A robust gesture detection and recognition algorithm for domestic robot interactions

Objectives: The primary objective of this paper is to evaluate the correctness of the algorithm Gesture Recognition that is one of the parts of our novel method ‘Braille coded gesture pattern interaction method”; and to study the performance of Eyedroid system using the statistical methodologies. Methods/Statistical Analysis: The algorithm for gesture recognition was developed and implemented in C program to check the correctness of each step of algorithm. The gesture input parameters Swipe-Minimum-Distance, Swipe-Threshold-Velocity are defined with range of values. Varying gesture input range of values within domain and out of domain results the expected solution space. Our gesture recognition algorithm is developed for 2D touch surface. The Real time experiments were conducted for 12 blind fold subjects to observe the different versions Eyedroid-A(E-A), Eyedroid-B(E-B) of gesture recognition algorithm perceptions, being implemented in smart phones. The mean WPM for 10 trials was recorded for 12 subjects with E-A and E-B. The two tailed paired t test is performed for the mean dependent sample data of wpm. Findings: The existing 3D gesture recognition algorithm Derivative Dynamic Time Warping (DDTW) analyzed by Katarzyna Barczewska, Aleksandra Drozd, 2013 was proven to be efficient method for air fly environment. Our gesture recognition algorithm is developed for 2D touch surface found to be effective by the results of algorithm test. Gesture Recognition algorithm finds the mal gestures and the results proved that there is a significant difference in respect of blind subjects for two different Eyedroid versions. The observed data reveals E-B (mean WPM=14.95) is easy to use than E-A (mean WPM=12) as the number of words entered per minute by E-B is higher than E-A. Our Eyedroid with WPM is 14.95 is obtained to be better interactive system than the existing “Adaptive interaction Technique” system developed by Georgios Y fantidis Grigori Evreinov, 2005 confirms WPM is 12. Applications/Improvements: The implication is our Braille coded gesture pattern that enables visually challenged ones to enter more number of words than by the existing system. Our interaction method is deployable at all kinds of touchable devices.Keywords: Eyedroid-Braille Coded Gesture Pattern, Gesture Recognition Algorithm, t Test for Mobile Interaction

In recent years, wearable sensor-based gesture recognition is proliferating in the field of healthcare. It could be used to enable remote control of medical devices, contactless navigation of X-ray display and Magnetic Resonance Imaging (MRI), and largely enhance patients' daily living capabilities. However, even though a few commercial or prototype devices are available for wearable gesture recognition, none of them provides a combination of (1) fully open API for various healthcare application development, (2) appropriate form factor for comfortable daily wear, and (3) affordable cost for large scale adoption. In addition, the existing gesture recognition algorithms are mainly designed for discrete gestures. Accurate recognition of continuous gestures is still a significant challenge, which prevents the wide usage of existing wearable gesture recognition technology. In this paper, we present Gemote, a smart wristband-based hardware/software platform for gesture recognition and remote control. Due to its affordability, small size, and comfortable profile, Gemote is an attractive option for mass consumption. Gemote provides full open API access for third party research and application development. In addition, it employs a novel continuous gesture segmentation and recognition algorithm, which accurately and automatically separates hand movements into segments, and merges adjacent segments if needed, so that each gesture only exists in one segment. Experiments with human subjects show that the recognition accuracy is 99.4% when users perform gestures discretely, and 94.6% when users perform gestures continuously.

Ultigesture: A wristband-based platform for continuous gesture control in healthcare

This paper presents a new algorithm, called Dynamic Arc-Length Warping algorithm (DALW) for hand gesture recognition based on the orientation data. In this algorithm, after calculating the quaternion for each orientation measurement, we use DALW algorithm to obtain a similarity measure between different trajectories. We present the benefits of using quaternion alongside the implementation of Dynamic Arc Length Warping to present an optimized tool for gesture recognition.We show the advantages of this approach compared with other techniques. This tool can be used to distinguish similar and different gestures. An experimental validation is carried out to classify a series of simple human gestures.

To address the problems of the traditional human motion gesture tracking and recognition methods, such as poor tracking effect, low recognition accuracy, high frame loss rate, and long-time cost, a dynamic human motion gesture tracking and recognition algorithm using multimode deep learning was proposed. Firstly, the collected human motion images are repaired in the three-dimensional (3D) environment, and the multimodal 3D human motion model is reconstructed using the processed images. Secondly, according to the results of model reconstruction, the camera gesture and other parameters of the keyframe are used to construct the target tracking optimization function so as to achieve the purpose of accurate tracking of human motion. Finally, for multimodal human motion gesture learning, a convolutional neural network (CNN) is developed. The trained CNN is utilized to complete dynamic human motion recognition after convolutional and pooling calculations. The results demonstrate that the proposed algorithm is effective in tracking human motion gestures. The average recognition accuracy is 96%, the average frame loss rate is 8.8%, the time cost is low, and the proposed algorithm has a high F-measure and much lower power consumption than other algorithms, indicating that the proposed algorithm is effective.

