Content uploaded by Ayoub Al-Hamadi
Author content
All content in this area was uploaded by Ayoub Al-Hamadi
Content may be subject to copyright.
A preview of the PDF is not available
Posture and Gesture Recognition for Human-Computer Interaction
Abstract
1
... The hand gestures used in the SL communication are the spatio-temporal patterns. These spatio-temporal hand patterns may be static or dynamic or both static and dynamic [1]. While the still hand shapes or postures are used to represent the message in the static SL, a series of hand postures are used to convey the message in dynamic SL. ...
... Figure 1 shows these twenty-six alphabets [9]. Out of these twenty-six alphabets, letters J and Z require the movement of hand for representation [1,10]. Therefore, these letters are termed as dynamic hand gestures. ...
... Depending on the length of usage, these interactions can burden users with physical constraints, like pressing power for a device and calibration, or using touch-screen and accelerometer-embedded wearable devices. Research must move away from these potential and existing constraints in favor of free-space interfaces (or "non-touch-based interfaces"): this will make gestures appropriate for control tasks in human-computer interaction (HCI); and advance human-computer communication by narrowing the distance between HCI performance and human-human interaction (Lindeberg, Sundblad, Bretzner, 2001;Elmezain, Al-Hamadi, Rashid 2009). Research must also consider user health conditions and long-term usage, which are some of the factors that contribute to the appropriateness of gestures for specific tasks within such interfaces. ...
... Finding and identifying a set of intuitive yet appropriate gestures for a system remains challenging. Across different users and even for one user, the same gesture varies in shape, trajectory, and duration (Elmezain et al. 2009). Gestures also vary in meanings from each participant (Morris et al., 2014). ...
Along with the number of technologies that have been introduced over a few years ago, gesture-based human-computer interactions are becoming the new phase in encompassing the creativity and abilities for users to communicate and interact with devices. Current gesture recognition systems lack sufficient consideration for the impact they have on user’s health and preference during usage. Thus, determining and defining low-stress and intuitive gestures for user interaction is necessary for improving human-computer communication and increasing the usability of gesture-driven devices. To find the relationship, if any, between stress and intuitive gestures, a Galvanic Skin Response instrument was used and worn by fifteen participants (10 females, 5 males) as a primary measurement of stress. The participants also engaged in creating and performing their own gestures for specified tasks, which include “activation of the display,” “scroll,” “page,” “selection,” “undo,” and “return to main menu.” They were then asked to repeat their gestures for around ten seconds each, giving them time and further insight into how their gestures would be appropriate or not for them and any given task. Surveys were given to the users at different times: one after they had defined their gestures and another after they had repeated their gestures. It is found that they have the tendency to rank their gestures based on comfort, intuition, and the ease of communication. On an average of more than half of the times, users were also found to change the gestures they have preferred in the beginning after practicing their gestures. Out of the user-ranked gestures, health-efficient gestures, given that the participants’ rankings were based on comfort and intuition, were chosen in regards to the highest ranked gestures: a single or double taps, single or double claps, circle arm, slide, wave, swipe, point, hold up numbers, draw in mid-air, and form a shape.
... In terms of domain difficulty, the SVM Technique can be reorganized using dualities features, margins, and kernel types. SVM is used to solve issues such as nonlinearity and local minima [39,40]. SVM can also differentiate between labels and properly separate them. ...
The total number of discovered plant species is increasing yearly worldwide. Plant species differ from one region to another. Some of these discovered plant species are beneficial while others might be poisonous. Computer vision techniques can be an effective way to classify plant species and predict their poisonous status. However, the lack of comprehensive datasets that include not only plant images but also plant species’ scientific names, description, poisonous status, and local name make the issue of poisonous plants species prediction a very challenging issue. In this paper, we propose a hybrid model relying on transformers models in conjunction with support vector machine for plant species classification and poisonous status prediction. First, six different Convolutional Neural Network (CNN) architectures are used to determine which produces the best results. Second, the features are extracted using six different CNNs and then optimized and employed to Support Vector Machine (SVM) for testing. To prove the feasibility and benefits of our proposed approach, we used a real case study namely, plant species discovered in the Arabian Peninsula. We have gathered a dataset that contains 2500 images of 50 different Arabic plant species and includes plants images, plant species scientific name, description, local name, and poisonous status. This study on the types of Arabic plants species will help in the reduction of the number of poisonous plants victims and their negative impact on the individual and society. The results of our experiments for the CNN approach in conjunction SVM are favorable where the classifier scored 0.92, 0.94, and 0.95 in accuracy, precision, and F1-Score respectively.
