Korea University · Neuroscience
Professor Young-Eun Lee's research lab specializes in brain-computer interfaces (BCIs) and neural signal processing, with a focus on translating brain activity—particularly imagined and overt speech—into speech output using non-invasive electroencephalography (EEG). The lab develops advanced deep learning models, such as transformer-based architectures and hybrid neural networks, to decode complex neural patterns and improve the accuracy and naturalness of speech reconstruction. A key emphasis is on robustness in real-world conditions, including artifact removal from noisy EEG signals and adapting models trained on spoken speech to imagined speech. The lab also explores applications in mobile and wearable BCI systems, integrating multimodal data from EEG, inertial sensors, and eye tracking for dynamic environments.
Figures are computed from collected data and may differ slightly.
Transformers are groundbreaking architectures that have changed a flow of deep learning, and many high-performance models are developing based on transformer architectures. Transformers implemented only with attention with encoder-decoder structure following seq2seq without using RNN, but had better performance than RNN. Herein, we investigate the decoding technique for electroencephalography (EEG) composed of self-attention module from transformer architecture during imagined speech and overt s
Translating imagined speech from human brain activity into voice is a challenging and absorbing research issue that can provide new means of human communication via brain signals. Efforts to reconstruct speech from brain activity have shown their potential using invasive measures of spoken speech data, but have faced challenges in reconstructing imagined speech. In this paper, we propose NeuroTalk, which converts non-invasive brain signals of imagined speech into the user's own voice. Our model
Recently, practical brain-computer interfaces (BCIs) have been widely investigated for detecting human intentions in real world. However, performance differences still exist between the laboratory and the real world environments. One of the main reasons for such differences comes from the user's unstable physical states (e.g., human movements are not strictly controlled), which produce unexpected signal artifacts. Hence, to minimize the performance degradation of electroencephalography (EEG)-bas
We present a mobile dataset obtained from electroencephalography (EEG) of the scalp and around the ear as well as from locomotion sensors by 24 participants moving at four different speeds while performing two brain-computer interface (BCI) tasks. The data were collected from 32-channel scalp-EEG, 14-channel ear-EEG, 4-channel electrooculography, and 9-channel inertial measurement units placed at the forehead, left ankle, and right ankle. The recording conditions were as follows: standing, slow
Surgeons cannot directly touch the patient’s tissue in robot-assisted minimally invasive procedures. Instead, they must palpate using instruments inserted into the body through trocars. This way of operating largely prevents surgeons from using haptic cues to localize visually undetectable structures such as tumors and blood vessels, motivating research on direct and indirect force sensing. We propose an indirect force-sensing method that combines monocular images of the operating field with mea
We propose three novel methods to evaluate a distance function for robotic motion planning based on semiinfinite programming (SIP) framework; these methods include golden section search (GSS), conservative advancement (CA) and a hybrid of GSS and CA. The distance function can have a positive and a negative value, each of which corresponds to the Euclidean distance and penetration depth, respectively. In our approach, each robot's link is approximated and bounded by a capsule shape, and the dista
Abstract We present simple and fast parallel proximity algorithms for rigid polygonal models. Given two polygon‐soup models in space, if they overlap, our algorithm can find all the intersected primitives between them; otherwise, it reports their Euclidean minimum distance. Our algorithm is performed in a parallel fashion and shows scalable performance in terms of the number of available computing cores. The key ingredient of our algorithm is a simple load‐balancing metric based on the penetrati
We present a novel algorithm to compute a gradient-continuous penetration depth (PhongPD) between two interpenetrated polygonal models. Our penetration depth (PD) formulation ensures separating the intersected models by translation, and the amount of such translation is close to an optimal motion to resolve interpenetration in most cases. In order to achieve the gradient-continuity in our algorithm, we interpolate tangent planes continuously over the contact space and then perform a projection a
Recently, practical brain-computer interface is actively carried out, especially, in an ambulatory environment. However, the electroencephalography signals are distorted by movement artifacts and electromyography signals in ambulatory condition, which make hard to recognize human intention. In addition, as hardware issues are also challenging, ear-EEG has been developed for practical brain-computer interface and is widely used. However, ear-EEG still contains contaminated signals. In this paper,
Recent advances in brain-computer interface (BCI) technology, particularly based on generative adversarial networks (GAN), have shown great promise for improving decoding performance for BCI. Within the realm of BCI, GANs find application in addressing many areas. They serve as a valuable tool for data augmentation, which can solve the challenge of limited data availability, and synthesis, effectively expanding the dataset and creating novel data formats, thus enhancing the robustness and adapta
We propose a novel method adaptive subdivision (AS) to evaluate the distance function for moving general polygonal models. The distance function can have a positive and a negative value, each of which corresponds to the Euclidean distance and penetration depth, respectively. In our approach, the distance between a pair of objects can be evaluated along any time interval of the object's trajectory; therefore it is called “continuous”, and a minimum of the continuous distance (MCD) is determined f
Brain imaging studies of human speech are an active and intriguing research topic that is generating novel ways of communication through human brain signals. Efforts to generate voice from human neural activity have demonstrated the potential based on invasive measurements of speech, but have encountered difficulties in recreating data from imagined speech. Here, we propose NeuroTalk, which non-invasively converts brain signals from spoken and imagined speech to voice. The proposed framework is
Recently, practical brain-computer interface is actively carried out,\nespecially, in an ambulatory environment. However, the electroencephalography\nsignals are distorted by movement artifacts and electromyography signals in\nambulatory condition, which make hard to recognize human intention. In\naddition, as hardware issues are also challenging, ear-EEG has been developed\nfor practical brain-computer interface and is widely used. However, ear-EEG\nstill contains contaminated signals. In this
Open papers in the app to read, cite, and organize with AI.
