithm, DS-CDMA.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1693

Authors: Urvee B. Trivedi, U. D. Dalal

Abstract:

As wireless communication services grow quickly; the seriousness of spectrum utilization has been on the rise gradually. An emerging technology, cognitive radio has come out to solve today’s spectrum scarcity problem. To support the spectrum reuse functionality, secondary users are required to sense the radio frequency environment, and once the primary users are found to be active, the secondary users are required to vacate the channel within a certain amount of time. Therefore, spectrum sensing is of significant importance. Once sensing is done, different prediction rules apply to classify the traffic pattern of primary user. Primary user follows two types of traffic patterns: periodic and stochastic ON-OFF patterns. A cognitive radio can learn the patterns in different channels over time. Two types of classification methods are discussed in this paper, by considering edge detection and by using autocorrelation function. Edge detection method has a high accuracy but it cannot tolerate sensing errors. Autocorrelation-based classification is applicable in the real environment as it can tolerate some amount of sensing errors.

Keywords: Cognitive radio (CR), probability of detection (PD), probability of false alarm (PF), primary User (PU), secondary user (SU), Fast Fourier transform (FFT), signal to noise ratio (SNR).

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1453

Authors: H. Özkan

Abstract:

In this study, a new and fast algorithm for Ascending Aorta (AscA) and Descending Aorta (DesA) segmentation is presented using Computed Tomography Angiography images. This process is quite important especially at the detection of aortic plaques, aneurysms, calcification or stenosis. The applied method has been carried out at four steps. At first step, lung segmentation is achieved. At the second one, Mediastinum Region (MR) is detected to use in the segmentation. At the third one, images have been applied optimal threshold and components which are outside of the MR were removed. Lastly, identifying and segmentation of AscA and DesA have been carried out. The performance of the applied method is found quite well for radiologists and it gives enough results to the surgeries medically.

Keywords: Ascending aorta (AscA), Descending aorta (DesA), Computed tomography angiography (CTA), Computer aided detection (CAD), Segmentation

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1817

Authors: Parvinder S. Sandhu, Iqbaldeep Kaur, Amit Verma, Prateek Gupta

Abstract:

The current research paper is an implementation of Eigen Faces and Karhunen-Loeve Algorithm for face recognition. The designed program works in a manner where a unique identification number is given to each face under trial. These faces are kept in a database from where any particular face can be matched and found out of the available test faces. The Karhunen –Loeve Algorithm has been implemented to find out the appropriate right face (with same features) with respect to given input image as test data image having unique identification number. The procedure involves usage of Eigen faces for the recognition of faces.

Keywords: Eigen Faces, Karhunen-Loeve Algorithm, FaceRecognition.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1724

Authors: Thomas Fuhrmann

Abstract:

In this paper an algorithm for fast wavelength calibration of Optical Spectrum Analyzers (OSAs) using low power reference gas spectra is proposed. In existing OSAs a reference spectrum with low noise for precise detection of the reference extreme values is needed. To generate this spectrum costly hardware with high optical power is necessary. With this new wavelength calibration algorithm it is possible to use a noisy reference spectrum and therefore hardware costs can be cut. With this algorithm the reference spectrum is filtered and the key information is extracted by segmenting and finding the local minima and maxima. Afterwards slope and offset of a linear correction function for best matching the measured and theoretical spectra are found by correlating the measured with the stored minima. With this algorithm a reliable wavelength referencing of an OSA can be implemented on a microcontroller with a calculation time of less than one second.

Keywords: correlation, gas reference, optical spectrum analyzer, wavelength calibration

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1400

Authors: Vincent King Soon Wong, Hong Seng Ng, Florinna Sim

Abstract:

Negative Bias Temperature Instability (NBTI) is one of the critical degradation mechanisms in semiconductor device reliability that causes shift in the threshold voltage (Vth). However, thorough understanding of this reliability failure mechanism is still unachievable due to a recovery characteristic known as NBTI recovery. This paper will demonstrate the severity of NBTI recovery as well as one of the effective methods used to mitigate, which is the minimization of measurement system delays. Comparison was done in between two measurement systems that have significant differences in measurement delays to show how NBTI recovery causes result deviations and how fast measurement systems can mitigate NBTI recovery. Another method to minimize NBTI recovery without the influence of measurement system known as Fast Wafer Level Reliability (FWLR) NBTI was also done to be used as reference.

