Hand Keypoints Research Articles

Research on Mixed Reality Hand Interaction Technology in Flight Simulation Teaching

Abstract The integration of mixed reality simulation training in institutional education and teaching processes can yield significant quality benefits. Non-wearable interactive methods represent a more natural form of hand interaction within flight cockpits, covering various emerging fields such as mixed reality, computer vision, and human-computer interaction. This paper is guided by a “human-centered” approach to natural human-computer interaction and proposes a virtual hand mixed reality interaction scheme based on hand natural feature points. This study introduces a hand detection, segmentation, and recognition algorithm based on skin colour and key hand feature points. Initially, the algorithm uses the Otsu adaptive threshold algorithm in the YCbCr space for skin colour detection to accommodate varying image brightness levels. Subsequently, it employs a hand Keypoint model for rapid hand detection within skin-like areas while excluding interference from other regions. During the process of hand tracking, optimization prediction is performed using the particle filter algorithm combined with the artificial fish swarm algorithm to achieve tracking accuracy within six pixels in both horizontal and vertical directions. Gesture recognition involves extracting Fourier descriptive contour features of the hand bone structure through identification of feature points. Furthermore, this study combines the optimization capabilities of the artificial fish swarm algorithm to enhance support vector machine model parameter optimization for improved recognition rates. Testing involved 800 processed images as part of a test set for gesture recognition; only five commonly used gestures within flight cockpits were recognized and classified with an accuracy rate reaching up to 98%. Finally, this research presents virtual representations of common operational gestures including natural hand states, control stick manipulation, and throttle handling positions, button activations, and rotational movements via head-mounted displays or screens. The proposed mixed reality-based hand gesture interaction system addresses challenges related to complex background interference during hand detection in cockpit environments as well as issues arising from lighting disturbances. Additionally it resolves problems associated with insufficient particles leading to tracking failures while significantly enhancing registration effects during tracking processes. Moreover it mitigates issues related to Fourier descriptive features being influenced by background variations or changes in pose along with limitations pertaining to expressing singular forms of gestures thereby improving overall classification accuracy.

An Actively Vision-Assisted Low-Load Wearable Hand Function Mirror Rehabilitation System

The restoration of fine motor function in the hand is crucial for stroke survivors with hemiplegia to reintegrate into daily life and presents a significant challenge in post-stroke rehabilitation. Current mirror rehabilitation systems based on wearable devices require medical professionals or caregivers to assist patients in donning sensor gloves on the healthy side, thus hindering autonomous training, increasing labor costs, and imposing psychological burdens on patients. This study developed a low-load wearable hand function mirror rehabilitation robotic system based on visual gesture recognition. The system incorporates an active visual apparatus capable of adjusting its position and viewpoint autonomously, enabling the subtle monitoring of the healthy side’s gesture throughout the rehabilitation process. Consequently, patients only need to wear the device on their impaired hand to complete the mirror training, facilitating independent rehabilitation exercises. An algorithm based on hand key point gesture recognition was developed, which is capable of automatically identifying eight distinct gestures. Additionally, the system supports remote audio–video interaction during training sessions, addressing the lack of professional guidance in independent rehabilitation. A prototype of the system was constructed, a dataset for hand gesture recognition was collected, and the system’s performance as well as functionality were rigorously tested. The results indicate that the gesture recognition accuracy exceeds 90% under ten-fold cross-validation. The system enables operators to independently complete hand rehabilitation training, while the active visual system accommodates a patient’s rehabilitation needs across different postures. This study explores methods for autonomous hand function rehabilitation training, thereby offering valuable insights for future research on hand function recovery.

Attention-Based 2-D Hand Keypoints Localization

Attention-Based 2-D Hand Keypoints Localization

ML-Based Edge Node for Monitoring Peoples' Frailty Status.