... Mapping of the input data to a feature space through kernel function. Adapted fromElmezain et al. (2009). ...
Support Vector Machine (SVM) is a supervised learning algorithm widely used in data classification problems. However, the quality of the solution is related to the chosen kernel function, and the adjustment of its parameters. In the present study we compare a genetic algorithm (GA), a particle swarm optimization(PSO), and the grid-search in setting the parameters
and C of SVM. After running some experimental tests based on the prediction of protein function, it is concluded that all algorithms are suitable to set the SVM parameters efficiently, yet grid-search runs up to 6 times faster than GA and 30 times faster than PSO.
... By transforming the original space (left) into a space of increased dimension (right) the two classes "circle" and "square" become linearly separable. Adapted from [32]. ...
Linear subspace projections are an important technique to reduce the dimensionality of data for automatic classification. Especially for large-scale and on-line systems, e.g. gesture recognition applications, this is important to guarantee near real-time processing. The linear subspace projections, however, fail if the classes are not linearly separable. Kernel methods, in contrast, have been widely applied to linear classification algorithms to solve problems of non-linearly separable classes. This technique, however, increases the computational complexity by introducing the evaluation of a possibly non-linear function. Here, we extend a linear subspace projection that has been applied to large-scale systems using a kernel function. The method is evaluated on Fisher's Iris dataset and a recorded gesture dataset. The results indicate that the proposed method yields an increased accuracy at a subspace of lower dimension while achieving a similar runtime at a subspace of the same dimension. The proposed method is thus expected to work well with online systems.
... By transforming the original space (left) into a space of increased dimension (right) the two classes circle and square become linearly separable. Adapted fromElmezain et al. (2009). ...
The automatic recognition of gestures is important in a variety of applications, e.g. human-machine-interaction.
Commonly, different individuals execute gestures in a slightly different manner and thus a fully
labelled dataset is not available while unlabelled data may be acquired from an on-line stream. Consequently,
gesture recognition systems should be able to be trained in a semi-supervised learning scenario. Additionally,
real-time systems and large-scale data require a dimensionality reduction of the data to reduce the processing
time. This is commonly achieved by linear subspace projections. Most of the gesture data sets, however, are
non-linearly distributed. Hence, linear sub-space projection fails to separate the classes. We propose an extension
to linear subspace projection by applying a non-linear transformation to a space of higher dimensional
after the linear subspace projection. This mapping, however, is not explicitly evaluated but implicitly used by
a kernel function. The kernel nearest class mean (KNCM) classifier is shown to handle the non-linearity as
well as the semi-supervised learning scenario. The computational expense of the non-linear kernel function
is compensated by the dimensionality reduction of the previous linear subspace projection. The method is
applied to a gesture dataset comprised of 3D trajectories. The trajectories were acquired using the Kinect
sensor. The results of the semi-supervised learning show high accuracies that approach the accuracy of a fully
supervised scenario already for small dimensions of the subspace and small training sets. The accuracy of the
semi-supervised KNCM exceeds the accuracy of the original nearest class mean classifier in all cases.
... Main methods for doing so utilize several types of visual features of the hand, such as skin color, anatomical form and specific motion patterns of the hand. Skin color segmentation is a common method for detecting the hand because of its fast implementation using lookup tables, but sometimes the hand can be confused with background objects and shadows [ 30 ]. Several hand segmentation algorithms [ 31 ], [ 32 ] and methods are described in the literature [33], [34]. It is getting common to simply use a depth threshold to isolate the hands from the background [28]. ...