Keywords: Fast vs slow BTI, Fast wafer level reliability, Negative bias temperature instability, NBTI measurement system, metal-oxide-semiconductor field-effect transistor, MOSFET, NBTI recovery, reliability.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1630

Authors: Mohammed Mahmood, Walid Tizani, Carlo Sansour

Abstract:

In this paper, experimental testing and numerical analysis were used to investigate the effect of tube thickness on the face bending for concrete filled hollow sections connected to other structural members using Extended Hollobolts. Six samples were tested experimentally by applying pull-out load on the bolts. These samples were designed to fail by column face bending. The main variable in all tests is the column face thickness. Finite element analyses were also performed using ABAQUS 6.11 to extend the experimental results and to quantify the effect of column face thickness. Results show that, the column face thickness has a clear impact on the connection strength and stiffness. However, the amount of improvement in the connection stiffness by changing the column face thickness from 5mm to 6.3mm seems to be higher than that when increasing it from 6.3mm to 8mm. The displacement at which the bolts start pulling-out from their holes increased with the use of thinner column face due to the high flexibility of the section. At the ultimate strength, the yielding of the column face propagated to the column corner and there was no yielding in its walls. After the ultimate resistance is reached, the propagation of the yielding was mainly in the column face with a miner yielding in the walls.

Keywords: Anchored bolted connection, Extended Hollobolt, Column faces bending and concrete filled hollow sections.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2537

Authors: Nur Syarafina Binti Othman, Tsubasa Jindo, Makato Yamada, Miho Tsuyama, Hitoshi Nakano

Abstract:

A novel method to produce a fast high voltage solid states switch using Insulated Gate Bipolar Transistors (IGBTs) is presented for discharge-pumped gas lasers. The IGBTs are connected in series to achieve a high voltage rating. An avalanche transistor is used as the gate driver. The fast pulse generated by the avalanche transistor quickly charges the large input capacitance of the IGBT, resulting in a switch out of a fast high-voltage pulse. The switching characteristic of fast-high voltage solid state switch has been estimated in the multi-stage series-connected IGBT with the applied voltage of several tens of kV. Electrical circuit diagram and the mythology of fast-high voltage solid state switch as well as experimental results obtained are presented.

Keywords: High voltage, IGBT, Solid states switch.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 5888

Authors: Sulaiman Al Amro, Ali Alkhalifah

Abstract:

The growing number of computer viruses and the detection of zero day malware have been the concern for security researchers for a large period of time. Existing antivirus products (AVs) rely on detecting virus signatures which do not provide a full solution to the problems associated with these viruses. The use of logic formulae to model the behaviour of viruses is one of the most encouraging recent developments in virus research, which provides alternatives to classic virus detection methods. In this paper, we proposed a comparative study about different virus detection techniques. This paper provides the advantages and drawbacks of different detection techniques. Different techniques will be used in this paper to provide a discussion about what technique is more effective to detect computer viruses.

Keywords: Computer viruses, virus detection, signature-based, behaviour-based, heuristic-based.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 4578

Authors: Michael Zimba

Abstract:

A duplicated image region may be subjected to a number of attacks such as noise addition, compression, reflection, rotation, and scaling with the intention of either merely mating it to its targeted neighborhood or preventing its detection. In this paper, we present an effective and robust method of detecting duplicated regions inclusive of those affected by the various attacks. In order to reduce the dimension of the image, the proposed algorithm firstly performs discrete wavelet transform, DWT, of a suspicious image. However, unlike most existing copy move image forgery (CMIF) detection algorithms operating in the DWT domain which extract only the low frequency subband of the DWT of the suspicious image thereby leaving valuable information in the other three subbands, the proposed algorithm simultaneously extracts features from all the four subbands. The extracted features are not only more accurate representation of image regions but also robust to additive noise, JPEG compression, and affine transformation. Furthermore, principal component analysis-eigenvalue decomposition, PCA-EVD, is applied to reduce the dimension of the features. The extracted features are then sorted using the more computationally efficient Radix Sort algorithm. Finally, same affine transformation selection, SATS, a duplication verification method, is applied to detect duplicated regions. The proposed algorithm is not only fast but also more robust to attacks compared to the related CMIF detection algorithms. The experimental results show high detection rates. 

Keywords: Affine Transformation, Discrete Wavelet Transform, Radix Sort, SATS.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1895

Authors: Matthew Montebello

Abstract:

The way e-learning effectiveness has been notoriously measured within an academic setting is by comparing the e-learning medium to the traditional face-to-face teaching methodology. In this paper, a simple yet innovative comparison methodology is introduced, whereby the effectiveness of next generation e-learning systems are assessed in contrast not only to the face-to-face mode, but also to the classical e-learning modality. Ethical and logistical issues are also discussed, as this three-way approach to compare teaching methodologies was applied and documented in a real empirical study within a higher education institution.