The development of contactless methods to assess the degree of personal hygiene in elderly people is crucial for detecting frailty and providing early intervention to prevent complete loss of autonomy, cognitive impairment, and hospitalisation. The unobtrusive nature of the technology is essential in the context of maintaining good quality of life. The use of cameras and edge computing with sensors provides a way of monitoring subjects without interrupting their normal routines, and has the advantages of local data processing and improved privacy. This work describes the development an intelligent system that takes the RGB frames of a video as input to classify the occurrence of brushing teeth, washing hands, and fixing hair. No action activity is considered. The RGB frames are first processed by two Mediapipe algorithms to extract body keypoints related to the pose and hands, which represent the features to be classified. The optimal feature extractor results from the most complex Mediapipe pose estimator combined with the most complex hand keypoint regressor, which achieves the best performance even when operating at one frame per second. The final classifier is a Light Gradient Boosting Machine classifier that achieves more than 94% weighted F1-score under conditions of one frame per second and observation times of seven seconds or more. When the observation window is enlarged to ten seconds, the F1-scores for each class oscillate between 94.66% and 96.35%.

Multimodal Gesture Recognition with Spatio-Temporal Features Fusion Based on YOLOv5 and MediaPipe

As a natural, intuitive and easy-to-learn mode of interaction, gesture plays an important role in communication. Hand detection, containing multimodal information, includes static and dynamic detection and involves intricate spatial relationship problems such as different hand sizes, complex joints, occlusion and self-occlusion. This study focused on a multimodal hand gesture recognition system based on YOLOv5 and MediaPipe with fused spatio-temporal features. First, the Mediapipe and OpenCV libraries were employed to implement hand keypoint detection. Subsequently, the human–computer interaction (HCI) of volume control was realized by identifying the distance between thumb and index. Finally, model training was conducted based on the YOLOv5 algorithm, and the recognition of different gesture categories was realized. The performance was evaluated and compared through YOLOv5s, YOLOv5m, and YOLOv5l. The gesture recognition system interface visualization was achieved through pyqt5. Experiments show that the average detection accuracy of the model is 99.4% and the recognition speed is around 0.2[Formula: see text]s.

Expressive Sign Language Recognition using CNN

A major shortcoming in our society is a social barrier between the differently abled members of the society and the abled folks. One of the most important aspects of human beings, being regarded as social animals, is in fact communication. Communication is also a major obstacle faced by the hearing and vocal disabilities people. This inability to communicate leads to frequent problems and hinders the daily activities of a person with hearing and vocal disabilities. The underlying reason for this disparity is that abled folks don’t learn and aren’t taught Sign Language which is the main means of communication for a person with hearing and vocal disabilities. Thus, abled folks are incapable of having a normally fluent conversation with these different sections of the society. Consequently, in a verbal exchange among hearing and speech impaired individuals and an able person the convenience of communique and consequently the consolation degree is hampered. So, in our project, we have proposed a cost-efficient solution to overcome this communication barrier. This solution can be easily used by everyone and can also be, with some modifications, made to work on most platforms which have a camera module. Our approach uses the integrated camera module to capture real time hand gestures based on hand key points or landmarks and the algorithm using machine learning techniques, displays the alphabet that the gesture is representing. KEY WORDS: Convolution Neural Network, Transfer Learning, Deep Learning, Artificial Neural Network, Support Vector Machines

Human Action Recognition and Note Recognition: A Deep Learning Approach Using STA-GCN.

Human action recognition (HAR) is growing in machine learning with a wide range of applications. One challenging aspect of HAR is recognizing human actions while playing music, further complicated by the need to recognize the musical notes being played. This paper proposes a deep learning-based method for simultaneous HAR and musical note recognition in music performances. We conducted experiments on Morin khuur performances, a traditional Mongolian instrument. The proposed method consists of two stages. First, we created a new dataset of Morin khuur performances. We used motion capture systems and depth sensors to collect data that includes hand keypoints, instrument segmentation information, and detailed movement information. We then analyzed RGB images, depth images, and motion data to determine which type of data provides the most valuable features for recognizing actions and notes in music performances. The second stage utilizes a Spatial Temporal Attention Graph Convolutional Network (STA-GCN) to recognize musical notes as continuous gestures. The STA-GCN model is designed to learn the relationships between hand keypoints and instrument segmentation information, which are crucial for accurate recognition. Evaluation on our dataset demonstrates that our model outperforms the traditional ST-GCN model, achieving an accuracy of 81.4%.

