An Efficacious Method for Face Recognition Using DCT and Neural Network

The recent advances in the feature extraction techniques in recognition of handwritten digits attract researchers to work in this area. The present study includes recognition of handwritten digits using hybrid feature extraction technique including static and dynamic properties of handwritten digit images. In this paper, static properties include number of non-zero (white) pixels in square, horizontal, vertical and diagonal styles as sub regions of a binary image. The dynamic properties include features from recovery of drawing order of original image. The extraction of dynamic features include two stages: first stage recover the drawing order of an image and second stage compute the chain code directions from recovered drawing order. The algorithm for recovery of drawing order uses properties of writing behavior. The support vector machine has been used as recognition method for the proposed feature extraction scheme. We have achieved an overall error rate of 0.73 % for mnist data set including 60,000 training images and 10,000 test images. Our feature extraction technique results in feature vector length of an image equals to 356. The achieved results strengthen our proposed technique usability as error rate achieved is at par with literature (\(<\)1 %) and the length of feature vector per image is small in comparison to input feature vector length of 784 which has been commonly used in previous work. The developed system is stable and useful in real life applications.

Content-based image retrieval (CBIR) systems use multiple image features to represent an image. These systems suffer from curse of dimensionality since a high-dimensional feature vectors are formed to represent images in the target dataset. Reduction in length of feature vector can speed-up the retrieval process, but it reduces the retrieval accuracy. To overcome the aforementioned problem, an efficient layer-wise feature incremental approach for the CBIR system has been proposed. The proposed approach uses three primitive image features namely color, texture, and shape. The retrieval process is accomplished in three layers, at first a layer-complete dataset is searched but only single-feature space is used. Top 10% of images most similar to query image are retained in the second layer. The second layer uses two features for similarity computation, and only 50% of the most similar images are passed to the third layer. Finally, the third layer uses all three features to compute the similarity. Our aim is to reduce the search space at subsequent layers and use multiple features for a reduced dataset at the final layer. The proposed approach is evaluated on four publicly available image datasets. The retrieval results on the basis of precision, recall, and f-score show the performance improvement in comparison to state-of-the-art CBIR systems.

Texture image retrieval system using contourlet transform has better performance than the same structure system based on wavelet transform due to contourlet's better directional information representation than wavelet transform. In order to improve the retrieval rate further, a dual-tree complex contourlet transform based texture image retrieval system was proposed in this paper. In the system, the dual tree contourlet transform was used to transform each image into contourlet domain and implemented multiscale decomposition, sub-bands energy, standard deviations and skewness in contourlet domain were cascaded to form feature vectors, and the similarity metric used here is Canberra distance. Experimental results on brodatz test images set show that dual tree contourlet transform based image retrieval system is superior to those of the original contourlet transform, non-subsampled contourlet transform under the same system structure with almost same length of feature vectors, retrieval time and memory needed; and contourlet decomposition structure parameter can make significant effects on retrieval rates, especially scale number.

There has been much interest recently in image content based retrieval, with applications to digital libraries and image database accessing. One approach to this problem is to base retrieval from the database upon feature vectors which characterize the image texture. Since feature vectors are often high dimensional, multi-dimensional scaling, or non-linear principal components analysis (PCA) may be useful in reducing feature vector size, and therefore computation time. We have investigated a variant of the non-linear PCA algorithm described by Webb (see Pattern Recognition, vol.28, no.6, p.753-9, 1995) and its usefulness in the database retrieval problem. The results are quite impressive; in an experiment using an aerial photo database, the feature vector length was reduced by a factor of 10 without significantly reducing the retrieval performance.

Nondominated sorting genetic algorithm II (NSGA-II) is well known for engine optimization problem. Artificial neural networks (ANNs) followed by multi-objective optimization including a NSGA-II and strength pareto evolutionary algorithm (SPEA2) were used to optimize the operating parameters of a compression ignition (CI) heavy-duty diesel engine. First, a multi-layer perception (MLP) network was used for the ANN modeling and the back propagation algorithm was utilized as training algorithm. Then, two different multi-objective evolutionary algorithms were implemented to determine the optimal engine parameters. The objective of the present study is to decide which algorithm is preferable in terms of performance in engine emission and fuel consumption optimization problem.