Keywords: E-learning effectiveness, higher education, teaching modality comparison.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1371

Authors: M. Ranjeeth, S. Anuradha

Abstract:

Spectrum sensing is the main feature of cognitive radio technology. Spectrum sensing gives an idea of detecting the presence of the primary users in a licensed spectrum. In this paper we compare the theoretical results of detection probability of different fading environments like Rayleigh, Rician, Nakagami-m fading channels with the simulation results using energy detection based spectrum sensing. The numerical results are plotted as Pf Vs Pd for different SNR values, fading parameters. It is observed that Nakagami fading channel performance is better than other fading channels by using energy detection in spectrum sensing. A MATLAB simulation test bench has been implemented to know the performance of energy detection in different fading channel environment.

Keywords: Spectrum sensing, Energy detection, fading channels, Probability of detection, probability of false alarm.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 3081

Authors: Rohana Musa, Yuzman Yusoff, Chia Chieu Yin, Hanif Che Lah

Abstract:

This paper presents the design and characterization of a low power Complementary Metal Oxide Semiconductor (CMOS) process sensor. The design is targeted for implementation using Silterra’s 180 nm CMOS process technology. The proposed process sensor employs a voltage threshold (V_th) extractor architecture for detection of variations in the fabrication process. The process sensor generates output voltages in the range of 401 mV (fast-fast corner) to 443 mV (slow-slow corner) at nominal condition. The power dissipation for this process sensor is 6.3 µW with a supply voltage of 1.8V with a silicon area of 190 µm X 60 µm. The preliminary result of this process sensor that was fabricated indicates a close resemblance between test and simulated results.

Keywords: CMOS Process sensor, Process, Voltage and Temperature (PVT) sensor, threshold extractor circuit, Vth extractor circuit.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 724

Authors: Svitov David, Alyamkin Sergey

Abstract:

The usage of convolutional neural networks (CNNs) in conjunction with the margin-based softmax approach demonstrates the state-of-the-art performance for the face recognition problem. Recently, lightweight neural network models trained with the margin-based softmax have been introduced for the face identification task for edge devices. In this paper, we propose a distillation method for lightweight neural network architectures that outperforms other known methods for the face recognition task on LFW, AgeDB-30 and Megaface datasets. The idea of the proposed method is to use class centers from the teacher network for the student network. Then the student network is trained to get the same angles between the class centers and face embeddings predicted by the teacher network.

Keywords: ArcFace, distillation, face recognition, margin-based softmax.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 606

Authors: M. A Anjum, M. Y. Javed, A. Basit

Abstract:

In this paper a new approach to face recognition is presented that achieves double dimension reduction making the system computationally efficient with better recognition results. In pattern recognition techniques, discriminative information of image increases with increase in resolution to a certain extent, consequently face recognition results improve with increase in face image resolution and levels off when arriving at a certain resolution level. In the proposed model of face recognition, first image decimation algorithm is applied on face image for dimension reduction to a certain resolution level which provides best recognition results. Due to better computational speed and feature extraction potential of Discrete Cosine Transform (DCT) it is applied on face image. A subset of coefficients of DCT from low to mid frequencies that represent the face adequately and provides best recognition results is retained. A trade of between decimation factor, number of DCT coefficients retained and recognition rate with minimum computation is obtained. Preprocessing of the image is carried out to increase its robustness against variations in poses and illumination level. This new model has been tested on different databases which include ORL database, Yale database and a color database. The proposed technique has performed much better compared to other techniques. The significance of the model is two fold: (1) dimension reduction up to an effective and suitable face image resolution (2) appropriate DCT coefficients are retained to achieve best recognition results with varying image poses, intensity and illumination level.

Keywords: Biometrics, DCT, Face Recognition, Feature extraction.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1479

Authors: Mustafa S. Fawad, Wang Wencheng, Wu Enhua

Abstract:

Shadows add great amount of realism to a scene and many algorithms exists to generate shadows. Recently, Shadow volumes (SVs) have made great achievements to place a valuable position in the gaming industries. Looking at this, we concentrate on simple but valuable initial partial steps for further optimization in SV generation, i.e.; model simplification and silhouette edge detection and tracking. Shadow volumes (SVs) usually takes time in generating boundary silhouettes of the object and if the object is complex then the generation of edges become much harder and slower in process. The challenge gets stiffer when real time shadow generation and rendering is demanded. We investigated a way to use the real time silhouette edge detection method, which takes the advantage of spatial and temporal coherence, and exploit the level-of-details (LOD) technique for reducing silhouette edges of the model to use the simplified version of the model for shadow generation speeding up the running time. These steps highly reduce the execution time of shadow volume generations in real-time and are easily flexible to any of the recently proposed SV techniques. Our main focus is to exploit the LOD and silhouette edge detection technique, adopting them to further enhance the shadow volume generations for real time rendering.

Keywords: LOD, perception, Shadow Volumes, SilhouetteEdge, Spatial and Temporal coherence.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1599