Ashwin Kothari

3-D face recognition: features, databases, algorithms and challenges

Face recognition is being widely accepted as a biometric technique because of its non-intrusive nature. Despite extensive research on 2-D face recognition, it suffers from poor recognition rate due to pose, illumination, expression, ageing, makeup variations and occlusions. In recent years, the research focus has shifted toward face recognition using 3-D facial surface and shape which represent more discriminating features by the virtue of increased dimensionality. This paper presents an extensive survey of recent 3-D face recognition techniques in terms of feature detection, classifiers as well as published algorithms that address expression and occlusion variation challenges followed by our critical comments on the published work. It also summarizes remarkable 3-D face databases and their features used for performance evaluation. Finally we suggest vital steps of a robust 3-D face recognition system based on the surveyed work and identify a few possible directions for research in this area.

Optimal Pattern Synthesis of Linear Antenna Array Using Grey Wolf Optimization Algorithm

The aim of this paper is to introduce the grey wolf optimization (GWO) algorithm to the electromagnetics and antenna community. GWO is a new nature-inspired metaheuristic algorithm inspired by the social hierarchy and hunting behavior of grey wolves. It has potential to exhibit high performance in solving not only unconstrained but also constrained optimization problems. In this work, GWO has been applied to linear antenna arrays for optimal pattern synthesis in the following ways: by optimizing the antenna positions while assuming uniform excitation and by optimizing the antenna current amplitudes while assuming spacing and phase as that of uniform array. GWO is used to achieve an array pattern with minimum side lobe level (SLL) along with null placement in the specified directions. GWO is also applied for the minimization of the first side lobe nearest to the main beam (near side lobe). Various examples are presented that illustrate the application of GWO for linear array optimization and, subsequently, the results are validated by benchmarking with results obtained using other state-of-the-art nature-inspired evolutionary algorithms. The results suggest that optimization of linear antenna arrays using GWO provides considerable enhancements compared to the uniform array and the synthesis obtained from other optimization techniques.

Linear antenna array optimization using flower pollination algorithm

Flower pollination algorithm (FPA) is a new nature-inspired evolutionary algorithm used to solve multi-objective optimization problems. The aim of this paper is to introduce FPA to the electromagnetics and antenna community for the optimization of linear antenna arrays. FPA is applied for the first time to linear array so as to obtain optimized antenna positions in order to achieve an array pattern with minimum side lobe level along with placement of deep nulls in desired directions. Various design examples are presented that illustrate the use of FPA for linear antenna array optimization, and subsequently the results are validated by benchmarking along with results obtained using other state-of-the-art, nature-inspired evolutionary algorithms such as particle swarm optimization, ant colony optimization and cat swarm optimization. The results suggest that in most cases, FPA outperforms the other evolutionary algorithms and at times it yields a similar performance.

Spanner shaped ultra wideband patch antenna

In recent years Cognitive radio and UWB technologies are emerging for efficient use of Spectrum for short range wireless communications. FCC also has released a wideband of 7.5 GHz (From 3.1GHz to 10.6GHz) as unlicensed to use it anytime anywhere. In cognitive radio for sensing purpose and in UWB transceiver, UWB antennas play a vital role. Since antenna dimensions are frequency dependent, designing an antenna for wideband is a challenging task. This paper discusses the step by step techniques to obtain ultra wide bandwidth (2.23GHz-11.4GHz) from narrowband patch antenna. Also it discusses the various antenna parameters of proposed antenna design, its simulated and experimental results.

Expression invariant face recognition using contourlet transform

The performance of many state-of-the-art expression invariant face recognition systems hampers when fewer faces are available in the gallery. This paper addresses the issue of expression invariant face recognition with small gallery set. The contourlet transform is an established tool for capturing contour-like edges. The contourlet transform generates prominent features as its local and directional properties have strong resemblance with human visual cortex. We have proposed a novel approach that fuses the features from spatial domain and contourlet transform domain. Feature extraction is performed by employing the LBP and WLD descriptors. The experiments are performed on the JAFFE face database and the Yale face database. The results indicate that the proposed feature level fusion approach yields a robust feature vector and exemplary recognition rates.

