A5100668167- Image and Video Quality Assessment
- Advanced Image Processing Techniques
- Advanced Vision and Imaging
- Adversarial Robustness in Machine Learning
- Video Coding and Compression Technologies
- Anomaly Detection Techniques and Applications
- Image Processing Techniques and Applications
- EEG and Brain-Computer Interfaces
- Visual Attention and Saliency Detection
- Image and Signal Denoising Methods
- Advanced Data Compression Techniques
- Speech and Audio Processing
- Neural Networks and Applications
- Advanced Image and Video Retrieval Techniques
- Human Pose and Action Recognition
- Music and Audio Processing
- Image Enhancement Techniques
- Neural dynamics and brain function
- Video Analysis and Summarization
- Generative Adversarial Networks and Image Synthesis
- Advanced Neural Network Applications
- Speech Recognition and Synthesis
- Emotion and Mood Recognition
- Color perception and design
- Visual perception and processing mechanisms
Pusan National University
2019-2025
Yonsei University
2015-2024
Electronics and Telecommunications Research Institute
2024
Deutsches Zentrum für Luft- und Raumfahrt e. V. (DLR)
2020-2023
Kwangwoon University
2020-2023
Samsung (South Korea)
2018-2021
Seoul National University
2021
Dalian Maritime University
2020
National University of Defense Technology
2020
Integer (United States)
2020
We present a multimodal data set for the analysis of human affective states. The electroencephalogram (EEG) and peripheral physiological signals 32 participants were recorded as each watched 40 one-minute long excerpts music videos. Participants rated video in terms levels arousal, valence, like/dislike, dominance, familiarity. For 22 participants, frontal face was also recorded. A novel method stimuli selection is proposed using retrieval by tags from last.fm website, highlight detection,...
Abstract Over the last decade, neural networks have reached almost every field of science and become a crucial part various real world applications. Due to increasing spread, confidence in network predictions has more important. However, basic do not deliver certainty estimates or suffer from over- under-confidence, i.e. are badly calibrated. To overcome this, many researchers been working on understanding quantifying uncertainty network’s prediction. As result, different types sources...
This paper reviews the 2nd NTIRE challenge on single image super-resolution (restoration of rich details in a low resolution image) with focus proposed solutions and results. The had 4 tracks. Track 1 employed standard bicubic downscaling setup, while Tracks 2, 3 realistic unknown downgrading operators simulating camera acquisition pipeline. were learnable through provided pairs high train images. tracks 145, 114, 101, 113 registered participants, resp., 31 teams competed final testing...
The adsorption equilibria of CO2 on zeolite 13X and X/activated carbon composite (Zeocarbon) were measured by a static volumetric method. equilibrium experiments conducted at (273.15, 293.15, 313.15, 333.15, 353.15) K pressures up to 102.0 kPa for 99.7 Zeocarbon. experimental data obtained correlated the Toth, UNILAN, Sips models, which are generally used microporous adsorbents such as zeolites activated carbon. isosteric enthalpies calculated both adsorbents.
This paper examines the conditions under which exploration of a new, incompatible technologyis conducive to firm growth in presence network externalities. In particular, this study is motivated by divergent evolutions PC and workstation markets response new technology: reduced instruction set computing (RISC). market, Intel has developed microprocessors maintaining compatibility with established architecture, whereas it was radically replaced RISC market. History indicates that unlike market...
With the advances in understanding perceptual properties of human visual system and constructing their computational models, efforts toward incorporating mechanisms video compression to achieve maximal quality have received great attention. This paper thoroughly reviews recent mainly terms three major components, namely, model definition, implementation coding, performance evaluation. Furthermore, open research issues challenges are discussed order provide perspectives for future trends.
Emotion recognition based on electroencephalography (EEG) has received attention as a way to implement human-centric services. However, there is still much room for improvement, particularly in terms of the accuracy. In this paper, we propose novel deep learning approach using convolutional neural networks (CNNs) EEG-based emotion recognition. particular, employ brain connectivity features that have not been used with models previous studies, which can account synchronous activations...
