Human-Machine Interface: Multiclass Classification by Machine Learning on 1D EOG Signals for the Control of an Omnidirectional Robot

Gaze Estimation Based on Convolutional Structure and Sliding Window-Based Attention Mechanism

The direction of human gaze is an important indicator of human behavior, reflecting the level of attention and cognitive state towards various visual stimuli in the environment. Convolutional neural networks have achieved good performance in gaze estimation tasks, but their global modeling capability is limited, making it difficult to further improve prediction performance. In recent years, transformer models have been introduced for gaze estimation and have achieved state-of-the-art performance. However, their slicing-and-mapping mechanism for processing local image patches can compromise local spatial information. Moreover, the single down-sampling rate and fixed-size tokens are not suitable for multiscale feature learning in gaze estimation tasks. To overcome these limitations, this study introduces a Swin Transformer for gaze estimation and designs two network architectures: a pure Swin Transformer gaze estimation model (SwinT-GE) and a hybrid gaze estimation model that combines convolutional structures with SwinT-GE (Res-Swin-GE). SwinT-GE uses the tiny version of the Swin Transformer for gaze estimation. Res-Swin-GE replaces the slicing-and-mapping mechanism of SwinT-GE with convolutional structures. Experimental results demonstrate that Res-Swin-GE significantly outperforms SwinT-GE, exhibiting strong competitiveness on the MpiiFaceGaze dataset and achieving a 7.5% performance improvement over existing state-of-the-art methods on the Eyediap dataset.

EOG Signal Classification with Wavelet and Supervised Learning Algorithms KNN, SVM and DT

The work carried out in this paper consists of the classification of the physiological signal generated by eye movement called Electrooculography (EOG). The human eye performs simultaneous movements, when focusing on an object, generating a potential change in origin between the retinal epithelium and the cornea and modeling the eyeball as a dipole with a positive and negative hemisphere. Supervised learning algorithms were implemented to classify five eye movements; left, right, down, up and blink. Wavelet Transform was used to obtain information in the frequency domain characterizing the EOG signal with a bandwidth of 0.5 to 50 Hz; training results were obtained with the implementation of K-Nearest Neighbor (KNN) 69.4%, a Support Vector Machine (SVM) of 76.9% and Decision Tree (DT) 60.5%, checking the accuracy through the Jaccard index and other metrics such as the confusion matrix and ROC (Receiver Operating Characteristic) curve. As a result, the best classifier for this application was the SVM with Jaccard Index.

Soft Wireless Headband Bioelectronics and Electrooculography for Persistent Human-Machine Interfaces

Recent advances in wearable technologies have enabled ways for people to interact with external devices, known as human-machine interfaces (HMIs). Among them, electrooculography (EOG), measured by wearable devices, is used for eye movement-enabled HMI. Most prior studies have utilized conventional gel electrodes for EOG recording. However, the gel is problematic due to skin irritation, while separate bulky electronics cause motion artifacts. Here, we introduce a low-profile, headband-type, soft wearable electronic system with embedded stretchable electrodes, and a flexible wireless circuit to detect EOG signals for persistent HMIs. The headband with dry electrodes is printed with flexible thermoplastic polyurethane. Nanomembrane electrodes are prepared by thin-film deposition and laser cutting techniques. A set of signal processing data from dry electrodes demonstrate successful real-time classification of eye motions, including blink, up, down, left, and right. Our study shows that the convolutional neural network performs exceptionally well compared to other machine learning methods, showing 98.3% accuracy with six classes: the highest performance till date in EOG classification with only four electrodes. Collectively, the real-time demonstration of continuous wireless control of a two-wheeled radio-controlled car captures the potential of the bioelectronic system and the algorithm for targeting various HMI and virtual reality applications.

Advances in Materials, Sensors, and Integrated Systems for Monitoring Eye Movements

Eye movements show primary responses that reflect humans' voluntary intention and conscious selection. Because visual perception is one of the fundamental sensory interactions in the brain, eye movements contain critical information regarding physical/psychological health, perception, intention, and preference. With the advancement of wearable device technologies, the performance of monitoring eye tracking has been significantly improved. It also has led to myriad applications for assisting and augmenting human activities. Among them, electrooculograms, measured by skin-mounted electrodes, have been widely used to track eye motions accurately. In addition, eye trackers that detect reflected optical signals offer alternative ways without using wearable sensors. This paper outlines a systematic summary of the latest research on various materials, sensors, and integrated systems for monitoring eye movements and enabling human-machine interfaces. Specifically, we summarize recent developments in soft materials, biocompatible materials, manufacturing methods, sensor functions, systems' performances, and their applications in eye tracking. Finally, we discuss the remaining challenges and suggest research directions for future studies.

eyeSay: Brain Visual Dynamics Decoding with Deep Learning & Edge Computing

Brain visual dynamics encode rich functional and biological patterns of the neural system, and if decoded, are of great promise for many applications such as intention understanding, cognitive load quantization and neural disorder measurement. We here focus on the understanding of the brain visual dynamics for the Amyotrophic lateral sclerosis (ALS) population, and propose a novel system that allows these so-called 'lock-in' patients to 'speak' with their brain visual movements. More specifically, we propose an intelligent system to decode the eye bio-potential signal, Electrooculogram (EOG), thereby understanding the patients' intention. We first propose to leverage a deep learning framework for automatic feature learning and classification of the brain visual dynamics, aiming to translate the EOG to meaningful words. We afterwards design and develop an edge computing platform on the smart phone, which can execute the deep learning algorithm, visualize the brain visual dynamics, and demonstrate the edge inference results, all in real-time. Evaluated on 4,500 trials of brain visual movements performed by multiple users, our novel system has demonstrated a high eye-word recognition rate up to 90.47%. The system is demonstrated to be intelligent, effective and convenient for decoding brain visual dynamics for ALS patients. This research thus is expected to greatly advance the decoding and understanding of brain visual dynamics, by leveraging machine learning and edge computing innovations.

Pattern Recognition of EMG Signals by Machine Learning for the Control of a Manipulator Robot

Human Machine Interfaces (HMI) principles are for the development of interfaces for assistance or support systems in physiotherapy or rehabilitation processes. One of the main problems is the degree of customization when applying some rehabilitation therapy or when adapting an assistance system to the individual characteristics of the users. To solve this inconvenience, it is proposed to implement a database of surface Electromyography (sEMG) of a channel in healthy individuals for pattern recognition through Neural Networks of contraction in the muscular region of the biceps brachii. Each movement is labeled using the One-Hot Encoding technique, which activates a state machine to control the position of an anthropomorphic manipulator robot and validate the response time of the designed HMI. Preliminary results show that the learning curve decreases when customizing the interface. The developed system uses muscle contraction to direct the position of the end effector of a virtual robot. The classification of Electromyography (EMG) signals is obtained to generate trajectories in real time by designing a test platform in LabVIEW.