A Hand Gesture-Operated System for Rehabilitation Using an End-to-End Detection Framework

We propose a hand gesture-operated system as an AI application to relieve discomfort and restore function in hand and arm movements caused by injuries and nerve and muscle complications. The system trains patients with hand exercises, such as performing hand gestures accurately, traversing within specified bounds, and operating a hand gesture calculator. However, the system requires accurate hand gesture detection, which is impeded by background clutter and variations in illumination and in the hand's size and angle. To address this, we developed a robust hand detection module that uses a single-stage transformer deep network. The transformer network encodes global information and uses bipartite matching to reduce the frequency of spurious detections. It drives a regression head and a classification head to localize the hand gesture in a bounding box and assign it a class label. Hand keypoints are also detected to support drawing, path traversal, and calculator use. The approach is evaluated on two benchmark datasets: OUHANDS and NUS. The method yields 89.6% accuracy for OUHANDS and 100% for NUS. These results indicate that precise hand detection can support a robust system for rehabilitation through hand exercises. Our experiments confirm that the users' hand function progressively improved.

Isolated Arabic Sign Language Recognition Using a Transformer-based Model and Landmark Keypoints

Pose-based approaches for sign language recognition provide light-weight and fast models that can be adopted in real-time applications. This article presents a framework for isolated Arabic sign language recognition using hand and face keypoints. We employed MediaPipe pose estimator for extracting the keypoints of sign gestures in the video stream. Using the extracted keypoints, three models were proposed for sign language recognition: Long-Term Short Memory, Temporal Convolution Networks, and Transformer-based models. Moreover, we investigated the importance of non-manual features for sign language recognition systems and the obtained results showed that combining hand and face keypoints boosted the recognition accuracy by around 4% compared with only hand keypoints. The proposed models were evaluated on Arabic and Argentinian sign languages. Using the KArSL-100 dataset, the proposed pose-based Transformer achieved the highest accuracy of 99.74% and 68.2% in signer-dependent and -independent modes, respectively. Additionally, the Transformer was evaluated on the LSA64 dataset and obtained an accuracy of 98.25% and 91.09% in signer-dependent and -independent modes, respectively. Consequently, the pose-based Transformer outperformed the state-of-the-art techniques on both datasets using keypoints from the signer’s hands and face.

Multiscale Attention-Based Hand Keypoint Detection

Multiscale Attention-Based Hand Keypoint Detection

Sign Language Recognition with Multimodal Sensors and Deep Learning Methods

Sign language recognition is essential in hearing-impaired people’s communication. Wearable data gloves and computer vision are partially complementary solutions. However, sign language recognition using a general monocular camera suffers from occlusion and recognition accuracy issues. In this research, we aim to improve accuracy through data fusion of 2-axis bending sensors and computer vision. We obtain the hand key point information of sign language movements captured by a monocular RGB camera and use key points to calculate hand joint angles. The system achieves higher recognition accuracy by fusing multimodal data of the skeleton, joint angles, and finger curvature. In order to effectively fuse data, we spliced multimodal data and used CNN-BiLSTM to extract effective features for sign language recognition. CNN is a method that can learn spatial information, and BiLSTM can learn time series data. We built a data collection system with bending sensor data gloves and cameras. A dataset was collected that contains 32 Japanese sign language movements of seven people, including 27 static movements and 5 dynamic movements. Each movement is repeated 10 times, totaling about 112 min. In particular, we obtained data containing occlusions. Experimental results show that our system can fuse multimodal information and perform better than using only skeletal information, with the accuracy increasing from 68.34% to 84.13%.