The classification of the sound heart into different valve-physiological heart disease categories is a complex pattern recognition task. This paper will purpose sound heart recognition for diagnosing heart disease with 4 type of Artificial Neural Network (ANN). We develop a simple model for the recognition of heart sounds, and demonstrate its utility in identifying features useful in diagnosis. We then present a prototype system intended to aid in heart sound analysis. Based on a wavelet decomposition of the sounds, feature vectors are formed and ANNs finds use in classification of Heart valve diseases for its discriminative training ability and easy implementation. The heart sound diseases classes considered for the purpose of this study were classified into normal heart sound and the other six valve physiological heart categories. 4 type of ANN which used for this approach are Multilayer perceptron!(MLP), Back Propagation Algorithm (BPA), Elman Neural Network (ENN) and Radial Basis Function (RBF) Network. Using these ANN classifiers would appeared ability of classifying heart disease and will be shown an accuracy of 81.25 % for MLP, 87.17% for BPA, 91.59% for ENN and 96.42% for RBF was achieved.

A distortion-agnostic video quality metric based on multi-scale spatio-temporal structural information

Face recognition has been a major research theme over the last two decades. There are several problems to be solved to improve the performance of face recognition. Such major problems involve appearance variation due to pose, illumination, expression, and aging. In particular, aging includes internal and external factors that cause facial appearance variation and, consequently, it is the most difficult problem to handle. In this paper, we propose a face recognition method that is robust against facial appearance variation due to aging. The proposed method employs segmentation verification of frontal face images that consists of the following three steps. (1) Face image segmentation generates three regional subimages from the input face image. (2) A matching score is calculated using gradient features from a pair consisting of the input image and a registered image for each of the three generated subimages and original (whole face) image. We obtain four matching scores. (3) The verifying classifier evaluates the matching score vector formed of the matching scores calculated for each of the four images and predicts the a posteriori probability that two matching images belong to the same person. The results of an experimental evaluation with the FGNET and MORPH face aging datasets clarify the effectiveness of the proposed method for age invariant face recognition.

The distinguishing characteristics belong to a person are personality traits, which is predicted from person’s behavioral pattern. As most of the people provide lots of information knowingly or unknowingly into their writings, it is possible to extract personality traits from those texts. Individual personality traits detection from texts, yields enormous possibilities toward various applications named as forensic department, mental health diagnosis, etc. Meanwhile, deep learning algorithm performs fairly well in text based personality detection; however, its performance may vary with activation functions. Hence, this paper proposed an empirical approach to find the best personality detection performance by comparing several activation functions named as sigmoid, tanh, and leaky ReLU. Here, text documents were pre-processed and vectorized for input in convolutional neural network. The input size was multiple to length of word, sentence, documents, and feature vectors. Five personality traits named as EXT, NEU, AGR, CON, and OPN have been used for experimental analysis. The result showed that tanh and leaky ReLU performs over sigmoid in all datasets. The average F1-score of sigmoid, tanh and leaky ReLU showed 33.11%, 47.25%, and 49.07% respectively. However, Fl-score of leaky ReLU was high only for CON, tanh showed better result for others datasets. The overall performance showed by tanh is better than sigmoid and leaky ReLU for personality detection from text.

This article presents an efficient and totally unsupervised content-based similarity search method for multimedia data objects represented by high-dimensional feature vectors. The assumption is that the similarity measure is applicable to feature vectors of arbitrary length. During the offline process, different sets of features are selected by a generalized version of the Laplacian Score in an unsupervised way for individual data objects in the database. Online retrieval is performed by ranking the query object in the feature spaces of candidate objects. Those candidates for which the query object is ranked highly are selected as the query results. The ranking scheme is incorporated into an automated query expansion framework to further improve the semantic quality of the search result. Extensive experiments were conducted on several datasets to show the capability of the proposed method in boosting effectiveness without losing efficiency.