We present a study of subjective and objective quality assessment compressed 4K ultra-high-definition (UHD) videos in an immersive viewing environment. First, we conduct evaluation experiment for UHD by three state-of-the-art video coding techniques, i.e., Advanced Video Coding, High Efficiency VP9. In particular, aim at investigating added values over conventional high definition (HD) terms perceptual quality. The results are systematically analyzed various viewpoints, such as scheme,...
Motion blur is a common photography artifact in dynamic environments that typically comes jointly with the other types of degradation. This paper reviews NTIRE 2021 Challenge on Image Deblurring. In this challenge report, we describe specifics and evaluation results from 2 competition tracks proposed solutions. While both aim to recover high-quality clean image blurry image, different artifacts are involved. track 1, images low resolution while compressed JPEG format. each competition, there...
Recently, learned image compression methods have out-performed traditional hand-crafted ones including BPG. One of the keys to this success is entropy models that estimate probability distribution quantized latent representation. Like other vision tasks, most recent are based on convolutional neural networks (CNNs). However, CNNs a limitation in modeling long-range dependencies due their nature local connectivity, which can be significant bottleneck where reducing spatial redundancy key...
Human-Object Interaction (HOI) detection, which localizes and infers relationships between human objects, plays an important role in scene understanding. Although two-stage HOI detectors have advantages of high efficiency training inference, they suffer from lower performance than one-stage methods due to the old back-bone networks lack considerations for perception process humans interaction classifiers. In this paper, we propose Vision Transformer based Pose-Conditioned Self-Loop Graph...
The involvement of TRPV1 and TRPA1 in mediating craniofacial muscle nociception mechanical hyperalgesia was investigated male Sprague-Dawley rats. First, we confirmed the expression masseter afferents rat trigeminal ganglia (TG), provided new data that is also expressed primary innervating masticatory muscles double-labeling immunohistochemistry experiments. We then examined whether activation each TRP channel evokes acute nocifensive responses leads to development hypersensitivity...
Scalable video coding is a powerful solution for content delivery in many interactive multimedia services due to its adaptability varying terminal and network constraints. In order successfully exploit such adaptability, it necessary understand users' preference among various scalability options consequently develop an optimal bit rate adaptation strategy. this paper, we present study of subjective quality assessment scalable coding, which investigates the influence combination on perceived...
Single-image super-resolution aims to generate a high-resolution version of low-resolution image, which serves as an essential component in many image processing applications. This paper investigates the robustness deep learning-based methods against adversarial attacks, can significantly deteriorate super-resolved images without noticeable distortion attacked images. It is demonstrated that state-of-the-art are highly vulnerable attacks. Different levels different analyzed theoretically and...
Several image-based biomedical diagnoses require high-resolution imaging capabilities at large spatial scales. However, conventional microscopes exhibit an inherent trade-off between depth-of-field (DoF) and resolution, thus objects to be refocused each lateral location, which is time consuming. Here, we present a computational platform, termed E2E-BPF microscope, enables large-area, of large-scale without serial refocusing. This method involves physics-incorporated, deep-learned design...
Single image super-resolution (SR) have recently shown great performance thanks to the advances in deep learning. In middle of networks for SR, a part that increases resolution is required, which subpixel convolution layer known as an efficient way. However, we argue method has room improvement, and propose enhanced upscaling module (EUM), achieves improvement by utilizing nonlinear operations skip connections. Employing our proposed EUM, novel residual network called EUSR. Our EUSR was...
This paper proposes a novel graph signal-based deep learning method for electroencephalography (EEG) and its application to EEG-based video identification. We present new methods effectively represent EEG data as signals on graphs, learn them using convolutional neural networks. Experimental results identification responses obtained while watching videos show the effectiveness of proposed approach in comparison existing methods. Effective schemes signal representation are also discussed.
Understanding music popularity is important not only for the artists who create and perform but also music-related industry. It has been studied well how can be defined, what its characteristics are, whether it predicted, which are addressed in this paper. We first define eight metrics to cover multiple aspects of popularity. Then, analysis each metric conducted with long-term real-world chart data deeply understand real world. build classification models predicting using acoustic data. In...