Rough Set Approaches to Unsupervised Neural Network Based Pattern Classifier

Unsupervised neural network based pattern classification is a widely popular choice for many real time applications. Such applications always face challenges of processing data with lot of consistency, inconsistency, ambiguity or incompleteness. Hence to deal with such challenges a strong approximation tool is always needed. Rough set is one such tool and various approaches based on Rough set, if are applied to pure neural (unsupervised) pattern classifier can yield desired results like faster convergence, feature space reduction and improved classification accuracy. The application of such approaches at respective level of implementation of neural network based pattern classifier for two case studies are discussed here. Whereas more emphasis is given on the preprocessing level based approach used for feature space reduction.

Rough Neuron Based Neural Classifier

Rough sets theory can be applied to the problem of pattern recognition using neural networks in three different stages: preprocessing, learning rule and in the architecture. This paper discusses the use of rough set theory in the architecture of the unsupervised neural network, which is implemented, by the use of rough neuron. The rough neuron consists of two neurons: upper boundary neuron and lower boundary neuron, derived on the upper and lower boundaries of the input vector. The proposed neural network uses the Kohonen learning rule. Problem of character recognition is taken to verify the usefulness of such a network. The data set is formed by the images of English alphabets of ten different fonts. The approximation quality of such a network is better compared to the traditional networks. The number of iterations reduce significantly for such a network and hence the convergence time.

Design and Implementation of Rough Set Algorithms on FPGA: A Survey

Rough set theory, developed by Z. Pawlak, is a powerful soft computing tool for extracting meaningful patterns from vague, imprecise, inconsistent and large chunk of data. It classifies the given knowledge base approximately into suitable decision classes by removing irrelevant and redundant data using attribute reduction algorithm. Conventional Rough set information processing like discovering data dependencies, data reduction, and approximate set classification involves the use of software running on general purpose processor. Since last decade, researchers have started exploring the feasibility of these algorithms on FPGA. The algorithms implemented on a conventional processor using any standard software routine offers high flexibility but the performance deteriorates while handling larger real time databases. With the tremendous growth in FPGA, a new area of research has boomed up. FPGA offers a promising solution in terms of speed, power and cost and researchers have proved the benefits of mapping rough set algorithms on FGPA. In this paper, a survey on hardware implementation of rough set algorithms by various researchers is elaborated.

Tumor segmentation by tolerance near set approach in mammography and lesion classification with neural network

The mammography is the most effective procedure for an early diagnosis of the breast cancer. In this paper, a new algorithm to detect suspicious lesions in mammograms is developed using tolerance near set approach. Near set theory provides a method to establish resemblance between objects contained in a disjoint set. Objects that have, in some degree, affinities are considered perceptually near each other. The probe functions are defined in terms of digital images such as: gray level, entropy, color, texture, etc. Objects in visual field are always presented with respect to the selected probe functions. Moreover, it is the probe functions that are used to measure characteristics of visual objects and similarities among perceptual objects, making it possible to determine if two objects are associated with the same pattern. The algorithm has been verified on mammograms from the CICRI (Central India Cancer Research Institute, Nagpur, India) and Mias database. Results of segmentation are compared with Otsu method of segmentation.. Once the features are computed for each region of interest (ROI), they are used as inputs to a supervised Back Propagation Neural Network. Results indicate that Tolerance Near sets segmentation method performs better than otsu method in terms of classification accuracy.

Expression invariant face recognition using semidecimated DWT, Patch-LDSMT, feature and score level fusion

This paper addresses the issue of human face recognition in presence of expression variations, which pose a great challenge to face recognition systems. Typically, the discriminant features lie in both spatial as well as transform domain. In this paper, we propose combination of Discrete Wavelet Transform (DWT) and proposed Semi-decimated Discrete Wavelet Transform (SDWT) to develop an expression invariant face recognition algorithm followed by a novel wavelet coefficients enhancement function. The wavelet coefficients are boosted using the proposed coefficients enhancement function and extracted using the Weber Local Descriptors (WLD). This enhances weak skin edges based features, resulting in increased probability of recognition. The proposed algorithm also exploits spatial domain features using our customized version of Complete Local binary patterns (CLBP) named Patch Local Difference Sign Magnitude Transform (Patch-LDSMT) applied on complete images and physiologically meaningful overlapping regions of human facial images for the first time. Feature level fusion of the wavelet based features and Patch-LDSMT yields a robust feature vector whose dimensionality is reduced using Linear Discriminant Analysis (LDA). Comprehensive experimentation is carried out on the JAFFE, CMU-AMP, ORL, Yale, Cohn-Kanade (CK) and database collected by us. Benchmarking analysis illustrates that the proposed face recognition algorithm offers much better rank one recognition performance when compared with the current state-of-the-art expression invariant face recognition approaches.