Explore the vast range of titles available.

Hand gesture recognition (HGR) is the most effective and intuitive way for the human–computer interface used in various applications, such as sign language recognition, robotics, and multimedia applications. The performance of the existing handcrafted techniques relies on the less generalized preprocessing and feature extraction steps. The convolutional neural networks (CNNs) can handle the less generalization characteristic of the handcrafted techniques. However, the architecture of CNN is complex and the extracted deep feature from these networks provides the global information only. Therefore, a CNN-based HGR paradigm can be developed with less number of layers and feature fusion with global and local information from different layers. Motivated by the above facts, this work proposes (i) a two-stage residual CNN (2RCNN) architecture for learning of features from the color hand gesture images which overcomes the need of a specific preprocessing step, (ii) a novel residual block intensity (RBI) feature to extract the global and local information from the hand gesture images. After extracting the RBI features, a linear kernel-based multi-class support vector machine classifier is used to recognize the gesture poses. The experimental results are evaluated using a subject-independent cross-validation test on three benchmarked datasets and compared with the earlier reported techniques.

Applications for gait recognition are numerous especially in security surveillance. However, due to the variety of individual walking behaviours and the complexities of external variables during data gathering, gait identification continues to face several obstacles. Among these, shallow learning-based gait recognition algorithms struggle to attain the correct rate of recognition crucial by numerous applications, while the volume of gait training data available cannot match the demands of deep learning-based model training. In order to comply with the problem outlined above, this work offers a visual gait detection system based on an attention-based multi-scale convolutional network with sequential learning. As a first step, the approach takes multi-recurrent gait energy images (MR-GEIs) as an input in a frame by frame manner and runs each frame through the convolutional network to extract entire gait features. In the second step, the key attributes of the extracted features from multi-scale convolutions are highlighted by the attention block in order to improve prediction performance. Thirdly, the bidirectional gated recurrent unit (Bi-GRU) layer is applied to obtain the temporal relationships among the different frames of MR-GEIs in the sequential learning block. The proposed network achieves average accuracies of 93.4% on CASIA-B and 97.4% on OULP gait datasets, demonstrating superior recognition performance and improved generalizability compared to previous state-of-the-art techniques.

Classification of electrocardiogram (ECG) signals plays an important role in clinical diagnosis of heart disease. This paper proposes the design of an efficient system for classification of the normal beat (N), ventricular ectopic beat (V), supraventricular ectopic beat (S), fusion beat (F), and unknown beat (Q) using a mixture of features. In this paper, two different feature extraction methods are proposed for classification of ECG beats: (i) S-transform based features along with temporal features and (ii) mixture of ST and WT based features along with temporal features. The extracted feature set is independently classified using multilayer perceptron neural network (MLPNN). The performances are evaluated on several normal and abnormal ECG signals from 44 recordings of the MIT-BIH arrhythmia database. In this work, the performances of three feature extraction techniques with MLP-NN classifier are compared using five classes of ECG beat recommended by AAMI (Association for the Advancement of Medical Instrumentation) standards. The average sensitivity performances of the proposed feature extraction technique for N, S, F, V, and Q are 95.70%, 78.05%, 49.60%, 89.68%, and 33.89%, respectively. The experimental results demonstrate that the proposed feature extraction techniques show better performances compared to other existing features extraction techniques.

A vision-based static hand gesture recognition method which consists of preprocessing, feature extraction, feature selection and classification stages is presented in this work. The preprocessing stage involves image enhancement, segmentation, rotation and filtering. This work proposes an image rotation technique that makes segmented image rotation invariant and explores a combined feature set, using localized contour sequences and block-based features for better representation of static hand gesture. Genetic algorithm is used here to select optimized feature subset from the combined feature set. This work also proposes an improved version of radial basis function (RBF) neural network to classify hand gesture images using selected combined features. In the proposed RBF neural network, the centers are automatically selected using k-means algorithm and estimated weight matrix is recursively updated, utilizing least-mean-square algorithm for better recognition of hand gesture images. The comparative performances are tested on two indigenously developed databases of 24 American sign language hand alphabet.