When designing products, networked computers are increasingly used to facilitate the collaboration among team members from remote locations. Design visualization plays a critical role in understanding the design concepts shared by the design team members. CAD systems have 3D visualization capabilities that are designed to help users to understand complex structures easily and to design better products. However, 3D visualization on a 2D screen has limitations in communicating complex structures. Furthermore, gestures play a significant role in face-to-face communication but are missing in remote collaboration. Object manipulation in 3D using gestures in the real physical space without a cumbersome hand-held peripheral device may ease the visualization and understanding of the concept model. Today, peripheral devices for human-computer interfaces are not only becoming more advanced but also less expensive. Some of the game controllers are now equipped with depth cameras (RGB-D cameras) and have a great potential for complementing or even replacing the traditional keyboard and mouse interface with hand or body gestures. In this paper, new low-cost mixed reality devices for 3D user inputs and visual outputs are investigated and their possible uses with CAD systems in the process of collaborative design are discussed.
In this study, a hand tracking system built based on two Kinect sensors has been used to track the position and orientation of the user’s hands. In this system, CAD models can be manipulated and disassembled using hand gestures. The new user interface provides for a user-friendly interaction between designer and CAD system.
... In order to have a high computational speed, the detection component analyzes each frame obtained from the video device by using the pattern classifier until a hand is detected. Starting from the next frame, the hand detected area is being tracked using Camshift algorithm [10], because the tracking is less time consuming than the detection algorithm. The tracking algorithm estimates for each pixel in the following frames the probability to be in the hand. ...
St. Gh. Pentiuc, E. G. Craciun, L. Grisoni. interface for Gestural interaction in Virtual Reality Environments // Electronics and Electrical Engineering. - Kaunas: Technologija, 2012. - No. 5(121). - P. 97-100. The main contribution of the paper is the concept, design and experimentation of an interface able to satisfy all the main function that the user needs to interact with VR environments. The interface was aimed to supply an unified gestural paradigm. It deals with video acquisition, image processing, features extraction and pattern classification. The results of the experiments with the posture and gesture recognizing application validated our approach. Ill. 3, bibl. 14 (in English; abstracts in English and Lithuanian).
One of the main research areas in the field of musical human–AI interactivity is how to incorporate expressiveness into interactive digital musical instruments (DMIs). In this study we analyzed gestures rooted in expressiveness by using AI techniques that can enhance the mapping stage of multitouch DMIs. This approach not only considers the geometric information of various gestures but also incorporates expressiveness, which is a crucial element of musicality. Our focus is specifically on multitouch DMIs, and we use expressive descriptors and a fuzzy logic model to mathematically analyze performers' finger movements. By incorporating commonly used features from the literature and adapting some of Rudolf Laban's descriptors—originally intended for full-body analysis—to finger-based multitouch systems, we aim to enrich the mapping process. To achieve this, we developed an AI algorithm based on a fuzzy control system that takes these descriptors as inputs and maps them to synthesis variables. This tool empowers DMI designers to define their own mapping rules based on expressive gestural descriptions, employing musical metaphors in a simple and intuitive way. Through a user evaluation, we demonstrate the effectiveness of our approach in capturing and representing gestural expressiveness in the case of multitouch DMIs.
In this paper, a method is proposed to maximize the accuracy during the feature extraction stage in a real time system for hand gesture recognition by escalating the number of parameters of the feature set for support vector machine. Numerous former researches utilized hu moments but they didn't correspond to the complete description of an image, and was suitable only for giving very rough estimation of possible match. Thus matching performance was not acceptable for image retrieval. On the other hand, the accuracy of the support vector machine (SVM) depends on the number of support vectors. Hence adding features that significantly improve the splitting probability of training images decrease the number of support vectors and improves the performance of the SVM. Therefore to enhance the harmonizing of images, together with hu moments, edge histogram descriptor and circularity shape parameter is used to compose the feature vector. Experiments on series of test images show that the proposed method yields better matching performance. Integrated feature based approach to hand gesture recognition has been tested over 23 gestures and it gave promising results.
ResearchGate has not been able to resolve any references for this publication.