As a mode of perpetuating and revitalizing traditional culture, cultural and creative products have garnered widespread affection and recognition from the public. In the context of evolving societal trends and advancements in science and technology, the digitization of traditional cultural and creative products has emerged as a prominent trend. This study undertakes the digital design of these products, primarily focusing on the innovative application of a style transfer algorithm to traditional motifs, supplemented by their visualization through digital platforms such as augmented reality (AR) and virtual reality (VR). Furthermore, it facilitates intelligent consumer interaction via gesture recognition algorithms, thereby enhancing user engagement and experience. During the implementation phase, this research conducts comparative analyses of style transfer and gesture recognition within digital cultural and creative products. It also employs the Kano questionnaire to categorize and analyze user needs effectively. Notably, while the recall rate of the style transfer algorithm documented in this study remains below 0.9, it consistently achieves high precision, significantly enhances feature extraction capabilities, and improves the quality of the style transfer effects produced. Moreover, the static gesture recognition algorithm achieves an impressive average recognition rate of 98.2%. The dynamic gesture recognition algorithm, meanwhile, maintains an average recognition rate of 94% with a processing time of 3.2 seconds, balancing the demands of real-time interaction and accuracy effectively. This study systematically analyzes the significance of each requirement element across six dimensions, classifying customer needs for digital cultural and creative products into four distinct categories. Additionally, it delineates a viable pathway for the integration of digital art elements into the design of artistic and innovative products, setting a foundation for future innovations in this field.

Many of the state-of-the-art algorithms for gesture recognition are based on Conditional Random Fields (CRFs). Successful approaches, such as the Latent-Dynamic CRFs, extend the CRF by incorporating latent variables, whose values are mapped to the values of the labels. In this paper we propose a novel methodology to set the latent values according to the gesture complexity. We use an heuristic that iterates through the samples associated with each label value, stimating their complexity. We then use it to assign the latent values to the label values. We evaluate our method on the task of recognizing human gestures from video streams. The experiments were performed in binary datasets, generated by grouping different labels. Our results demonstrate that our approach outperforms the arbitrary one in many cases, increasing the accuracy by up to 10%.

Gesture recognition is a field of research that consists in recognizing patterns of movement performed by a human using his body parts with or without the help of a comprehensive device (a mouse, a laser, etc.). This particular area as attracted a number of researchers over the years that applied such algorithms in a broad range of disciplines. In particular, it was exploited on early research initiatives with pervasive environments to enable simple communication with automation systems. Nowadays, those environments are used for more than automation. Many researchers, in fact, believe it is one of the most promising solutions to the problems related to ageing of the population. Smart homes are seen as an alternative to the full-time support of a semi-autonomous person by healthcare professional and thus also a potentially economically viable solution to the rising cost of such support. However, researchers are still facing many challenges in that regards, such as the comprehension of the context and of the ongoing activity of daily living. In that equation, gesture recognition could help extract more information from the collected data and thus reinforce the context modeling. The knowledge extracted could even help with monitoring of more fine-grained activities and with the understanding of normal or abnormal behaviors. Gesture recognition is often considered as a solved problem since the techniques to perform it work well as soon as we can track accurately. However, in smart environment, tracking is very imprecise and hard to achieve with limited technology (i.e. noninvasive sensors). In this paper, we present a novel gesture recognition algorithm that works under uncertainty, and that is scalable to the precision of the tracking system. The algorithm is based on the tracking of passive RFID tags installed on all everyday life objects in a smart environment. A set of experimentation in simulation and in a real smart home environment is presented. The results are very encouraging despite the very low accuracy of the passive RFID tracking system.