This study presents a variable length multi-path delay commutator fast Fourier transform (FFT)/inverse FFT (IFFT) architecture for a multiple input multiple output orthogonal frequency division multiplexing system. It supports the FFT/ IFFT lengths of 512/256/128/64 samples to process each symbol carried by eight spatial streams and achieves a speed of 160 MHz to meet the IEEE 802.11ac timing requirements. A resource scheduling methodology to minimise the hardware complexity of the design is proposed and adopted in the architecture presented. A novel stagger word length strategy is also proposed and applied to achieve the better accuracy with lesser hardware. Here, the signal to quantisation noise ratio of 57.23 dB is obtained. The twiddle coefficient storage space is significantly compressed to achieve the coefficient generation with reduced hardware. The design is implemented using the TSMC-65 nm complementary metal oxide semiconductor technology with a supply voltage of 1 V at 160 MHz. The implementation results show that the architecture has a gate count of 3,48,013 with power consumption of 105.1 mW and area of 0.492 mm2. The hardware complexity and performance of the design are compared with earlier reported architectures. It is observed that the proposed design achieves better performance in terms of hardware complexity and normalised energy for the given specifications.

Edge detection is an important step for finding the discontinuities of images and detecting the boundaries of objects. This work presents a novel algorithm for image edge detection using ant colony optimization and Fisher ratio ( F ratio)-based techniques. Ants generally search the food from the nest to the food source in the way that maximizes the intensity of pheromone (a chemical secretion). The proposed technique considers that the movements of the artificial ants are steered by the local intensity variation in the image pixel. The directions of ants movements in the image are determined using a direction probability matrix, computed by pheromone and heuristic information of possible directions. In this work, F ratio technique is utilized to determine the optimum threshold value from updated pheromone matrix. This threshold value is further used to extract binary edge map from pheromone matrix. The experiment is conducted on the different test images, i.e., Cameraman, Lena, Coins, Peppers, House and Pillsetc image. The proposed edge detection algorithm is evaluated on the basis of statistical parameters such as kappa, figure of merit, Baddeley’s delta metric and Hausdorff distance, and the experimental results show that the proposed method performs better as compared to earlier reported techniques in most of the cases.

The IEEE 802.11ac is the recently ratified standard developed for the fifth generation wireless fidelity technology, in which the multi-user (MU) multiple-input multiple-output orthogonal frequency division multiplexing (MIMO-OFDM) technique is adopted for the high data rate communication. In an MIMO-OFDM System, the forward/inverse fast Fourier transform (FFT/IFFT) processor is a key component. On proper reception, the reordering and scheduling of data is important for the optimal utilisation of butterfly resources in the pipelined FFT/IFFT processor. In this study, a mathematical model for an eight-parallel multimode (N = 512/256/128/64) multi-path delay commutator-based FFT/IFFT processor which is suitable for the IEEE 802.11ac compliant MU-MIMO-OFDM system is presented. On the other hand, the data reordering, scheduling methodologies and its architectures are proposed for the pre-, post-FFT/IFFT process are proposed. The design implementations are done using TSMC 65 nm complementary metal–oxide–semiconductor technology at 160 MHz. The power and area metrics with and without clock gating are compared. The clock gated implementation reports show that the power consumption is 17.44 mW for the pre-transformed data reordering and 11.64 mW for the post-transformed data reordering with an area occupation of 0.7694 mm2 and 0.5111 mm2, respectively.

Electrocardiogram (ECG) is generally used for diagnosis of cardiovascular abnormalities and disorders. An efficient method for analysing the ECG signal towards the detection of heart rate (HR) and ischemic episodes follows mainly five stages: pre-processing, feature extraction, heart rate detection, beat classification and ischemic episode recognition. The heart rate is calculated using the extracted features of the ECG signal. The calculated HR value can be analysed for the detection of various cardiovascular abnormalities. The ability of the method was validated on European ST-T database. The performance of ischemic episode detection shows 88.08% sensitivity (Se) and 92.42% positive predictive accuracy (PPA).