Automatic Multiview Alignment of RGB-D Range Maps of Upper Limb Anatomy.

Digital representations of anatomical parts are crucial for various biomedical applications. This paper presents an automatic alignment procedure for creating accurate 3D models of upper limb anatomy using a low-cost handheld 3D scanner. The goal is to overcome the challenges associated with forearm 3D scanning, such as needing multiple views, stability requirements, and optical undercuts. While bulky and expensive multi-camera systems have been used in previous research, this study explores the feasibility of using multiple consumer RGB-D sensors for scanning human anatomies. The proposed scanner comprises three Intel® RealSenseTM D415 depth cameras assembled on a lightweight circular jig, enabling simultaneous acquisition from three viewpoints. To achieve automatic alignment, the paper introduces a procedure that extracts common key points between acquisitions deriving from different scanner poses. Relevant hand key points are detected using a neural network, which works on the RGB images captured by the depth cameras. A set of forearm key points is meanwhile identified by processing the acquired data through a specifically developed algorithm that seeks the forearm's skeleton line. The alignment process involves automatic, rough 3D alignment and fine registration using an iterative-closest-point (ICP) algorithm expressly developed for this application. The proposed method was tested on forearm scans and compared the results obtained by a manual coarse alignment followed by an ICP algorithm for fine registration using commercial software. Deviations below 5 mm, with a mean value of 1.5 mm, were found. The obtained results are critically discussed and compared with the available implementations of published methods. The results demonstrate significant improvements to the state of the art and the potential of the proposed approach to accelerate the acquisition process and automatically register point clouds from different scanner poses without the intervention of skilled operators. This study contributes to developing effective upper limb rehabilitation frameworks and personalized biomedical applications by addressing these critical challenges.

3D hand mesh recovery through inverse kinematics from a monocular RGB image

Recovering 3D hand mesh from a monocular RGB image has a wide range of application scenarios such as VR/AR. The parametric hand model provides a good geometric prior to the shape of hand, and is commonly used to recover the 3D hand mesh. However, the rotation parameters of hand model are not easy to learn, which influences the accuracy of model-based methods. To address this problem, we take advantage of the inverse kinematic chains of hand to derive an analytical method, which can convert hand joint locations into rotation parameters. By integrating such analytical method into the neural network, we propose an end-to-end learnable model named IKHand to recover the 3D hand mesh. IKHand comprises detection module and mesh generation module. Detection module predicts the 3D hand keypoints while mesh generation module takes these keypoints to generate the 3D hand mesh. Experimental results show that our proposed method can generate impressive and robust 3D hand meshes under several challenging conditions, and can achieve competitive results on FreiHAND dataset as well as RHD.

Bias-Compensated Integral Regression for Human Pose Estimation.

In human and hand pose estimation, heatmaps are a crucial intermediate representation for a body or hand keypoint. Two popular methods to decode the heatmap into a final joint coordinate are via an argmax, as done in heatmap detection, or via softmax and expectation, as done in integral regression. Integral regression is learnable end-to-end, but has lower accuracy than detection. This paper uncovers an induced bias from integral regression that results from combining the softmax and the expectation operation. This bias often forces the network to learn degenerately localized heatmaps, obscuring the keypoint's true underlying distribution and leads to lower accuracies. Training-wise, by investigating the gradients of integral regression, we show that the implicit guidance of integral regression to update the heatmap makes it slower to converge than detection. To counter the above two limitations, we propose Bias Compensated Integral Regression (BCIR), an integral regression-based framework that compensates for the bias. BCIR also incorporates a Gaussian prior loss to speed up training and improve prediction accuracy. Experimental results on both the human body and hand benchmarks show that BCIR is faster to train and more accurate than the original integral regression, making it competitive with state-of-the-art detection methods.