Levenberg-Marquardt back-propagation (LMBP) algorithm is one of the most widely used alternative algorithms for training of artificial neural networks (ANN). In some cases, learning in LM can become a challenging machine-learning task owing to poor selection of weights and biases. In this manner, the network may become unstable and cause the search process to get stuck in local minima (i.e. resulting into network stagnancy). Recently, meta-heuristic algorithms have proven themselves to alleviate such problem and improve the convergence slow convergence in ANN. In this paper, Wolf Search (WS) that simulates the hunting behavior of wolves in their natural habitat is implemented with LMBP algorithm to optimize the weight selection in ANN. The performance of the proposed Wolf Search with Levenberg-Marquardt (WSLM) is compared with back-propagation and other stochastic LM algorithms by means of simulations on benchmark datasets. The simulation results show that the proposed WSLM performs superior to BPNN, ABC-BP, Bat-BP, and ABC-LM algorithms in terms of convergence speed and local minima avoidance.

Neural network classifier optimization using Differential Evolution with Global Information and Back Propagation algorithm for clinical datasets

In this paper, we present a system based on a feature selection approach for solving the recognition problem in a multimodal biometric system combining both left and right palmprints of the same subject. In particular, the fusion of two or more traits, at feature-level, results in a long feature vector that needs large storage space, makes the execution time of the recognition task very long, and may include redundant and irrelevant features that can affect the recognition accuracy. To overcome these problems, feature selection is performed using genetic algorithms (GAs) and backtracking search algorithm for a comparison purpose. The experimental results show the usefulness of feature selection, especially the use of genetic algorithms, on the robustness of the multimodal biometric system as regards the feature vector length and run-time reduction, and the significant increase of the recognition rate.

Iris recognition is a means of biometric identification. A key part of the recognition system using iris is the extraction of prominent texture information or features in the iris. The identification delay in iris recognition can be reduced by reducing the feature vector generated from the feature extraction of iris images. A new form of local binary pattern (LBP) termed LBPX is proposed in this paper as an iris feature extraction method. For this, input eye images are processed and converted to normalized iris images employing circular Hough transformation and Daugman’s rubber sheet model. Next, LBPX is applied to the normalized images. In this LBPX stage, rotation-invariant LBP operation takes place. The performance of LBPX based recognition system adopting iris image is evaluated in terms of accuracy and feature vector length. This is done for three datasets CASIA-IRIS-V4, UBIRIS and IITD. Results indicate that LBPX can achieve acceptable accuracy values of 97.20%, 96% and 96.40% for CASIA-IRIS-V4, UBIRIS and IITD datasets, respectively. Furthermore, results show LBPX outperforms existing feature extraction methods in terms of reduced feature-length, ensuring faster iris recognition.

Artificial Neural Networks are a computational paradigm formed based on the neural structure of intelligent organisms to gain better knowledge. Artificial neural networks are often used for various computing purposes. One of them is for prediction (forecasting) data. The type of artificial neural network that is often used for prediction is the artificial neural network backpropagation because the backpropagation algorithm is able to learn from previous data and recognize the data pattern. So from this pattern backpropagation able to analyze and predict what will happen in the future. In this study, the data to be predicted is Human Development Index data from 2011 to 2015. Data sourced from the Central Bureau of Statistics of North Sumatra. This research uses 5 architectural models: 3-8-1, 3-18-1, 3-28-1, 3-16-1 and 3-48-1. From the 5 models of this architecture, the best accuracy is obtained from the architectural model 3-48-1 with 100% accuracy rate, with the epoch of 5480 iterations and MSE 0.0006386600 with error level 0.001 to 0.05. Thus, backpropagation algorithm using 3-48-1 model is good enough when used for data prediction.