In this article, a real-time dynamic finger gesture recognition using a soft sensor embedded data glove is presented, which measures the metacarpophalangeal (MCP) and proximal interphalangeal (PIP) joint angles of five fingers. In the gesture recognition field, a challenging problem is that of separating meaningful dynamic gestures from a continuous data stream. Unconscious hand motions or sudden tremors, which can easily lead to segmentation ambiguity, makes this problem difficult. Furthermore, the hand shapes and speeds of users differ when performing the same dynamic gesture, and even those made by one user often vary. To solve the problem of separating meaningful dynamic gestures, we propose a deep learning-based gesture spotting algorithm that detects the start/end of a gesture sequence in a continuous data stream. The gesture spotting algorithm takes window data and estimates a scalar value named gesture progress sequence (GPS). GPS is a quantity that represents gesture progress. Moreover, to solve the gesture variation problem, we propose a sequence simplification algorithm and a deep learning-based gesture recognition algorithm. The proposed three algorithms (gesture spotting algorithm, sequence simplification algorithm, and gesture recognition algorithm) are unified into the real-time gesture recognition system and the system was tested with 11 dynamic finger gestures in real-time. The proposed system took only 6 ms to estimate a GPS and no more than 12 ms to recognize the completed gesture in real-time.

A gesture recognition system using stereo vision and arm model fitting

Classification algorithms require training data initially labelled by classes to build a model and then to be able to classify the new data. The amount and diversity of training data affect the classification quality and usually the larger the training set, the better the accuracy of classification. In many applications only small amounts of training data are available. This article presents a new time series classification algorithm for problems with small training sets. The algorithm was tested on hand gesture recordings in tasks of person identification and gesture recognition. The algorithm provides significantly better classification accuracy than other machine learning algorithms. For 22 different hand gestures performed by 10 people and the training set size equal to 5 gesture execution records per class, the error rate for the newly proposed algorithm is from 37% to 75% lower than for the other compared algorithms. When the training set consists of only one sample per class the new algorithm reaches from 45% to 95% lower error rate. Conducted experiments indicate that the algorithm outperforms state-of-the-art methods in terms of classification accuracy in the problem of person identification and gesture recognition.

In the immersive transformer training simulator, with inertia motion capture system, the operator's motion data was captured and analyzed. The motions capture data of hand movements and gesture recognition algorithm were studied. A method of continuous space-time tracking of the hand joints was realized using skeleton graph. The hand movement features of natural human-machine interaction were defined. A method of hand segmentation from the depth graph background using hand joints of skeleton graph was proposed. Three kinds of gesture recognition method were realized with SVM and studied about stability and robustness. Semantics of gesture, mapped to virtual environments input events. The immersive transformer training simulator was realized based on the research.

In the course of our research work, the American, Russian and Turkish sign languages were analyzed. The program of recognition of the Kazakh dactylic sign language with the use of machine learning methods is implemented. A dataset of 5000 images was formed for each gesture, gesture recognition algorithms were applied, such as Random Forest, Support Vector Machine, Extreme Gradient Boosting, while two data types were combined into one database, which caused a change in the architecture of the system as a whole. The quality of the algorithms was also evaluated. The research work was carried out due to the fact that scientific work in the field of developing a system for recognizing the Kazakh language of sign dactyls is currently insufficient for a complete representation of the language. There are specific letters in the Kazakh language, because of the peculiarities of the spelling of the language, problems arise when developing recognition systems for the Kazakh sign language. The results of the work showed that the Support Vector Machine and Extreme Gradient Boosting algorithms are superior in real-time performance, but the Random Forest algorithm has high recognition accuracy. As a result, the accuracy of the classification algorithms was 98.86 % for Random Forest, 98.68 % for Support Vector Machine and 98.54 % for Extreme Gradient Boosting. Also, the evaluation of the quality of the work of classical algorithms has high indicators. The practical significance of this work lies in the fact that scientific research in the field of gesture recognition with the updated alphabet of the Kazakh language has not yet been conducted and the results of this work can be used by other researchers to conduct further research related to the recognition of the Kazakh dactyl sign language, as well as by researchers, engaged in the development of the international sign language

Ultrasound imaging, as a non-invasive method which can recognize morphological changes of forearm muscles, has been gradually applied in the field of gesture recognition. In this paper, we proposed a B-mode ultrasound-based gesture database to provide a standard platform for the evaluation of different gesture recognition algorithms. The ultrasonic signals of the forearm according to the preset paradigm, and the classification results of a variety of features and classifiers were presented. Our results showed that the vast majority of feature algorithms and classifiers could achieve a satisfying recognition accuracy (about 90%). The analysis and results described herein are intended to provide a strong benchmark for the database and pave the way for subsequent researches.

Gesture recognition is a quite promising field in robotics and many Human-Computer Interaction (HCI) related areas. This research uses Microsoft® Kinect to capture the 3D position data of joints, and uses Gaussian Mixture Model (GMM) and Hidden Markov Model (HMM) to model full-body gestures. We propose a gesture recognition algorithm to segment gestures from real-time data flow, and finally achieved to recognize predefined full-body gestures in real-time. This proposed method gives a high recognition rate of 94.36%, indicating the capability of the new method.