Consistent 3D Hand Reconstruction in Video via Self-Supervised Learning.

We present a method for reconstructing accurate and consistent 3D hands from a monocular video. We observe that the detected 2D hand keypoints and the image texture provide important cues about the geometry and texture of the 3D hand, which can reduce or even eliminate the requirement on 3D hand annotation. Accordingly, in this work, we propose S2HAND, a self-supervised 3D hand reconstruction model, that can jointly estimate pose, shape, texture, and the camera viewpoint from a single RGB input through the supervision of easily accessible 2D detected keypoints. We leverage the continuous hand motion information contained in the unlabeled video data and explore S2HAND(V), which uses a set of weights shared S2HAND to process each frame and exploits additional motion, texture, and shape consistency constrains to obtain more accurate hand poses, and more consistent shapes and textures. Experiments on benchmark datasets demonstrate that our self-supervised method produces comparable hand reconstruction performance compared with the recent full-supervised methods in single-frame as input setup, and notably improves the reconstruction accuracy and consistency when using the video training data.

Recognizing gestures of the ukrainian dactylic alphabet

Sign language is a visual way of communicating used by people who are deaf or hard of hearing. It involves handshapes, facial expressions, and body movements to convey meaning. Sign language helps the deaf community interact with each other and the hearing world, allowing them to participate fully in society. According to the WHO (World Health Organization) over 5 % of the world’s population – or 430 million people — experience problems with hearing. More than 44,000 people with hearing impairments are registered with the Ukrainian Society of the Deaf, an all-Ukrainian public organization for the disabled. Therefore, it is extremely important to develop new software, available to the public, that would allow quickly and effectively learn and understand sign language. This work aims to review gesture recognition techniques and develop a system for detecting and classifying gestures of the Ukrainian dactylic alphabet. Two main approaches to gesture recognition, glove-based and computer vision-based (CV), are explained, with the latter being preferred due to its flexibility and widespread usage. The text elaborates on deep learning-based approaches, particularly LSTM networks, and the advantages they offer in automatically learning features from raw image data. The process of creating a dataset for training the gesture classification model is described, which involves recording videos of hand gestures and extracting keypoints using Google MediaPipe. The mo­del training phase is detailed, covering the architecture of the LSTM-based classifier, optimization algorithms, and loss functions. The resulting model achieves an accuracy of 98.4% on the test dataset. A program for real-time gesture recognition is developed using Python and relevant libraries. The program utilizes a webcam feed to detect and classify hand gestures, displaying the top three predicted letters of the Ukrainian dactylic alphabet. The scientific novelty of the obtained results: the paper presents a method that utilizes hand keypoints for recognizing hand gestures of the Ukrainian dactyl alphabet. Also, as part of the development of the gesture recognition system, a data set was collected, where each gesture corresponds to 50 videos of 65 frames. The practical significance of the results obtained: the model obtained as a result of the study can be used to interpret the gestures of the Ukrainian dactylic alphabet. The dataset collected for training this model can be used in other works to train or validate similar models. The paper might be of use to the ones who are interested in developing similar systems for gesture recognition.

Multi-Stream General and Graph-Based Deep Neural Networks for Skeleton-Based Sign Language Recognition

Sign language recognition (SLR) aims to bridge speech-impaired and general communities by recognizing signs from given videos. However, due to the complex background, light illumination, and subject structures in videos, researchers still face challenges in developing effective SLR systems. Many researchers have recently sought to develop skeleton-based sign language recognition systems to overcome the subject and background variation in hand gesture sign videos. However, skeleton-based SLR is still under exploration, mainly due to a lack of information and hand key point annotations. More recently, researchers have included body and face information along with hand gesture information for SLR; however, the obtained performance accuracy and generalizability properties remain unsatisfactory. In this paper, we propose a multi-stream graph-based deep neural network (SL-GDN) for a skeleton-based SLR system in order to overcome the above-mentioned problems. The main purpose of the proposed SL-GDN approach is to improve the generalizability and performance accuracy of the SLR system while maintaining a low computational cost based on the human body pose in the form of 2D landmark locations. We first construct a skeleton graph based on 27 whole-body key points selected among 67 key points to address the high computational cost problem. Then, we utilize the multi-stream SL-GDN to extract features from the whole-body skeleton graph considering four streams. Finally, we concatenate the four different features and apply a classification module to refine the features and recognize corresponding sign classes. Our data-driven graph construction method increases the system’s flexibility and brings high generalizability, allowing it to adapt to varied data. We use two large-scale benchmark SLR data sets to evaluate the proposed model: The Turkish Sign Language data set (AUTSL) and Chinese Sign Language (CSL). The reported performance accuracy results demonstrate the outstanding ability of the proposed model, and we believe that it will be considered a great innovation in the SLR domain.

A Computational Approach to Hand Pose Recognition in Early Modern Paintings.

Hands represent an important aspect of pictorial narration but have rarely been addressed as an object of study in art history and digital humanities. Although hand gestures play a significant role in conveying emotions, narratives, and cultural symbolism in the context of visual art, a comprehensive terminology for the classification of depicted hand poses is still lacking. In this article, we present the process of creating a new annotated dataset of pictorial hand poses. The dataset is based on a collection of European early modern paintings, from which hands are extracted using human pose estimation (HPE) methods. The hand images are then manually annotated based on art historical categorization schemes. From this categorization, we introduce a new classification task and perform a series of experiments using different types of features, including our newly introduced 2D hand keypoint features, as well as existing neural network-based features. This classification task represents a new and complex challenge due to the subtle and contextually dependent differences between depicted hands. The presented computational approach to hand pose recognition in paintings represents an initial attempt to tackle this challenge, which could potentially advance the use of HPE methods on paintings, as well as foster new research on the understanding of hand gestures in art.

Fast and Accurate ROI Extraction for Non-Contact Dorsal Hand Vein Detection in Complex Backgrounds Based on Improved U-Net.

In response to the difficulty of traditional image processing methods to quickly and accurately extract regions of interest from non-contact dorsal hand vein images in complex backgrounds, this study proposes a model based on an improved U-Net for dorsal hand keypoint detection. The residual module was added to the downsampling path of the U-Net network to solve the model degradation problem and improve the feature information extraction ability of the network; the Jensen-Shannon (JS) divergence loss function was used to supervise the final feature map distribution so that the output feature map tended to Gaussian distribution and improved the feature map multi-peak problem; and Soft-argmax is used to calculate the keypoint coordinates of the final feature map to realize end-to-end training. The experimental results showed that the accuracy of the improved U-Net network model reached 98.6%, which was 1% better than the original U-Net network model; the improved U-Net network model file was only 1.16 M, which achieved a higher accuracy than the original U-Net network model with significantly reduced model parameters. Therefore, the improved U-Net model in this study can realize dorsal hand keypoint detection (for region of interest extraction) for non-contact dorsal hand vein images and is suitable for practical deployment in low-resource platforms such as edge-embedded systems.

Robot Programming from a Single Demonstration for High Precision Industrial Insertion

We propose a novel approach for robotic industrial insertion tasks using the Programming by Demonstration technique. Our method allows robots to learn a high-precision task by observing human demonstration once, without requiring any prior knowledge of the object. We introduce an Imitated-to-Finetuned approach that generates imitated approach trajectories by cloning the human hand's movements and then fine-tunes the goal position with a visual servoing approach. To identify features on the object used in visual servoing, we model object tracking as the moving object detection problem, separating each demonstration video frame into the moving foreground that includes the object and demonstrator's hand and the static background. Then a hand keypoints estimation function is used to remove the redundant features on the hand. The experiment shows that the proposed method can make robots learn precision industrial insertion tasks from a single human demonstration.